Editing text with mined

Mined is always in insert mode. Commands are single control characters, double key commands starting with ESCAPE, and a collection of function keys (for various types of keyboards and terminals). As a specialty, note the prefixing 'HOP KEY' which amplifies the effect of certain commands "just as you would expect"; this provides for more command flexibility without having to remember too many keys. It is described in a separate section below.

Keypad layout

Control key layout for basic movement functions is topographic on the left-hand side of the keyboard (an idea originating from early editors, when keyboards didn't have cursor keypads). (Although using a cursor block is more comfortable, a simple set of control key assignments is useful as a fallback on terminals or remote connections with reduced functionality.)

The right-hand cursor block of typical keyboards is assigned the most important movement and paste buffer functions.

Keypad assignment features:

central placement of HOP key (see below)

integration of frequently used copy/paste functions

7 Mark	8 ^	9 PgUp
4 <–	5 HOP	6 –>
1 Copy	2 v	3 PgDn
0 Paste		. Cut

<p> <pre> +------+------+------+ | (7) | (8) | (9) | | Mark | ^ | PgUp | +------+------+------+ | (4) | (5) | (6) | | &lt;- | HOP | -> | +------+------+------+ | (1) | (2) | (3) | | Copy | v | PgDn | +------+------+------+ | (0) | (.) | | Paste | Cut | +------+------+------+ </pre>

Note that the mined keypad function assignment as shown here deviates from the more usual assignment of Home/End to "move to beginning/end of line" and Del to "delete character". This is deliberately designed to provide more useful functions to easily available keys, while e.g. line movement can also easily be achieved with HOP cursor-left or HOP cursor-right, respectively, and character deletion can still be done with the Del key on the smaller keypad.
This keypad function assignment gives you the best benefit of keypad usage and is thus considered much more useful than the commonly expected "standard assignment" although now and then a user is irritated by it.

As there is often a conflict between the mined keypad assignment and commonly expected function assignments of some keypad keys, mined tries to conciliate this issue as follows:

• The more common Home/End/Del function assignments (line navigation and character deletion) to the respective keypad keys are also easily accessible (Alt-Home/End/Del).
• The -k option switches to the more common assignment by exchanging the unshifted keypad assignments with the Alt- assignments, keeping the mined assignment available with Alt.
• Using Del without a paste buffer gives an additional hint on alternative usage.
• Mined assigns different functions to the Home/End/Del keys on the numeric keypad and the similar keys on the small keypad (whenever possible with the terminal) in order to avoid the waste of resources by these usual redundancy of these two keypad blocks.
  By default, the numeric keypad will be assigned the paste buffer functions while the small keypad will be assigned the line navigation and character deletion functions.
• Regardless of -k mode, the more common function assignments (line navigation and character deletion) are always assigned to Control-Home/End/Del, while the paste buffer functions are always assigned to Shift-Home/End/Del (at least on the small keypad).
• Note: Shifted keypad key assignments may require proper keyboard configuration, see Keypad configuration for details.

The HOP function

This function, triggered by any of the HOP keys, amplifies (or modifies) functions as listed below. To achieve the combined function, first press any key that is assigned the HOP function, then any key assigned the second function:

HOP char left	move cursor to beginning of current line
HOP char right	move cursor to end of current line
HOP line up	move cursor to top of screen
HOP line down	move cursor to bottom of screen
HOP scroll up	scroll half a screen up
HOP scroll down	scroll half a screen down
HOP page up	move to beginning of file
HOP page down	move to end of file
HOP word left	move cursor to previous ";" or "."
HOP word right	move cursor to next ";" or "."
HOP delete tail of line/line end	delete whole line
HOP delete whole line	delete tail of line
HOP delete previous character	delete beginning of line
HOP set mark	go to mark
HOP search	search for current identifier
HOP search next	repeat previous (last but one) search
HOP copy/cut	copy or cut, but append to buffer
HOP save buffer	save buffer, but append to file
HOP paste buffer	paste "inter-window buffer", which is the last saved buffer by any invocation of mined on the same machine by the same user.
HOP edit next file	edit last file
HOP edit previous file	edit first file
HOP exit current file	exit mined
HOP suspend	suspend without writing file
HOP show status line	toggle permanent status line
HOP enter HTML tag	embed copy area in HTML tags

While a pull-down or pop-up menu is opened, any HOP key or the Space key or the middle mouse button toggles the HOP amplifier for a function subsequently invoked in the menu; the menu redisplays with function names changed where applicable.

Character-oriented navigation and editing

From the traditional restriction of Unix tools to the line as a unit of operation, other editors have derived a line-oriented movement and insertion paradigm which is a nuisance for anyone who wants an editor with decently intuitive operation.
Mined handles the end-of-line character like any ordinary character during movement and editing operations. Also search and replace strings can contain line ends.

Mouse control and menus

All versions of mined (Unix, DOS/Windows) support mouse operation.
Mouse control operates on pull-down and pop-up menus, flags, the text area, the bottom line, and the scroll bar, in order to provide the most useful functions and menu-driven command selection at hand.

Summary of mouse functions:

In text area:
• left click
  moves the text cursor to the mouse position
• left click-drag-release
  selects a text area and copies it to the paste buffer
• middle click
  display the text status line
• right click
  pops up the quick menu
On scroll-bar:
• left click
  moves one page down
• middle click
  moves to text position corresponding to cursor
• right click
  move one page up
On bottom line (status line):
• left click
  moves one page down
• middle click
  displays the text status line
• right click
  move one page up
On pull-down menu header (in left menu area of upper line):
• left or right click or mouse scroll
  opens menu
• middle click
  opens menu with HOP-modified functions
On flag indication (in right flag area of upper line):
• middle click
  toggles flag
• left click (deprecated)
  toggles flag (should open menu in a future version)
• right click or mouse scroll
  opens flag menu

Configuration hint: To enable mouse operation in a Windows DOS box, deactivate "QuickEdit mode" in the properties menu.

Menus

Mined provides three kinds of menus, all can be opened with either mouse clicks or commands. The menus offer the most important editing functions (apart from simple movement). Some menus have their items grouped into sections, some of which have subtitles.
The HOP flag can be toggled while a menu is open with any of the HOP key, ^G, Space, or the middle mouse button. When a pull-down menu is opened with the middle mouse button, the HOP variation is initially triggered, offering the HOP variations of the menu items.
The three menu groups are used as follows:

• A pull-down menu is opened by clicking the mouse on the menu header (in the left part of the top screen line) or scrolling the mouse wheel on this header.
  Shortcut: Each pull-down menu can also be opened with ESC or Alt and the small initial letter of the menu header (Alt-f or ESC f for the file menu etc.).
• A flag menu is opened by clicking the right mouse button on a flag indication in the flags area (right part of the top screen line) or scrolling the mouse wheel on it. The flag menus have optional markers in front of each item showing which items are currently active.
  Shortcut: The Info menu, Input Method (Keyboard Mapping) menu, Smart Quotes menu, Encoding menu can also be opened with Alt-F10, ESC I or Alt-I, ESC K or Alt-K, ESC Q or Alt-Q, ESC E or Alt-E, respectively.
• The pop-up menu is placed above the text area and can be opened with a right-click or Alt-Space (ESC Space).

Menu navigation

When a menu is open, the cursor-left or cursor-right keys cycle through the pull-down and flag menus. Alt-cursor-left and Alt-cursor-right navigate quickly between the two sets of menus (pull-down or flag menus).
When a sub-menu is open, cursor-left goes back to the parent menu, cursor-right opens its next menu to the right.

There are three methods to navigate within a menu:

• With the keyboard: open menu as described above, navigate with cursor keys or by typing the first letter of the desired menu item (which cycles through all items starting with that letter), activate with Enter key.
• With mouse clicks: open menu with click (and release) mouse button, switch to other menu with another click, click on item to activate it. The mouse wheel may be used to navigate menu items.
• With mouse dragging: open menu with mouse button (left or right), browse menus and items with button held down, activate selected item with releasing mouse button.

Methods may be mixed, e.g. open a menu with either mouse click or keyboard, navigate with mouse wheel, then select with Enter.

Scrollable menus: In a low-height terminal (e.g. 24 lines), longer menus (especially the Encoding menu and the Input Method menu) may not fit on the terminal. All menus are scrollable with cursor keys, including Page Down/Up, Home, End keys.
When the window size is changed, open menus are closed in order to prevent resizing and repositioning problems; this is planned to be enhanced in a future version.

Hints

Note: Your mouse driver may be configured to generate multiple (e.g. 3) mouse wheel events on one mouse wheel movement (e.g. with Windows).

Layout configuration: See Menu display below for configuration of menu appearance.

Configuration hint: On Unix, in order to make Alt work as a modifier, set the xterm resource metaSendsEscape to true and the rxvt resource meta8 to false as suggested in the example file Xdefaults.mined in the Mined runtime support library. (With older versions of xterm, setting eightBitInput to false may be required instead; this xterm option doesn't actually disable 8 bit input as its name might suggest.) With xterm, this setting can also be enforced dynamically with the +D option.

Inter-window paste buffer

Mined can perform copy/paste operations within different editing sessions (parallel or subsequent invocations of mined): The command HOP Ins (e.g. ^G ^P) will insert the most recent paste buffer copied or cut in any of the user's mined sessions. This can also work remotely in a network; to configure this features, see Common paste buffer configuration.

Multiple paste buffers

Mined provides emacs-style multiple paste buffers that are organised as a buffer ring. Every buffer cut or copy operation (that places the text between the marked and the current position to the buffer) creates a new buffer and stacks it to the list of buffers. If the feature "deleted word/line appends to buffer" is enabled (+VV) the commands delete-end-of-line (^K), delete-word (^T) and delete-end-of-sentence (currently emacs mode only) append to the top buffer (disabled with the option -VV).
To paste a non-top-most buffer, paste the most recent buffer first as usual, then use the buffer-ring command (Alt-Ins-key or Control-F4, or M-y in emacs mode) to exchange the pasted text with the previous buffer. This can be repeated, going down the stack of buffers, and at its bottom, starting over from the top again.

Text position markers

A default marker for quick use and additional 10 numbered text markers are available.
Marker 0 has a special function: 1. it is set when opening a file at the memorized position, 2. whenever a new current marker is set, the previous one is pushed to marker 0.

Text position marker stack

In addition to the explicit text markers, mined implicitly maintains a marker stack to support navigation and orientation when browsing files. Whenever a command moves the position by a far distance (Go to marker, Go to line, Go to file beginning/end, Go to next/previous file, Search functions including Search identifier definition across files, Replace with confirm), the current position is first pushed to this stack. Later, in order to return to the previous position, use the command ESC Enter (Alt-Enter) to move along the positions in the marker stack. The command HOP ESC Enter (HOP Alt-Enter) moves again forward along the stack.

Paragraph justification / word wrap

Manual paragraph line/word wrap is invoked with the justify command (ESC j or ESC J); it justifies the current paragraph (wraps its lines/words) according to the effective margins and paragraph termination mode.
Clever justification: With ESC j, mined automatically determines left margins depending on the current paragraph and line contents. Heuristic detection of numbered items will trigger automatic indentation.
Normal justification: With ESC J, mined justifies strictly according to the margin values currently configured.
See commands listing below "ESC j" for margin setting commands.

Paragraph termination modes: Two different definitions of paragraph end are available.

• The primary mode is to add a space at the end of each line when the paragraph continues and to end the line without space where the paragraph ends. This seems an intuitive way and as a big advantage over other approaches, it is transparent with respect to visual formatting, i.e. no text property is required that would affect visual layout of the text.
  Note: Additional visual support of paragraph end detection is available with the mined option -p that distinguishes paragraph/line end display.
• The other word-wrap mode is to add an empty (blank-only) line after each paragraph. Obviously this imposes more additional requirements on text formatting discipline and reduces freedom of text layout.

The mode in effect is indicated in the mode indication display; see description there.

Auto indentation

By default, mined acts in auto-indent mode: When you enter a newline, the following line will be filled with the same prefix of space characters (Space or Tab) as the current one. This option can be toggled from the Options menu. A new line without auto indentation can be entered with the ^O command.

Auto indentation is automatically suppressed if text is entered very fast (by heuristic detection of input speed) in order to allow unmodified copy and paste using terminal mouse functions.

Structure input commands

A pair of parentheses with matched indentation can be entered by prefixing a parenthesis character with HOP. For example, HOP "{" would enter a pair of "{" "}", both auto-indented on their respective new line. Other pairs are "(" ")", "[" "]", "<" ">".
HOP "/" enters an indented Javadoc comment frame.

Back-Tab (Undent function / reverse indent)

A Backarrow key from a position that is only preceded by white space on the line and on the line above will revert the input position to the previous matching indentation level. To avoid auto-undentation ("Delete single"), use Control-Backarrow or F5 Backarrow to delete only one character left. (Control-Backarrow only works if configured in your X configuration, see the example configuration file Xdefaults.mined in the Mined runtime support library.)

Tab expansion

With one of the options -+4 or -+8, a Tab key input will be expanded to an appropriate number of Space characters instead of inserting a Tab character. You can still insert a literal Tab character with Control-V Tab.

Search and replace multiple lines

Mined has overcome the typical Unix tool limitation of line orientation in search operations. Search and replacement patterns can contain embedded newlines. Enter a linefeed character in the search string with ^V^J or \n. (In some cases there are still display problems; then update the screen with the ESC "." command.)

Header line underlining

The command HOP "-" (e.g. Control-G -) underlines the header line before the cursor position with as many "-" characters as needed; it applies to the current line unless the cursor is at a line beginning in which case it applies to the previous line.

Automatic backspace mode adaptation

There is much confusion about what character codes are delivered by the Backarrow and Del keyboard keys in different operating environments and configurations. For proper operation, the "stty erase CHAR" configuration should generally be set correctly to reflect the actual code emitted by the terminal. Mined detects this setting and adjusts its handling accordingly, so that the "Backarrow" key should normally work as expected (delete a character left).

Overview: input support features

Character input

Mined provides several methods to support input of special characters that may not be easily available on the keyboard.

• Accented and mnemonic input support defines Accent prefix keys to compose accent combinations with subsequently entered characters.
• It also provides Character input mnemonics for easily memorisable input of a wide range of characters, including most composed Unicode characters.
• Input support commands include a quick shortcut for two-character mnemonics.
• Input support commands also provide for character input by hexadecimal / octal / decimal character code or Unicode value, including support for subsequent entry of multiple numeric characters according to ISO 14755.
• Keyboard mapping switching the keyboard to support another script. This mechanism also provides CJK input methods.

Structured input

• HTML tag input (starting/closing or embedding marked text).
• Auto indentation and Back-Tab.
• Structure input commands: Input of indented matching parentheses and Javadoc frames.
• Paragraph justification (line/word wrap).
• Header line underlining

Special features

• Smart quotes automatic transformation of entered straight quote marks into typographic quotation marks (style can be selected in flags area), as well as smart dashes and other smart text replacements.
• Right-to-left script input support.

Handling files with mined

Tags file support

The ESC t command moves to the definition of an identifier (on which the cursor should be placed) using the tags file (generated by the ctags command). HOP ESC t prompts for an identifier. (Also available from search or popup menu.) If a new file is opened for this purpose, the current file is saved automatically.
Like with a number of positioning commands, ESC t places the current position on the position marker stack before going to the location of the identifier definition. The command ESC Enter (Alt-Enter) moves back to that position, also saving the current file if needed first.

Data security

Mined has a robust and defensive concept of handling edited text and file contents in case of any kind of program or system errors.

Edited text

Every care has been taken to prevent loss of the edited text in case of save errors or accidental quit commands etc; mined always prompts before discarding any modified text (not all popular editors are so careful about this, e.g. emacs when editing text without associated filename).
In the rare case of an unrecoverable error (out of memory or terminal I/O error) or if mined is interrupted by an unexpected signal, mined needs to terminate but it tries to save the edited text (if modified) into a panic file in one of the directories $MINEDTMP, $TMPDIR, $TMP, $TEMP, /usr/tmp, or /tmp (whichever variable is defined first and directory is writable in this order). If possible, mined also tries to continue normally after panic handling unless multiple external signals are nested. Only if the temporary area happens to be full and mined cannot continue either you would be out of luck.
If mined is sent an explicit SIGTERM signal it tries to terminate normally, writing modified text to the file being edited (this would involve normal interactive handling if that file is read-only or the file name was changed).

Files

Also, if any command is issued to write to a file not previously read in (after change of file name or working directory, or with a Copy to file command), mined prompts for confirmation.

File access permissions

When creating a new file, its access permissions are set according to the default behaviour set in the user environment (umask setting in Unix). However, when cloning a file (with Save As / Set Name / ESC n / ESC d), file access permissions of the originally opened file are preserved and cloned.
The +x command line option adds executable permission to newly created files but only to those users that are also given read permission by the rules above.

Pipe output

In the "write to standard output" mode (i.e. when invoked within a pipe), only one "file save" operation can be performed writing to standard output. If more than one such operations are issued (e.g. using the ESC w / F2 , F3, or suspend command) only the first one will write the text buffer to standard output; any subsequent one is treated as usual (with empty file name).

Line end modes and binary-transparent editing

Mined is binary transparent. It can handle all types of line ends (Unix, DOS, optionally Mac, and Unicode separators) simultaneously in the same editing session. They are indicated by different visible line end indications. Files without trailing line end can be edited and created (using the delete character right function on the last line end). NUL characters are handled as virtual line ends. Lines too long for internal handling are split transparently (with a "none" virtual line end).
Note: Character codes that are illegal in the currently selected text encoding are maintained transparently and are clearly indicated (e.g. illegal UTF-8 sequences in Unicode text).
Files with mixed encoding (e.g. UTF-8 / 8 bit sections) can be edited comfortably.
Input: To enter a NUL character, use ^V # 0 or ^V < NUL or Control-Space > (if the keyboard supports the latter).

Memory of file position and editing style parameters

If the current directory contains a file named @mined.mar , file position memory is enabled.
The current cursor position is stored with every file save command (even if no write is performed because the text has not been edited). When editing that file again, mined will automatically move to that position (and set text marker 0 to it). (The association of the position is not with the file itself but with its relative name from the current directory.)
This mechanism is enabled in each directory by using the command "Save Position" from the File menu, or by using Control-F2 to save a file or by prefixing any file writing command with HOP. This enforces creation of the marker file.

In addition to the current position, mined also stores the paragraph justification margins (only if automatic paragraph justification is active) and the selected Smart Quotes style.

Page length

The command ESC P sets the number of lines that mined assumes to be on a page. So the status line can contain the page number to make finding the current position in a print-out easy. Also the Goto Line/% command (^G etc.) accepts a final 'p' or 'P' in which cases it positions to the top of the given page. This information will be associated and stored with the file name if file position storing is active (i.e. if the file @mined.mar exists in the current directory).

File names

When entering file names, the leading ~/ notation to refer to one's home directory is accepted.

Restricted mode (tool mode)

Restricted mode is activated with

		mined -- [ filenames ... ]

In restricted mode, only the file opened when mined was started can be edited, no commands changing file name reference, involving other files (copy/paste), or escaping to a shell command will be allowed. (When mined is invoked without filename argument, a file name will be prompted for despite restricted mode, however.)

Version control integration

From the File menu, checkout and checkin commands are available that invoke "co" or "ci" scripts, respectively (which must reside in the user's command search path). This offers a gateway to ClearCase or other version control systems; mined applies automatic save or screen update as appropriate.

Printing

From the File menu, a print command is available that prints the text currently being edited. If the script uprint is installed and configured properly, printing works in any selected character encoding. See Printing configuration for further details.

Working with mined

Mode indication flags

The right side of the top menu bar displays a number of one-letter or two-letter indications for certain modes; the associated flag menus can be opened from here with a mouse right-click, or the modes can be toggled quickly with a middle-click. (Keyboard shortcuts for handling flags and menus are also available.)

• Information display mode
  • "?": this flag menu offers options for permanent File info, Char info, or Han character information display. For the latter, further options can be selected to configure the information shown.
• (In non-Latin-1 text and terminal mode only) Input Method (Keyboard Mapping)
  • "--": no keyboard mapping is active.
  • "...": a two-letter input method tag indicates that an according keyboard mapping is active, mapping keyboard input to characters of the selected Unicode script range, or using a more complex CJK input method involving "pick list" selection menus. See Keyboard Mapping and Input Methods below.
  • Right mouse button on this indication opens a menu for selection of the desired keyboard mapping.
  • Left mouse button on this indication toggles between the current and the previous selected keyboard mapping.
• Smart Quotes
  • Two quote marks are displayed that act as automatic "smart quotes": When you type a «"» or «'» character (straight double or single quote), it is replaced by an opening or closing typographic quote mark (double or single, respectively), depending on the text context.
  • Right mouse button on these indications opens a menu for selection of the desired quotation marks style.
  • Left mouse button on this indication toggles between the current and the previous style selected with the menu.
• Character encoding (used for text interpretation)
  • A two-letter character encoding tag indicates the text encoding currently assumed for display. Changing the encoding changes the interpretation of the text which is otherwise handled transparently; it does not recode the text.
  • Right mouse button on these indications opens a menu for selection of the desired quotation marks style.
  • Left mouse button on this indication toggles between the current and the previous selected encoding.
  Note: See Character encoding support below for a list of encodings that are auto-detected.
  Note: For hints on pre-selecting preferred text encoding (as well as terminal encoding) and a note on adjusting the available encodings and configuring the Encoding menu, see Locale configuration.

  • "U8": Unicode/ISO 10646 character set / UTF-8 encoding
  • "16" or "61": Unicode character set / UTF-16 encoding (big-endian or little-endian, respectively)
    In contrast to the other encodings, UTF-16 has no separate entry in the Character encoding menu as its internal handling is UTF-8 and cannot be switched while editing; these two flag values only indicate that the file being edited was found to be encoded and will be saved in UTF-16.
  • "L1": Western "Latin-1" character set / ISO 8859-1
  • "WL": Windows Latin character set / "codepage" 1252 (superset of Latin-1)
  • "L9": Western "Latin-9" character set (with Euro sign) / ISO 8859-15
  • "Cy": Cyrillic character set / KOI8-RU encoding (Russian, Ukrainian, Bjelorussian)

  sub-menu more Cyrillic:

  • "Ru": Cyrillic / Russian KOI8-R encoding; used if locale environment indicates this as terminal encoding, not in menu, use "Cy" instead which combines KOI8-R and KOI8-U
  • "Uk": Cyrillic / Ukrainian KOI8-U encoding; used if locale environment indicates this as terminal encoding, not in menu, use "Cy" instead which combines KOI8-R and KOI8-U
  • "I5": Cyrillic / ISO 8859-5 encoding
  • "WC": Cyrillic / Windows Cyrillic encoding
  • "Tj": Cyrillic / Tadjikistan encoding
  • "Kz": Cyrillic / Kazachstan encoding
  • "GP": Georgian character set (not Cyrillic) / Georgian-PS encoding

  sub-menu Greek/Oriental:

  • "I7": Greek / ISO 8859-7 encoding
  • "I6": Arabic / ISO 8859-6 encoding
  • "Ar": Arabic / MacArabic encoding (superset of ISO 8859-6)
  • "I8": Hebrew / ISO 8859-8 encoding
  • "He": Hebrew / Windows codepage 1255 (superset of ISO 8859-8)

  sub-menu more Latin:

  • "MR": Mac-Roman character code
  • "PC": PC DOS character code ("codepage 437")
  • "PL": PC Latin character code ("codepage 850")
  • "LN" where N is 2..8 or "0": Latin-N or Latin-10 encodings / ISO 8859-2/3/4/9/10/13/14/16

  CJK encodings:

  • "B5": Traditional Chinese character set / Big5 encoding with HKSCS extensions
  • "GB": Simplified Chinese character set / GB18030 encoding, includes GBK encoding, includes GB 2312 / EUC-CN encoding
  • "CN": Traditional Chinese character set / CNS / EUC-TW encoding (including 4-byte code points)
  • "JP": Japanese character set / JIS X 0208 / 0212 / 0213 / EUC-JP encoding (including 3-byte code points)
  • "sJ": Japanese character set / Shift-JIS encoding (including single-byte mappings to Halfwidth Forms)
  • "KR": Korean Unified Hangul character set / UHC encoding, includes KS C 5601 / KS X 1001 / EUC-KR encoding
  • "Jh": Korean Johab character set and encoding

  Further Asian encodings:

  • "VI": Vietnamese character set / VISCII encoding
  • "TV": Vietnamese character set / TCVN encoding
  • "TI": Thai character set / TIS-620 encoding

  Combining display (available only if the current text encoding contains combining characters)

  • "ç": combined display mode
  • "`": separated display mode: combining characters are separated from their base character and displayed with coloured background

  HOP key active

  • "H": HOP applies to next command
  • "h": HOP not active

  Edit mode vs. View only mode

  • "E": text is being edited
  • "V": text is being viewed (modification inhibited)
  • Note: this is not related to a file being read-only; if you "edit" and modify the text of a read-only file, you will have to save to a different file name (or discard)

  Paste buffer / append mode

  • "=": cut/copy replaces (overwrites) paste buffer
  • "+": cut/copy appends to paste buffer
  • "=": like "=", and indicates Unicode paste buffer mode
  • "+": like "+", and indicates Unicode paste buffer mode

  Auto-indent mode

  • "»": auto-indentation enabled: entering a newline indents the following line like the current one
  • "¦": auto-indentation disabled

  Automatic paragraph justification levels

  • "j": justification only on request (ESC j command)
  • "j": justification is performed whenever text is entered beyond the right margin
  • "J": justification is performed whenever text is inserted and the line exceeds the right margin (slightly buggy)

  Paragraph termination definition effective for justification

  • " ": non-blank line end terminates paragraph (blank space at line end continues paragraph)
  • "«": empty line terminates paragraph

Scrollbar

By default, mined displays a scrollbar at the right side. It may be used for position indication within the text and for relative or absolute positioning with the three mouse buttons.
In a UTF-8 terminal, mined uses Unicode character cell vertical eighths characters U+2581..U+2587 for a fine-grained scrollbar display. If your Unicode font doesn't include those block characters, you may switch to the cell-grained scrollbar with the -o1 option.

Text position marker stack

On commands that jump away from the current position (HOP Mark, File Begin/End, Search, Search identifier definition, Search current character, Goto Line/%, Goto Next/Previous File), the current position is remembered in a position stack. The command ESC Enter goes backward, HOP ESC Enter forward in this "stack", even if this means switching the file being edited.

Structured editing support

HTML support: syntax highlighting and tag entry/matching

HTML tag entry: With the ESC H commands, opening and closing HTML tags can be entered or (with HOP) a marked area can be enclosed into HTML tags.
Syntax highlighting: HTML tags are displayed in light blue colour to set them back from the actual text contents. Other highlighting modes apply to HTML comments and JSP code. This option is activated if the file name suffix is one of .html, .htm, .xml, .jsp, .sgml; it can be toggled from the Options menu.
HTML tag matching: With the ESC ( or ESC ) command, mined searches for the opening / closing HTML tag corresponding to the current one.
Note: While you edit within a line and change its HTML ending status (by entering or deleting '<' or '>'), the display status of subsequent lines is not changed. (You may refresh the display with ESC ".")
Configuration hint: The colour used for displaying HTML tags can be configured with the environment variable MINEDHTML using an ANSI sequence, e.g. MINEDHTML=34 (the default).

Search structure match

With the ESC ( or ESC ) commands, mined searches for a matching end of various structures, like opening/closing HTML/XML tags (see above), matching parentheses or brackets, matching comments (/* */), matching conditional macros (#if...), mail messages (in a mailbox file), MIME attachments. See the ESC ( command in the command reference for details.

Structure input

A structure template with opening and closing ends can be inserted with the structured input feature. HOP followed by one of { , ( , [ , < enters a corresponding bracket pair, HOP / enters a Javadoc comment frame. HOP – enters an underlining line matching the previous line.

Visual structure input is supported by Auto indentation

Password hiding

With the option -P, mined hides one word (separated by white space) behind the string "assword" in a line (to accommodate for "password" or "Password") and displays reverse "*" instead. Password hiding can be disabled with +P.
By default (without any P option), password hiding is activated when editing a file whose file name starts with "." (Unix "hidden" file convention).

Long line splitting

Mined has an internal line length limit (> ca. 1024 characters). When opening a file, longer lines are split. This is handled transparently as virtual "none" line ends are used and indicated. When saving the file, lines will be joined again.

Visible indication of line contents and display

Various options are available to indicate line control characters (Tab and line-feed) as well as shifted line display (of lines longer than the screen width). (So you can see how many dummy blank spaces there are before the line ends or how many superfluous blank spaces precede a Tab character.)
Environment variables can be used to modify these indications. See Display of contents indications and scrollbar for details.
Default indications and according configuration variables:

«	LF (Unix-type line end) change with MINEDRET or MINEDUTFRET (may contain up to 3 characters to configure different appearance)
µ	CRLF (MSDOS-type two-character line end) change with MINEDDOSRET or MINEDUTFDOSRET
@	CR (Mac-type line end) transparently handled and displayed with +R command line option
º	NUL character (pseudo line end)
¬	"none" line end (virtual line end as used to split input lines too long for internal handling; will be joined into a single line when saving the file)
·	non-breaking space (character code hex A0)
«	Unicode line separator
¶	Unicode paragraph separator change with MINEDPARA or MINEDUTFPARA
¶	end of paragraph (if enabled by -p) change with MINEDPARA or MINEDUTFPARA
»	line extending the end of the screen line (move cursor right to shift line display) change with MINEDSHIFT or MINEDUTFSHIFT
«	line shifted out left of the screen line (move cursor left to shift line display back) change with MINEDSHIFT or MINEDUTFSHIFT
·	position spanned by Tab character change with MINEDTAB or MINEDUTFTAB (may contain up to 3 characters to configure different appearance within the Tab span)

Configuration: Display colour of the indications which are by default red can be changed with the environment variable MINEDDIM, display colour for Unicode line end indications with MINEDUNIMARK. Their values should be the numeric part of an ANSI terminal control sequence, e.g. 31 for red, "33;44" for yellow text on blue background. For more details and recommended settings see the example script file profile.mined in the Mined runtime support library. Default values are compiled in and can be overridden by setting the variables to empty values.

Function key help bars

For quick reference of functions attached to function keys, modified function keys, and other modified keys (as used for accent prefix functions), a number of help bars can be displayed in the bottom line.
F1 followed by another F1, optionally modified by a combination of Control/Shift/Alt, displays a help line with function attachments to the respectively modified function keys; F1 followed by Control/Alt-1 or Control-, displays a help line for the respective accent prefix functions attached. See the F1 help bars command reference for details.

Menu display

Menu borders are displayed using Unicode Box Drawing characters in a UTF-8 terminal, using VT100-mode block graphics characters if they are detected to be available, or using ASCII graphics otherwise.
Configuration hint: The menu style option -Q is available to configure your style preference; see also Terminal interworking problems for configuration hints to deal terminal-related graphics display trouble.
In addition to round or rectangular corners, also fancy item selection display style can be selected (-Q).
With a non-UTF-8 terminal, if your system's termcap/terminfo database does not indicate the VT100 block graphics capability for the terminal you use but you know (or want to try if) your terminal has that capability, use of graphical borders can be enforced with the -Qv command line option.
Configuration hint: The marker of selected items in flag menus can be changed with the environment variable MINEDMENUMARKER. If it is empty, the default marker is changed to a nice preconfigured alternative.

Character handling support

This chapter describes mined features for character manipulation and display of characters and character properties. Unicode and CJK specific features are described in the respective chapters. Character input support is described separately in the subsequent chapter.

Script highlighting

It may be desirable to distinguish characters in different script by displaying their glyphs in different colours. (This especially allows to distinguish easier between similar glyphs as they occur in Latin/Greek/Cyrillic scripts.)
Script highlighting is currently pre-configured for Greek and Cyrillic. It uses the terminal's 256-colour mode if available.
The scripts to highlight and the colour values to use can be configured at compile-time. See Mined configuration below.

Combining characters

When editing text in Unicode or any encoding that contains combining characters, mined supports display and editing of combining and combined characters.

(Note: Terminal support for combining characters is auto-detected; additional command line options are available in case this fails.)
If mined operates on a terminal that handles combining characters, it offers two editing modes: combined or separated. They can be toggled by clicking the Combining display flag in the Mode indication flags area (right part of the top screen line), or by the menu entry "Options – Combined display"; separated display mode can also be selected by the command line option -c.

Combined display and editing mode (Combining display flag ç)

Combined characters are displayed as intended (i.e., combined).

Micro movement into combined characters:

• The cursor can be moved into a combined character with Control-cursor-left and Control-cursor-right, or ^V cursor-left and ^V cursor-right.
• You can determine the exact position of the cursor if permanent character info is switched on (by HOP ESC u or with HOP "Toggle Char info" in the Options menu).

Partially editing combined characters:

• If the cursor is on a combined character, delete next character will delete the whole combined character, with all combining accents.
• If the cursor is within a combined character, delete next character will delete the current combining accent only.
• Control-Backarrow or F5 Backarrow ("Delete single") behind or within a combined character will only delete the rightmost combining accent (preceding the cursor position) while Backarrow would delete the whole combined character.
• You can also position the cursor as described above and use copy-and-paste operations.

Note: Control-cursor-left and Control-cursor-right only work if these keys are configured to emit distinguished escape sequences with Control key held down. With xterm, this works by default. With rxvt, use the small keypad cursor keys, or enable Control on the right keypad with the sample configuration file Xdefaults.mined in the Mined runtime support library. With mlterm, enable this with the sample configuration file mlterm_key in the Mined runtime support library. Control-Backarrow can also be configured to work with xterm but doesn't appear to work with rxvt or mlterm, use F5 Backarrow instead.

Separated display and editing mode (Combining display flag `)

Combined characters are separated into base character and combining character(s) for display and editing. Combining characters are indicated with coloured background.

• In separated display mode, all cursor and text modification operations work on the combining parts as displayed.

Note: Unicode combining characters (according to the most recent version of Unicode known to mined) that are not handled as combining characters by the terminal (which might implement an older version of Unicode) are always shown like in separated display mode.

Character information display

The command ESC u displays character encoding information in the bottom status line (conforming to ISO 14755); it displays the character code in the selected encoding (UTF-8 byte sequence in UTF-8 mode) and the ISO-10646 (Unicode) value of the current character, as well as Unicode script range and character category, width, and combining information. The Unicode value is displayed with 4 hexadecimal digits if the character is in the Unicode BMP (Basic Multilingual Plane, 16 bit), with 6 digits if it is a Unicode character outside of the BMP, and 8 digits if it is an ISO-10646 character outside of the Unicode range. The information displayed also indicates all kinds of encoding irregularities.
For the Unicode data version used for character properties see the mined change log.

Permanent display of character information is toggled with HOP ESC u or by selecting "Char info" in the Info menu (or with HOP "Toggle Char info" in the Options menu).

Han character information display

CJK-specific character information (semantic and pronuciation hints) is described below in section Han character information display.

Character conversion features

Case conversion

The case conversion functions (ESC C, HOP ESC C, F11, HOP F11, Shift-F3) cover the full Unicode range. They also handle special cases like Greek final sigma, optionally Turkish "i", case mapping to multiple characters, and Lithuanian special conditions. Japanese characters are toggled between Hiragana and Katakana by the same functions.
Shift-F3 cycles casing of a word between all small, title case (beginning capital), and all capitals. It handles title casing, using Unicode title case characters for the first character when appropriate. For Japanese script, it toggles the word between Hiragana and Katakana.
The case mapping is based on the most recent Unicode version compiled into mined (for the actual version see the mined change log). It is applicable in all text encodings (since mined 2000.12).

Numeric conversion

Commands are available to insert characters corresponding to a hexadecimal character code or hexadecimal/octal/decimal Unicode value contained in the text, or to insert a respective value corresponding to the current character. For details, see the section in the command summary.

Mnemonic conversion

A character mnemonic at the cursor position can be replaced with its associated character. Use one of the Mnemonic character substitution commands (ESC _ or national variants) described below.

Encoding conversion support

A special feature offers interactive conversion between Latin-1 and Unicode character encoding, see Encoding conversion support in chapter Unicode support below.

Unicode Copy/Paste buffer

The Copy/Paste buffer can be operated in Unicode mode in which case it converts between text edited in different character encodings. See Unicode Copy/Paste buffer conversion below.

Smart quotes

Straight (double or single) quote characters «"» or «'» can be replaced automatically with an opening or closing typographic quotation mark, depending on the text context. Select the quotation marks style to be applied from the Smart Quotes selection menu (open with ESC Q or Alt-Q or right-click on the smart quotes indication in the flags area in the top screen line), or middle-click on the smart quotes flag to toggle between the current and the previous smart quotes style selected with the menu.
When a file is loaded, mined tries to determine the applicable quotation marks style in two ways: If mined edited the file before and noted the last cursor position (in the file @mined.mar, which can be created using the HOP F2 command, or the File menu "Save Position" command), this information also includes the last selected smart quotes mode for the file. If that information is not available, mined auto-detects existing quotation marks in the file and adjusts its smart quotes mode accordingly.
The smart quotes left/right selection algorithm considers the text context to automatically support smart quotes also in CJK text.
A typographic apostrophe can be inserted with HOP ' (^G ').
In smart quotes mode, straight quotes can be inserted with mnemonic compose pairs (^V ^ " or ^V ^ ' , or ^V "# or ^V '# respectively).
Smart quotes are applicable in all text encodings provided the desired quote marks are contained in the selected encoding.
Smart quotes style can also be preselected with the environment variable MINEDQUOTES which should then contain the opening/closing quote pair or just the opening quote mark (UTF-8 encoded, double or single quotes); this overrides both auto-detection and the preference saved with the cursor position.

Smart text replacements: smart dashes and arrows

If smart quotes are active, some other smart input text replacements are applied to sequently entered characters (unless during a repeat command entering multiple characters):

--	if preceded by a Space character: en dash (U+2013) otherwise: em dash (U+2014)
-	if an adjacent character is in the Hebrew script range: Hebrew hyphen mark Maqaf (U+05BE)
<-	leftwards arrow (U+2190)
->	rightwards arrow (U+2192)
<>	left right arrow (U+2194)

Character input support

Some character input support features support international scripts (especially with Keyboard Mapping and Input Methods), others mainly address composite characters. For the latter, it is useful to explain a few notions:

Combining character:

A character (usually in Unicode) that is defined to combine with the previous character into a combined character, to be displayed as a single glyph (visual unit).

Combined character:

The glyph combination of a Unicode character (base character) with one or more Unicode combining characters.

Composed character (or composite character):

A character that has one or more accents composed into it, or is otherwise composed of components, like the ae ligature, to be displayed as a single glyph. It can be a single Unicode character or a Unicode combined character consisting of a Unicode base character and one or two Unicode combining characters.

Accented character (or diacritic character):

A special case of a composite character where a letter is composed with one or more accents.

Compose key:

A number of system and keyboard vendors have equipped their keyboards with a "Compose" or "Combine" key. This key – when configured and interpreted properly by the operating environment – produces a composed character which is then provided as input to the application.

Accented and mnemonic input support

Function keys or character mnemonics can be used to enter accented or other composite characters. (This is also known as digraph function with some editors.)
These character composition functions also work on the prompt line.
(Any composite character configured on your keyboard can of course also be entered directly or using the Compose/Combine key of your keyboard.)

Note that mnemonic input and accent prefix keys can be combined in flexible ways, e.g.

^V ' Control-F6 e

or

F6 ^V e ^

which both enter U+1EBF (e with circumflex and acute)

Mnemonic input can be applied recursively to compose a character for further composition, e.g.

^V ' ^V a e

enters U+01FD (æ with acute)

Accent prefix keys can use an already precomposed base character for further composition; if this does not match an explicitly known mnemonic, the base character is decomposed first to find a match, e.g.

F6 ü

or

F5 ú

which both enter U+01D8 (u with diaeresis and acute)

Up to three accent prefix keys can be combined by entering them in sequence in order to compose characters with multiple accents, e.g.

F5 F6 u

enters U+01D8 (u with diaeresis and acute)

Control-2 Control-7 a

enters U+1EB1 (a with grave and breve)

Control-– Control-: u

enters U+1E7B (u with macron and diaeresis)

Control-, Control-( e

enters U+1E1D (e with cedilla and breve)

Alt-7 Control-, < alpha >

Alt-F6 Shift-F6 Control-, < alpha >

Control-< Control-` Control-, < alpha >

all enter U+1F82 (alpha with psili and varia and ypogegrammeni)

where < alpha > indicates the Greek letter alpha, which may e.g. be entered by selecting the Greek input method and typing the "a" key

Accent prefix keys

Preliminary note on using keys with Control, Shift, Alt modifiers: Especially for accented character input, mined makes use of key combinations modified with Control, Shift, Alt, or a combination of them. Some of these key combinations may be limited by local environment, especially the window system, or may need extra configuration to be enabled.

• Modified function keys are often intercepted by window systems for special functions. To enable modified function keys with KDE, e.g., like Control-F5/F6, Control-Shift-F5/F6, Alt-F5/F6, you have to press the "Window key" additionally at the same time.
• Modified digit keys as well as Control-modified punctuation keys are used as extended and intuitive accent prefix keys. To enable them, either use a recent version of xterm (216) or configure them with your terminal; configuration instructions for older versions of xterm can be found in the sample file Xdefaults.mined in the Mined runtime support library.
• Warning: The Alt-F4 key combination should not accidently be hit as many window managers use it to kill the terminal window!

The following table lists the accent prefix keys:

F5 (Sun: R4/-)	diaeresis (umlaut) / dialytika
Shift-F5 (Sun: R5/÷)	tilde / perispomeni
Control-F5 (Sun: R6/×)	ring / cedilla / iota (ypogegrammeni)
Alt-F5	stroke
Control-Shift-F5	ogonek / prosgegrammeni
Alt-Shift-F5	breve / vrachy
F6 (Sun: R3)	acute (accent d'aigu) / tonos
Shift-F6 (Sun: R1)	grave / varia
Control-F6 (Sun: R2)	circumflex / oxia
Alt-F6	caron / psili
Control-Shift-F6	macron / descender
Alt-Shift-F6	dot above / dasia
Control-1	acute
Control-2	grave
Control-3	hook above
Control-4	tilde
Control-5	dot below
Control-6	circumflex
Control-7	breve
Control-8	horn
Control-9	stroke
Control-0	ring / cedilla
Alt-1	circumflex and acute
Alt-2	circumflex and grave
Alt-3	circumflex and hook above
Alt-4	circumflex and tilde
Alt-5	circumflex and dot below
Control-Alt-1	breve/horn and acute (composes following A/a with breve and acute, or following O/o or U/u with horn and acute)
Control-Alt-2	breve/horn and grave
Control-Alt-3	breve/horn and hook above
Control-Alt-4	breve/horn and tilde
Control-Alt-5	breve/horn and dot below
Alt-6	psili and oxia
Control-Alt-6	dasia and oxia
Alt-7	psili and varia
Control-Alt-7	dasia and varia
Alt-8	psili and perispomeni
Control-Alt-8	dasia and perispomeni
Control-' (Ctrl-apostrophe)	acute (d'aigu) / tonos
Control-´ (Ctrl-acute)	acute (d'aigu) / oxia
Control-` (Ctrl-grave)	grave / varia
Control-^	circumflex / oxia
Control-~	tilde / perispomeni / titlo
Control-:	diaeresis (umlaut) / dialytika
Control-"	diaeresis (umlaut) / dialytika
Control-,	cedilla / ring / iota (ypogegrammeni)
Control-/	stroke
Control-– (Ctrl-minus)	macron / descender
Control-<	caron / psili
Control-.	dot above / dasia
Control-(	breve / vrachy
Control-;	ogonek / prosgegrammeni / tail / tick / upturn
Control-)	inverted breve
Control-&	hook / middle hook

Note: If your keyboard assignment provides its own accent prefix keys ("dead keys"), pressing the key twice usually delivers the corresponding spacing character which can then be used for the extended accent prefix functionality of mined; e.g. hold Control, then press ´ (acute key) twice, to invoke the acute/oxia prefix function of mined.

Note: For combining multiple accents, in most cases their order does not matter. As an exception, to combine dot above and macron, enter prefix keys in this order, as macron and dot above will be interpreted as dot below.

Note: Some accent prefix keys, when applied twice in sequence, are mapped to a single accent as follows:

acute acute e.g. F6 F6	double acute accent
grave grave e.g. Shift-F6 Shift-F6	double grave accent
macron macron e.g. Control-– Control-–	bar/topbar
cedilla cedilla e.g. Control-, Control-,	psili/comma below

Special character input shortcuts

Unicode combining characters can be entered by applying accent prefix keys to the Tab key. They will be visually combined with the previous character by rules of Unicode (and by terminal implementation). Examples:

Control-, Tab	combining cedilla
F6 F6 Tab	combining double acute accent

Typographic quotation marks can be entered by applying accent prefix keys to the space key as follows, or using certain input mnemonics (see below):

(twice) grave space	(double) left quotation mark
(twice) acute space	(double) right quotation mark
acute space e.g. F6 space	also serves for input of typographic apostrophe (or HOP ')
(twice) cedilla space	(double) low-9 quotation mark
(twice) dot above space	(double) high-reversed-9 quotation mark
^V < < or ^V > >	double angle quotation marks « »
^V < space or ^V > space	single angle quotation marks

Some characters are specifically mapped to special key combinations or specific applications of accent prefix keys for convenience or for Windows compatibility:

Control-Shift-space	no-break space (U+00A0)
Control-@ a/A	å/Å
Control-& a/A	æ/Æ
Control-& o/O	oe/OE ligature
Control-& s	ß
Control-?	¿
Control-!	¡

As with modified keys in general, these shortcuts may depend on proper terminal configuration according to the sample files in the Mined runtime support library.

Character input mnemonics

The enter-control-code prefix (^V, or ^Q in emacs mode, or ^P in WordStar mode) can be used for mnemonic character composition. This covers accented characters and other mnemonics. The available mnemonics include RFC1345 mnemonics (extended to provide generic accent mnemonics for Unicode characters), mnemonics known from HTML and TeX and useful supplementary mnemonics. See Character Mnemos reference on the mined web site for a listing.
With mined 2000.10, supplementary character mnemonics have been revised and made consistent with generic RFC1345 mnemonics, redundant mnemonics have been removed, and coverage of all Latin characters (esp. with multiple accents) has been completed.

For accent compositions, mnemonic patterns (generic accent mnemonics) are listed in the following table; the respective letter to place the accent(s) on is indicated with an "x" below.

For Greek and Cyrillic accented characters, mnemonics combining accents with Greek or Cyrillic base characters are generated automatically from the UnicodeData.txt database. They are intended for use with the Greek and Cyrillic accent prefix keys and are therefore not explicitly listed as generic mnemonics here.

generic mnemonic	accent placed on the base character ("x")
x: or "x	diaeresis (umlaut)
x' or ´x	acute (accent d'aigu)
x! or `x	grave
x> or ^x	circumflex
x? or ~x	tilde
x0 or °x	ring above
x,	cedilla
x-	macron
x(	breve
x.	dot above / middle dot
x_ or _x	line below
x/	stroke
x" or x''	double acute
x;	ogonek
x<	caron
x2	hook above
x9	horn
x-> or >x	circumflex below
x-. or .x	dot below
x--. or .x-	dot below and macron
x.-. or .x.	dot below and dot above
x7 or x.-	dot above and macron
x~- or x?-	tilde and macron
x;-	ogonek and macron
x:-	diaeresis and macron
x-:	macron and diaeresis
x-'	macron and acute
x-!	macron and grave
-x or x--	topbar
--x or x--	bar
,x or x-,	comma below / left hook
x# or x!!	double grave
x)	inverted breve
x&	hook
%x	retroflex hook
x,,	palatal hook
x~~	middle tilde
x}	curl
x-? or ?x	tilde below
x--: or :x	diaeresis below
x-0 or ox	ring below
x-( or (x	breve below
x(-. or .x(	breve and dot below
x>-. or .x>	circumflex and dot below
x9-. or .x9	horn and dot below
x'.	acute and dot above
x('	breve and acute
x(!	breve and grave
x(2	breve and hook above
x(?	breve and tilde
x<.	caron and dot above
x,'	cedilla and acute
x,(	cedilla and breve
x>'	circumflex and acute
x>!	circumflex and grave
x>2	circumflex and hook above
x>?	circumflex and tilde
x:'	diaeresis and acute
x:<	diaeresis and caron
x:!	diaeresis and grave
x9'	horn and acute
x9!	horn and grave
x92	horn and hook above
x9?	horn and tilde
x0'	ring above and acute
x/'	stroke and acute
x?'	tilde and acute
x?:	tilde and diaeresis

See also the description of the ^V function below for more input options.
Two-letter mnemonics can also be entered in reverse order if this is unambiguous. Detection of reverse order mnemomics (two letters or one letter and multiple accents) as well as the generic accent mnemonics " ^ ` ~ ¨ ¯ ´ ¸ ° (which are available for convenience in addition to the less intuitive > ! etc) works with both short mnemonic entry (two-letter "^Vxy") and full mnemonic entry ("^V xy... ").

Mnemonic character substitution commands (ESC _ and national variants) replace the two characters (or the HTML character tag) at the cursor position with a suitable composite character (e.g. accented character) if possible.

Keyboard Mapping and Input Methods

Mined supports optional keyboard mapping which is especially useful for Unicode or CJK editing. When a keyboard mapping is selected, input characters or sequences are transformed to other characters or sequences, typically of a certain Unicode script range.
Keyboard mappings for Greek, Cyrillic, Hebrew, Arabic, and major CJK input methods are preconfigured (they have been ordered in the Input Method menu according to the order of their respective basic ranges in the Unicode character set, or to the order of the letters of the usual abbreviation CJKV for East Asian text processing – Chinese, Japanese, Korean, Vietnamese). The Radical/Stroke input method provides additional functionality as a special case.
Mined provides compile-time configuration of additional input methods; for this aim, further mappings can be generated using the mkkbmap script (from tables in various formats as used by other editors or supplied by the m17n multilingualization package) and then compiled into mined. See Mined configuration below for details.

Keyboard mapping works as follows: You enter a key sequence that is mapped to a character sequence in the selected keyboard mapping table. The transformed character sequence is used as input.
As some typical keyboard mappings contain ambigous key sequences where one may be a prefix of another, a short delay is applied in these cases to allow recognition of any such sequence to be mapped. After a timeout, the shorter sequence already matching will be used; the timeout can be cut short by typing a Space key, the Space character itself will then be discarded. (The timeout value is 900 ms by default and can be configured with the environment variable MAPDELAY.)

Pick lists

Some keyboard mappings, especially for CJK input methods, contain multiple choice mappings. In these cases, a selection menu is displayed that offers a "pick list" to select a character from. A character can be picked with a mouse click, or by navigation to the desired choice with the cursor keys (down/up, right/left, page down/up) or the '<'/'>' keys , or by just selecting the menu row first (cursor-up/down), then typing a digit 1-9 or 0 to select the numbered character.
The Space key can be configured to either navigate to the next choice, the next row, or to select the current choice; see option -K.
If the pick list is too large to fit on the screen, the menu will be scrollable or pageable (using cursor keys).

While navigating through the pick list, the line and the selected item in the line are highlighted accordingly; if the current item is a CJK character, also its character information (description and optionally pronunciations as configured with the Han info option of the '?' information flag menu) is displayed on the status line. If the item is a word comprising multiple CJK characters, the information for only the first of them is shown. The available information is derived from the Unihan database.

Keyboard mapping data are based on Unicode. So in CJK text mode, the selection menu (the pick list) may contain symbols that are not mapped to the active CJK text encoding. In a UTF-8 terminal, these will still be displayed but cannot be inserted. In a CJK terminal, these are not displayed; an empty entry is shown instead.

Input method selection

An active and a standby input method (keyboard mapping) are maintained. They can be toggled quickly for text input, also on the prompt line.
The current mapping is indicated as the Input Method flag by its two-letter script tag in the flags area, showing "--" if no mapping is active.

The active mapping can be selected in the following ways:

ESC k or Alt-k or Alt-F12 or left click on Input Method flag

toggles between current (active) and previously selected (standby) input method (keyboard mapping)
(Alt- toggle functions also work on prompt line)

HOP ESC k (or HOP Alt-k)

clears input method, i.e. resets keyboard mapping to none (unmapped input)

ESC I or Alt-I or ESC K or Alt-K or Control-F12

opens the Input Method (Keyboard Mapping) selection menu
(Alt-I or Alt-K or Control-F12 also work on prompt line)

right click on Input Method flag

opens the Input Method selection menu

HOP ESC K or HOP Alt-K

cycles through available input methods / keyboard mappings

Note: For preselecting the active or standby input method by environment configuration, see about usage of the environment variable MINEDKEYMAP below.

Note: Keyboard mapping is implicitly suppressed temporarily where it is not useful: during mnemonic character input, HTML marker input, command letter entry, help selection, yes/no prompting.

Character encoding support

A character encoding for interpretation and handling of text is selected in one of the following ways:

• One of the command line options -E... with a number of options to specify the desired text encoding (see the encoding options above).
• From the Encoding Menu (one of the flag menus), the encoding interpretation can be changed while editing; to open it, click with the right mouse button on the encoding indication in the flags area of the top line, or type Alt-E. See also Mode indication flags for an overview. To toggle between the current and the previously selected encoding, click the Encoding flag with the left mouse button.
• Auto-detection (by heuristic counting of valid character codes).
  Note: The encodings to be taken into account for auto-detection can be configured with the MINEDDETECT environment variable. Set it to the desired list of encoding indications (capital letters as listed for the -E parameter, especially for CJK encodings) to disable auto-detection of other encodings. UTF-8 auto-detection cannot be disabled this way.

  Locale indication in environment variables (see Locale configuration), especially the variable TEXTLANG which does not affect the locale-related assumption of terminal encoding.

The following encodings are auto-detected unless overridden with a -E command line option (or -l or -u):

• UTF-8
• UTF-16 encoding (big or little endian) with or without BOM (byte order marker)
• 8 bit encoding is auto-detected in a generic way; the actual 8 bit encoding assumed corresponds to the terminal encoding if it is an 8 bit terminal; otherwise, Latin-1 is assumed
• CJK encoding (with unspecified mapping) is pre-auto-detected in a generic way; usually the actual CJK encoding is determined, too
• GB18030
• Big5
• EUC-JP
• Shift-JIS
• UHC
• VISCII

CJK and mapped 8 bit encoding support

Mined supports major CJK encodings as well as mapped 8 bit encodings ("character sets"). With 2000.12, mined has built in support for a large number of 8 bit encodings which appear to be in use or unique for a region. The Encoding menu has been structured with sub-menus to provide a concise menu selection feature.

Combining characters

In all character encodings handled by mined that contain combining characters, mined handles them and provides partial editing and an optional separated display mode as described above in section Combining characters. (CJK encodings EUC-JP, Shift-JIS and GB18030, Vietnamese TCVN and Thai TIS-620, ISO Arabic, Mac Arabic, ISO Hebrew, Windows Hebrew). Handling of combining text characters is properly coordinated with the set of combining characters supported by the terminal.

For Japanese, the JIS characters that map to two Unicode characters are supported.

Character code related commands

The command ESC u displays character encoding information in the bottom status line (conforming to ISO 14755); this includes the character code, the mapped Unicode character value, script and character category, and combining information. See Character information display for details.
With HOP ESC u, permanent display is toggled.

Other commands insert the code of the current character or insert a character taking its character code or Unicode value from the text. For details, see the command summary.

Terminal environment for CJK encoding support

Mined supports handling of CJK text encoding in any terminal (see Terminal encoding support below). However, proper display of a wide range of CJK characters can obviously only work in either a Unicode terminal (recommended) or in a native CJK terminal that runs the same encoding as the selected text encoding.

CJK terminals:

For terminals that support native CJK encodings (e.g. cxterm, kterm, hanterm), the terminal encoding assumed by mined can be specified with a command line option or by proper locale indication in one of the environment variables LC_ALL, LC_CTYPE or LANG. For available encodings, see Mode indication flags. For usage of the +E options, see the description of the Terminal encoding options above. For usage of the locale environment variables, see Locale configuration.

Note: In native CJK terminals, it is often troublesome to find a working encoding configuration and font setup, and the locale environment is not automatically set by the terminals. A collection of wrapper scripts is available ( http://towo.net/mined/terminals.tar.gz) to help with this setup problem and demonstrate the invocation of a number of different CJK and 8 bit encoded terminal windows, along with selection of suitable fonts and proper locale environment setting.

Note: Native CJK terminals have a different assumption of the range of character codes supported in an encoding family, e.g. Big5 / Big5 with HKSCS, GB2312 / GBK / GB18030, EUC-KR / UHC, EUC-JP without/with 3 byte codes. For compact handling, mined always assumes the largest superset of these encoding families. It does, however, have some features to prevent display garbage in most cases when a terminal supports a smaller character set:

By default, mined does not display the following CJK character codes in a native CJK terminal, i.e. it displays a substitute indication for them (see CJK character display above):

• Unknown characters: CJK characters that have no defined mapping to a valid Unicode character. Use the +C option to override this display suppression and enforce transparent display of unknown characters in a CJK terminal.
• Invalid characters: CJK characters that do not match the encoding scheme (e.g. wrt. to specified byte ranges) of the selected encoding. Use the +CC option to override this display suppression and enforce transparent display of invalid character codes in a CJK terminal.
• Extended characters: CJK characters encoded with 3 or 4 bytes. Use the +CCC option to override this display suppression and enforce transparent display of extended character codes in a CJK terminal.

Regardless of all these features and options, it may not always be possible to prevent display garbage, especially if the font used by the terminal does not cover the needed character range. To avoid these problems in general, it is recommended to use a Unicode terminal for editing CJK encoded files.

See also Terminal interworking problems for special hints about certain terminals.

Unicode support

Introduction: handling Unicode encodings

Mined interprets UTF-8 which is a multi-byte character encoding of the ISO-10646 character set, part of which is also known as Unicode. When reading a file, it detects UTF-8 encoding automatically (unless overridden by explicitly selecting a text encoding with a command line option -u or -l or -E...). It can also edit UTF-16 encoded Unicode files (UTF-16 can represent the complete 21 bit Unicode subset of ISO-10646). UTF-16 big or little endian with or without BOM (byte order mark U+FEFF) is auto-detected or can be selected with a command-line option (see notes under Locale configuration below).
Since mined 2000.10, UTF-16 is maintained transparently, i.e. a UTF-16 encoded file is written back in UTF-16, and if it was beginning with a BOM this is maintained. No explicit UTF-16 entry exists, however, in the Encoding menu since the text is internally handled in UTF-8. However, the character encoding flag indicates UTF-16 file encoding with either "16" (big endian) or "61" (little endian).

UTF-8 internal representation, transparent handling of other text

Mined handles UTF-8 representation internally and also edits and keeps illegal UTF-8 sequences. This way, if you happen to open a Latin-1 or CJK or any other encoded file in UTF-8 mode, or switch encoding while editing, or edit a file with mixed encoding, the text contents can still be edited and you will not loose any file contents information.

Character encoding indication

The upper-right flags area has a character encoding indication which shows "U8" if UTF-8 text interpretation is selected. For Latin-1 text interpretation "L1" is shown, for others see Mode indication flags. You may click on the indication flag to toggle between the current and the previous selected encoding.

Character information display

The Character information display command ESC u is described above. In UTF-8 mode, information shown includes the UTF-8 encoding byte sequence.

Character input support

With ^V, mined's special character input support is invoked (both while editing text and entering text on the prompt line, e.g. as a search expression). With this feature, (in addition to plain control characters) a composite character can be entered by its accent combination or other mnemonic character description; a more-than-two letter character mnemonics would be embedded in space characters after the ^V. In addition, numeric character codes or values can be entered with leading ^V#, octal/decimal with ^V##/^V#=, Unicode with optional u/U/+. (For examples, see description of the ^V function below.) With numeric character input, mined supports successive multiple character entry according to ISO 14755; if the numeric code is terminated by a Space key, another numeric character can be entered subsequently; an Enter key terminates numeric character input.

See also the generic section Character input support above for input support for accented characters and keyboard mapping.

Encoding conversion support

Two functions support interactive character encoding conversion (Latin-1 / UTF-8) to partially fix files with mixed encoding. In either UTF-8 or Latin-1 text mode, search for characters encoded in the other encoding with the command HOP ESC ( or Alt-F11 . Then, convert the character with ESC _ or its national variant (see mnemonic character substitution commands in the Command reference).
For repeated interactive conversion, both functions can be combined into Alt-Shift-F11 (convert current character, then search next).

Unicode Copy/Paste buffer conversion

For the Copy/Paste buffer, Unicode mode can be selected which maintains its contents always in Unicode, so that Copy/Paste of text works between differently encoded files (or sections of a file, if encoding is switched while editing) with automatic character code conversion. This mode is only effective while editing with non-Unicode encoded text interpretation.
Select this mode with the command line option -Eu or in the Paste buffer menu (righ-click on the Buffer mode flag "=" or "+") and select "Unicode".
Unicode buffer mode is indicated by cyan background of the Paste buffer flag (then "=" or "+"), except in Unicode text mode.

Smart quotes and dashes

If smart quotes mode is enabled (see the Quotes style menu under the Quotes flag left to the Encoding flag and menu), quote mark keys will enter typographic smart quotes instead. Smart dashes also apply. See Smart quotes above for more details.

Bidirectional terminal support

A bidirectional terminal (such as mlterm) will probably also apply Arabic LAM/ALEF ligature joining. Mined auto-detects this feature and enables bidi terminal handling automatically. Otherwise, bidi terminal handling can be configured with the option +UU.
In this mode, when displaying a menu, underlying text lines that contain right-to-left characters are cleared first in order to prevent display confusion between the terminal's bidi algorithm and the menu position.
Also, with bidi terminal handling enabled, mined assumes that the terminal applies Arabic LAM/ALEF ligature joining and properly accounts for this feature in display position handling.
In separated display mode, the joining part of the ligature is indicated similar to the handling of combining characters.

Input support for right-to-left scripts ("poor man's bidi" mode)

This support feature for input of right-to-left text pieces is enabled by default unless the terminal is detected to be in bidi mode itself (e.g. mlterm). "Poor man's bidi" mode is intended for quick entry of right-to-left text without having a right-to-left terminal; it is similar to the "revins" (reverse insert) option of vim and works as follows:
After entering a right-to-left Unicode character, the cursor position is moved left of it, so subsequent characters will be appended left and the text shifted right. Characters are stored in visual order while input support is implicit, based on the characters being typed. Entering a left-to-right character will automatically skip behind the previously entered right-to-left text on the line (changed in mined 2000.10) and switch to left-to-right direction; this behaviour optimises inserting small pieces of right-to-left text into basically left-to-right text; this priority is justified by the assumption that this mode (with visual storing order) is only useful for inserting small right-to-left quotations into left-to-right text and not for editing right-to-left documents (which should be stored in logical order).
Newline, Space, Tab, and combining characters attempt to behave well according to what was entered before; however, intermediate cursor movement is not considered.

Unicode line ends

Mined detects and handles Unicode line separators and paragraph separators; they are displayed as shown above (unless disabled with +u-u).
HOP Enter will insert a Unicode paragraph separator, Enter in a line that already has a Unicode line end will insert a Unicode line separator. Also the keys Shift-Enter or Control-Enter insert a paragraph separator or line separator respectively.
Configuration: In order to enable shift and control with the Enter keys, xterm must be configured as shown in the example configuration file Xdefaults.mined in the Mined runtime support library.

Unicode display

In UTF-8 terminal mode, mined displays all Unicode characters if they are contained in the font used by the terminal. Fonts usually have a substitute glyph to indicate characters not contained in the font. Wide characters (double-width glyphs) are displayed in a double-width character cell of the terminal. Combining characters are displayed either combined or separated (see Combining characters below).

Illegal UTF-8 sequences are displayed with highlighted background, using the following indications. Furthermore, control characters encoded as a UTF-8 sequence and control characters in the "C1" range (values 0x80..0x9F) will be displayed similar to normal control characters but with coloured highlighting.

8	for an unexpected UTF-8 continuation byte (range 80-BF)
4	for a 0xFE (254) byte
5	for a 0xFF (255) byte
«	for a too short UTF-8 sequence if followed by a single-byte character (00..7F)
»	for a too short UTF-8 sequence if followed by a multi-byte character (C0..FF)

Legal Unicode characters that cannot be displayed are indicated with the following replacements:

e	the character code U+FFFE
f	the character code U+FFFF
¤ or ¤  (if wide) or % (if the terminal cannot display ¤)	a Unicode character that cannot be displayed in the (non-UTF-8) terminal (non-combining)
' or '	a Unicode combining character that cannot be displayed in the (non-UTF-8) terminal
E	the Euro character U+20AC (in a non-UTF-8 terminal)
"	a quotation mark character (typographic quote mark)
-	a dash character
0 ..9 , A ..Z  etc	a corresponding fullwidth ASCII character (in a non-UTF-8 terminal)

Configuration: Display colour of special or illegal UTF-8 indications can be changed with the environment variable MINEDUNI, the value should be the numeric part of an ANSI terminal control sequence; optionally, the value can be preceded by a character to be used for Unicode character indication in non-UTF-8 (Latin-1) terminal mode.
(The default configuration value is "¤ 46").

Combining and joining characters

Mined supports handling of combining characters, featuring optional separate display and partial editing, as described above in section Combining characters.

Joining characters

If mined assumes that the terminal applies LAM/ALEF ligature joining (either configured with the +UU right-to-left display option or auto-detected), the joined character width will be handled correctly in cooperation with the terminal.
Mined supports ligature joining in both combining character display modes:

• In combined display mode, the screen position is accounted properly.
• In separated display mode, the joining part of the ligature is indicated using the appropriate isolated form on Unicode special indication background colour (similar to the handling of combining characters).

Search expression limitations

Unicode search ranges can not be very large as all included characters are listed in an internal buffer which is limited to ca. 1 KB.

UTF-8 preservation and byte-transparent editing

When splitting lines that are too long for internal handling, consistency of UTF-8 sequences is preserved (they are not split); combining characters may get split off their base characters, however, they will join seemlessly as lines are joined again (e.g. when saving the file). Note that combining characters at the beginning of a line are not displayed in combined display mode.

Terminal environment

Unicode text can be edited in any terminal encoding (UTF-8, 8 bit, CJK), however, a UTF-8 terminal is preferable.

UTF-8 terminal operation can be configured in either of these ways:

• Auto-detection: If the terminal emits cursor position reports, mined can uniquely recognise UTF-8 terminal encoding and further UTF-8 features (see Terminal encoding support below).
• Environment: By proper environment variable settings. For more details, see Locale configuration.
  Note: In general, it is advisable to start a terminal window using a wrapper script that sets a suitable locale environment at the same time, in order to support all kinds of applications that are more dependent on proper environment setting than mined is. The mined installation also provides the script uterm for this purpose, with its own manual page. (In case uterm is not installed, it is also included in the Mined runtime support library.)
• Parameter: +EU selects UTF-8 terminal mode.

See also Terminal interworking problems for special hints about certain terminals.

CJK support (Chinese/Japanese/Korean Han character features)

Mined provides CJK support features uniformly in Unicode and in major CJK encodings. For information relating to CJK character encoding see Character encoding support below.

CJK input method support

Input methods for CJK characters are supported with the keyboard mapping mechanism. A number of popular input methods for CJK text input are pre-configured, others can be added at compile-time with the mkkbmap script.

Radical/Stroke input method

Mined provides a Radical/Stroke input method for CJK characters with specific functionality in addition to keyboard mapping; it works at two-levels, selecting a radical first, then a character from a list sorted by stroke count. If this input method is active, a selection menu for the 214 CJK radicals is displayed (without prior keyboard input). The menu displays all variations of each radical. After selecting a radical from this menu, a second-level menu is displayed, showing all CJK characters based on the selected radical, sorted by the number of strokes. Many of these menus will not fit on the screen and can be scrolled. Pressing Escape here would return to the radical menu; pressing Escape there would disable the input method. To enter a non-mapped character (e.g. a line end), you need to disable Radical/Stroke input method temporarily; just toggle it back on with Alt-k (or Esc k) or Alt-F12 and the radical menu will be displayed again for continued input.
For the Unicode version used as the character data source, see the mined change log.

CJK character display

Combining characters (in both JIS encodings and GB18030) are handled and the combined characters are displayed properly in either combined or separated display mode in a UTF-8 terminal (like for UTF-8 encoded text).

The following special CJK character indications apply:

¤	CJK character that cannot be displayed in (8-bit) terminal
% or %	CJK character that cannot be displayed in (8 bit or CJK) terminal because the terminal encoding does not support it
' or '	CJK combining character that cannot be displayed in (8-bit) terminal
? or ?	CJK character code that has no known mapping to Unicode (to enforce display on CJK terminal use option +C)
# or #	invalid CJK character code that is outside of the code range assigned to the encoding scheme (to enforce display on CJK terminal use option +CC)
#	CJK character in extended code range (esp. 3 and 4 byte codes, or codes with 0x80...0x9F byte range) that cannot be displayed on CJK terminal due to terminal capability limitations (to enforce display on CJK terminal use option +CCC)
<	incomplete or otherwise illegal CJK code

Han character information display

When the cursor is on a Han character and either descriptive or pronunciation information about this character is available in the Unihan database (from unicode.org), mined can optionally display this information, with a selection of display details which may include semantic information and various pronunciations.
To enable Han info, select it in the Info menu. To open the Info menu, type Alt-F10 or right-click the "?" flag.
The information can optionally be shown on the status line (where it may be truncated if too long) or in a pop-up menu next to the character. Pronunciation information to be displayed can be selected in the Info menu.

The same information is always shown while you are browsing an input method pick list (then on the status line).

The information includes the character code (in CJK encoding, both CJK code and corresponding Unicode value are shown). The amount of descriptive information (from the Unihan database) to be shown can also be preconfigured with the environment variable MINEDHANINFO; see Han info configuration below.
(For the Unicode version used for the Unihan data source, see the mined change log.)

Terminal encoding support

Mined supports all kinds of UTF-8, CJK, Latin-1, and (with mined 2000.12) other 8-bit encoded terminals.

Terminal feature detection

Mined performs auto-detection of a number of terminal features:

• For UTF-8 terminals, mined performs auto-detection of terminal features (detection of UTF-8 terminal, different width data and combining data versions, handling of double-width, combining and joining characters).
• For CJK terminals, mined performs some auto-detection of specific CJK terminal features (handling of non-EUC code points, handling of extended code range, GB18030, 3-byte and 4-byte encodings, detection of kterm JIS encoding, detection of rxvt emulating CJK encoded terminal, special CJK width properties, and terminal support of combining characters).
• For mapped 8-bit terminals, mined performs auto-detection of terminal support of combining characters.
• For the Unicode version used for width and combining character properties, see the mined change log.
• CJK terminals cannot always be distinguished from 8-bit terminals by auto-detection. Neither can the encoding of either CJK or 8-bit terminals be auto-detected. It is thus advisable to setup proper settings of locale environment variables (LC_ALL, LC_CTYPE, LANG). Alternatively, the effective terminal encoding can be indicated to mined with a command line option (+EX). For configuration details, see Locale configuration below.

Specific terminal properties

For more specific configuration hints (especially for PC-based terminals), and for description of the handling of certain terminal interworking problems, see the Terminal environment configuration hints below.

Environment interworking and configuration hints

A number of configuration options have already been addressed throughout the manual page. A few more configuration features are mentioned here. For more details, examples, and other display settings see the example script profile.mined in the Mined runtime support library.

Mined runtime support library

The mined distribution provides a collection of runtime support files (in subdirectory usrshare); if mined is installed into standard locations, they are copied to one of the directories /usr/share/mined, /usr/share/lib/mined, /usr/local/share/mined, /opt/mined/share, $HOME/opt/mined/share (depending on operating system and installation options).

Mined runtime support includes the following files:

Package documentation

package_doc/README

mined package overview and introduction

package_doc/VERSION

version of the installed mined release

package_doc/CHANGES

mined change log

package_doc/LICENSE.GNU

the GNU license applicable to mined

Web documentation

doc_user/*

copy of the web documentation including the HTML version of the mined manual page

Online help

help/mined.hlp

online help file

Example files: environment configuration patterns

conf_user/profile.mined

shell commands to set environment variables for mined, template for inclusion in $HOME/.profile

conf_user/Xdefaults.mined

xterm configuration entries suitable for mined, template for inclusion in $HOME/.Xdefaults or $HOME/.Xresources

conf_user/xinitrc.mined

shell commands to activate Xdefaults.mined, template for inclusion in $HOME/.profile

conf_user/mlterm/main

mlterm configuration to enable Alt-key detection, for inclusion in $HOME/.mlterm/main

conf_user/mlterm/key

mlterm configuration for modified (shifted etc) function keys, for inclusion in $HOME/.mlterm/key

Scripts to be used at runtime

bin/uprint

script for printing a Unicode file, using either paps or uniprint for formatting

bin/minedmar

script to clean up the @mined.mar file position file

bin/minedmar.bat

DOS/Windows version of minedmar

Scripts to start mined

bin/uterm

script to invoke xterm in UTF-8 mode; it should also be installed into the system binary path and has its own manual page

bin/mterm

script to invoke mlterm with suitable options (for bidi support)

bin/umined

script to start mined in a separate xterm window, using UTF-8 mode with most recent version of Unicode width data (specifying wide and combining characters) as built-in to xterm

bin/xmined

script to start mined in a separate xterm window, using same encoding mode as currently set

bin/wmined

script to start mined in a separate rxvt window, using Windows look-and-feel, especially on Windows where rxvt is run in stand-alone mode

bin/wmined.bat

DOS/Windows version of wmined

Files to setup a mined installation

setup_install/mined.desktop

KDE desktop entry to start mined in an xterm from a menu entry, using the uterm script

setup_install/mined.ico

Cygwin/X desktop icon for adding mined to the Cygwin-X Editors section in the Windows Start menu

Scripts to configure an environment for mined

setup_install/bin/configure-xterm

sample configuration script to build xterm with recommended configuration options

setup_install/bin/makeprint

script to search for or retrieve and build the uniprint program from the yudit package

setup_install/bin/installfonts

script for downloading the Unicode-enhanced X screen fonts and installing them with your X server

setup_install/bin/bdf18to20

script to transform an 18x18 pixel double-width screen font into a corresponding 20x20 pixel font matching the 10x20 single-width font (which is much nicer than the 9x18)

setup_install/bin/mkicon

script to install mined with Cygwin/X by creating an entry (with icon) in the Cygwin-X Editors section in the Windows Start menu

setup_install/bin/postinstall

script to invoke mkicon after installation on Cygwin

Terminal environment

The Unix terminal type is determined from the environment variable TERM.

Recognition of some special terminal features or restrictions is associated with the setting of TERM (xterm, linux, vt100, sun*, cygwin, rxvt, *ansi*, 9780*, hp*, xterm-hp, superbee*, sb*, microb*, scoansi*, xterm-sco, cons*, att605-pc, ti_ansi, mgterm).

For detection of function keys and cursor keys, the escape sequences being used by terminals are often not known to an operating system environment because they are poorly and incompletely configured. Because this does usually not work as expected (see this bug report just for an example), mined does not rely on the termcap/terminfo configuration of function key codes; rather it always accepts a wide variety of typical codes. A few ambiguous codes are resolved according to the TERM variable.

In an xterm, window headline and icon text are set to the current filename and "(*)" is added if the text has been modified.

Locale configuration

The locale mechanism as implemented on modern systems has a number of design problems, one being that there is no explicit distinction between text encoding and terminal encoding although this is obviously a very different thing and mixed combinations of both may occur and are actually supported by mined.
For this reason, mined extends the locale environment variable mechanism with the variable TEXTLANG which is only considered for assumed text encoding (with precedence over the other locale variables). Also mined provides additional features to specify both terminal and text encodings.

• For text encoding, mined checks the variables TEXTLANG, LC_ALL, LC_CTYPE, LANG in this order.
• For terminal encoding, mined checks the variables LC_ALL, LC_CTYPE, LANG in this order.
• Explicit command line parameters are available to specify either terminal encoding (+E) or text encoding (-E). They override environment variable settings.
• UTF-8 terminal auto-detection overrides other terminal encoding settings.
• Text encoding auto-detection overrides environment settings but not command line settings.
• Assumed text encoding can be switched while editing.

For encoding recognition from locale environment variables, mined recognises locale specifications typically found in system installations, including those which do not include an explicit encoding suffix. Known encoding suffixes ("charset" or "codeset" component of locale name, starting with ".") are recognised regardless of whether the given locale is installed or not. Other encodings are recognised by region suffix (starting with "_") or full locale name or alias.
In addition to hard-coded locale recognition (especially for CJK), locale values and associated encodings are configured in the compile-time configuration file locales.cfg which especially lists locale names that do not have an explicit encoding suffix. You can use these settings (known locale name or generic locale name suffix) even on legacy systems without locale support to indicate the terminal environment properly to mined.

For encoding recognition from command-line parameters, mined provides three options:

• -EX or +EX with a single-letter encoding tag as listed with the description of the -E options; further encoding tags are configured in the compile-time configuration file charmaps.cfg.
• -E=charmap or +E=charmap with a character encoding name ("charmap" as reported by the locale charmap command).
• -E.suffix or +E.suffix with a locale encoding suffix ("charset"/"codeset" of locale name).
• -E:flag or +E:flag with a 2-letter indication used by mined to indicate the respective text encoding in the Encoding flag.
In each of these options, -E specifies text encoding while +E would specify terminal encoding to be assumed.

The following table lists major encodings and generic locale suffix values by which they are recognised; in addition (as mentioned above), a large number of locale names without encoding suffix as found on various systems in known to mined and will cause it to assume the corresponding terminal encoding.

Unicode: UTF-8	suffixes: .UTF-8 / .utf8
Traditional Chinese (Hongkong): Big5 with HKSCS	suffixes: .BIG5* / .Big5* / .big5* / _HK / _TW (_TW ambiguous, following encoding overrides)
Simplified Chinese: GB18030 (includes GBK and GB2312)	suffixes: .GB* / .gb* / .EUC-CN / .euccn / _CN.EUC / _CN
Traditional Chinese (Taiwan): CNS (EUC-TW)	suffixes: .EUC-TW / .euctw / .eucTW / _TW.EUC
Japanese: JIS / EUC-JP	suffixes: .EUC-JP / .eucjp / .eucJP / .ujis / _JP.EUC / _JP / .euc (.euc ambiguous, more specific string overrides)
Japanese: Shift-JIS	suffixes: .Shift_JIS / .shiftjis / .sjis / .SJIS
Korean Unified Hangul: UHC (includes EUC-KR)	suffixes: .UHC / .EUC-KR / .euckr / .eucKR / _KR.EUC / _KR
Korean: Johab	suffixes: .JOHAB
Vietnamese: VISCII	suffixes: .viscii
Vietnamese: TCVN	suffixes: .tcvn
Thai: TIS-620	suffixes: .tis* / .TIS* / _TH / .iso8859[-]11 / .ISO8859[-]11
Latin-9: ISO 8859-15	suffixes: @euro / .iso8859[-]15 / .ISO8859[-]15
Cyrillic: ISO 8859-5	suffixes: @cyrillic (unless preceded by uz_UZ which indicates UTF-8)
Latin or other: ISO 8859 encodings	suffixes: .iso8859[-]N / .ISO8859[-]N (with number N)
Russian Cyrillic: KOI8-R	suffixes: .koi8r
Ukrainian Cyrillic: KOI8-U	suffixes: .koi8u
Tadjikistan Cyrillic: KOI8-T	suffixes: .koi8t
Russian, Ukrainian, Bjelorussian Cyrillic: KOI8-RU	suffixes: .koi
MacRoman:	suffixes: .roman
Windows Latin: CP1252	suffixes: .cp1252
Windows Cyrillic: CP1251	suffixes: .cp1251
PC Latin: CP850	suffixes: .cp850
Windows Hebrew: CP1255	suffixes: .cp1255
Georgian: Georgian-PS	suffixes: .georgianps
Kazachstan Cyrillic: PT154	suffixes: .pt154

Examples: To indicate that mined is running in a UTF-8 terminal (normally auto-detected, included here for demonstration) and should assume GB18030 text encoding by default, invoke either of:

LC_ALL=whatever.UTF-8 TEXTLANG=zh_CN.gbk mined

LC_CTYPE=whatever.UTF-8 TEXTLANG=chinese mined

LANG=whatever.UTF-8 mined -EG

LC_ALL=en_IN mined -E.gbk

mined +EU -E.EUC-CN

mined +EU -E=GB18030

mined +EU -E:GB

Selecting UTF-16: To tell mined to interpret a file (or make a new file) in UTF-16 encoding, use the following command line options (first two little endian, then big endian):

mined -E:61

mined -E=UTF-16LE

mined -E:16

mined -E=UTF-16BE

mined -E=UTF-16

Selecting ASCII: To tell mined to assume that a terminal cannot display anything but ASCII characters, use the command line option +E:AS. Mined implicitly assumes this setting if the environment variable TERM indicates a VT52 terminal.

PC terminals

Character encoding of PC terminals is an even greater mess than on Unix systems. Mined provides heuristic best-guess assumptions about terminal encoding, supporting both local invocation as well as remote login from a PC (e.g. to a Unix machine).

The following assumptions are made based on environment variables or command-line parameters:

encoding ("codepage")	environment	option	examples
CP850 (PC mapping of Latin-1 character set)	TERM=ansi, ansi-nt, pcansi*, hpansi*, interix* or TERM=cygwin and CYGWIN contains "codepage:oem"	+EP	Windows console (DOS box), Windows console mode telnet (even if called from cygwin console, sets TERM=ansi), special cygwin mode (including telnet)
CP437 (IBM PC VGA encoding)	TERM=nansi*, ansi.*, opennt*, *-emx*	+Ep	plain DOS (without remapping)
CP1252 (Windows ANSI extension of Latin-1)	TERM=cygwin (unless CYGWIN contains "codepage:oem")	+EW	cygwin console (emulation in DOS box), cygwin telnet (even if called from DOS box with different encoding – does however not change TERM if set already), cygwin mined (even if called from non-cygwin DOS box with different encoding), older Windows GUI telnet (sets TERM=ansi)

Note: It is not unlikely that the assumption about the terminal encoding taken by mined does not match the actual terminal encoding (e.g. mined cannot determine the encoding based on the ambiguous setting TERM=ansi). Also the environment variable CYGWIN is not maintained through telnet. Explicitly setting TERM to a suitable value after remote login may help but may not always work either (e.g. pcansi is not known on SunOS).
In these cases, use the explicit +E option to force mined to assume a specific terminal encoding; see the option values listed above for the main DOS encodings.

Note: As indicated above, cygwin emulates the Windows Latin ("ANSI") character encoding (extension of ISO Latin-1) in a DOS box window (unless reconfigured with the CYGWIN variable) instead of the IBM PC character set that a DOS box runs by default. This may produce unexpected changing of initial appearance when the cygwin-compiled mined version and the djgpp-compiled mined version are used one after the other. Since mined 2000.12, text encoding interpretation can be changed while editing in the PC versions as well as in the Unix version by clicking the Encoding flag. As Windows Latin emulation is limited to the characters provided by the Windows DOS box, emulation of some characters is poor (e.g. the Euro sign). Since 2000.13, mined assumes an accordingly reduced character set for cygwin, improving display by substituting more suitable replacement indications.

Note: The environment setting "CYGWIN=codepage:oem" which is properly detected and handled by mined is not conveyed with remote login.

Note: Running mined in a dosemu session (DOS emulator on Linux) works fine, even in an xterm-embedded session although not perfectly in that case: ^S and ^Q are interpreted for flow control (thus ^S will hold all output until ^Q is entered), and the mined option -Qa should be used to tune menu borders right.

Terminal setup

The Mined runtime support library includes a configuration file Xdefaults.mined which lists settings that should be applied to the terminal for proper operation of several features as described throughout this manual.

In some terminals, the cursor may not be well visible or not visible at all if the cursor is on a character with reverse background (control character, occurs e.g. in xterm) or highlighted background (invalid character code, occurs e.g. in xterm and rxvt). See the X resource parameters for "cursorColor" in the example configuration file Xdefaults.mined for remedy.

If your terminal scrolls down one line when you click the left mouse button in the text area, the terminal type is not properly set up. This occurs, e.g., when you run inside a cygwin or rxvt terminal but the environment variable TERM is incorrectly set to xterm. Set it to the correct value for remedy.

If mouse wheel movement moves more than expected, especially if it cannot move by single items in a menu, this is probably a configuration issue with your mouse driver. You are probably running a Windows-based X server which is (often by default) configured to generate multiple mouse wheel events on each actual mouse wheel movement. Often not even in the Control Panel mouse section, but only in a configuration menu of mouse-specific setup software (e.g. "Browser Mouse Settings"), configure the scroll unit to 1.

Terminal interworking problems

With some terminals, problems are known due to missing terminal features or terminal bugs:

any terminal: menu border display

If the borders of mined menus appear as letters rather than graphic borders, the terminal can unexpectedly not handle VT100 block graphics. Use the option -Qa to switch to ASCII borders.
In a UTF-8 terminal, mined uses Unicode Box Drawing characters by default. If they don't display they are missing in the font used by the terminal. Use the option -Qv to switch to VT100 block graphics or -Qa to switch to ASCII graphics. If borders are visible but without corners, use -Qs to switch to rectangular borders.

any terminal: slow terminal feature auto-detection

Occasionally, when starting mined, you may receive a message "Late screen mode response – set ESCDELAY=2000 or higher for proper detection".
This happens if there is a large delay (> 700 ms) in the interaction mined uses to detect terminal properties. There are two possible reasons for this:

• A slow remote terminal connection. In this case, set up your environment variable ESCDELAY to a value (in milliseconds) large enough to cover the anticipated delay, e.g.:
  export ESCDELAY; ESCDELAY=3000

  Font loading. Especially with rxvt and mlterm, X fonts seem to be loaded partially on demand. While this speeds up initial terminal operation, it also results in unexpected delays of terminal responses. In this case, exiting mined and starting again will normally resolve the issue for one session of the terminal. For a more permanent remedy, also use the environment variable ESCDELAY when using those terminals, e.g.:

  export ESCDELAY; ESCDELAY=1200

Automatic handling of the situation is planned for the next release of mined.

mlterm

Bidirectional display handling of mlterm is based on the final display, not regarding any context (such as positioning control, that's why mined implements a workaround for menu display on mlterm). This also affects mouse cursor position reports which do not match over right-to-left text, so the cursor will be placed somewhere else in the line.

The Mined runtime support library includes a configuration file mlterm/key which defines enhanced escape sequences for function keys and other modified keys in order to enable the functionality described in this manual.

xterm

Although it is a waste of keyboard resources to have two indistinguishable sets of keypad keys, most terminals provide no means of distinguish them towards the applications, at least not by default. Especially for a text editor, it is highly desirable to distinguish them in order to have a rich intuitive function key mapping at disposition which mined tries to achieve.
Remapping keypad keys in a useful way is sensitive because it may create incompatibilities with other programs that rely strictly on installed terminfo information. Mined provides remapping recommendations for shifted keypad keys (with Shift, Control, Alt and combinations of them) in the configuration sample file Xdefaults.mined in the Mined runtime support library.
Due to the compatibility limitations mentioned above, however, the two Ins keys remain indistinguishable, and the two Del keys are only distinguishable if the xterm configuration resource *VT100*deleteIsDEL is set. Also, keypad and function key modification with the Alt is ensured with the xterm resource *VT100*metaSendsEscape. Both resource are set to true in the same configuration sample file just mentioned. These two can also be set dynamically with xterm. Mined is told to do so with the command line option +D. (Unfortunately this cannot be done by default as it cannot be undone because the previous state cannot be detected.)

Mined interactively determines the xterm version in order to apply certain workarounds conditionally.

If you run xterm in VT220 mode (configuration resource decTerminalID) you should make sure to also set the environment variable TERM=vt220 so mined can properly set up the keypad functions.

xterm legacy CJK width mode

Mined auto-detects and supports xterm legacy CJK width compatibility mode (xterm -cjk_width); character width and menu border layout are properly adjusted, stylish menu borders (-QQ) and fine-grained scroll bar display are disabled by default.

rxvt

In rxvt, the two keypad Del keys (small keypad, numeric keypad) are automatically distinguished from each other and invoke the Delete character (small keypad) and Cut (numeric keypad) functions, respectively (Control-/Shift-/Alt- alternatives are supported as described in this manual). This works, however, only if mined can recognise rxvt; it is generally a bad idea to set TERM=xterm in rxvt, see also hint below.

Also in rxvt, the two keypad Ins keys (small keypad left, numeric keypad right) are distinguished. The left Ins key positions the cursor left of the pasted region, the right Ins key positions it right.

The recent rxvt-unicode release provides a CJK terminal emulation. CJK display is buggy for characters that rxvt thinks cannot be displayed, especially for GB18030 (LC_CTYPE=zh_CN.gb18030 rxvt); single bytes are then interpreted instead which amounts to an unpredictable screen width and cannot be correctly handled. (This applies to character codes that are not mapped to Unicode but also to many that are mapped.)
Moreover, CJK width handling is inconsistent for many characters in rxvt CJK mode (rxvt claims to adhere to the locale mechanism in this respect but that's not the case here – character widths are inconsistent with the locale, too).
Remedy: Don't use rxvt in CJK-encoded mode; mined CJK terminal support is tailored to native CJK terminals (such as cxterm, kterm, hanterm) where it works fine – if you use a UTF-8-capable terminal, use it in UTF-8 mode! Mined can edit CJK-encoded files well in a UTF-8-encoded terminal.

In rxvt, Unicode characters that are Not Assigned are always displayed as a single-width replacement character. This is not consistent with xterm behaviour which would display them as a double-width replacement if they are located within a double-width Unicode range (which sounds reasonable). This would cause display positioning inconsistencies. Mined has a workaround for some of these cases (assuming that rxvt runs the most recent Unicode width data version available; or actually the same as mined assumes – handling of multiple auto-detected terminal Unicode versions does not cover this special case).

Due to the scrollbar display workaround for hanterm (see above), the scrollbar position may be shown as blank space instead of coloured (only in rxvt CJK mode with Korean encoding and if you explicitly set TERM=xterm which you shouldn't anyway in rxvt). In this case, coloured scrollbar foreground can be enabled with the environment variable MINEDSCROLLFG="44;36" or MINEDSCROLLFG="38;5;45".

As a workaround for an xterm bug on cygwin, mined applies terminal size re-adjustment. This may confuse rxvt (being resized to an unexpectedly large window) if it pretends to be xterm.
Remedy: in rxvt, make sure that the environment variable TERM=rxvt (or rxvt-unicode); the according X resource (Rxvt.termName: rxvt) is also listed in the file Xdefaults.mined in the Mined runtime support library.

Mined interactively determines the rxvt version in order to use certain features conditionally.

cxterm

EUC-JP half-width characters (8EA1-8EDF) are not properly displayed by cxterm in EUC-JP mode (cxterm -JIS, not available in "classic" cxterm).

Due to the scrollbar display workaround for hanterm (see above), the scrollbar position may be shown as blank space instead of coloured (only in Korean encoding mode which is probably rarely used with cxterm anyway). In this case, coloured scrollbar foreground can be enabled with the environment variable MINEDSCROLLFG="44;36" or MINEDSCROLLFG="38;5;45".

kterm

Auto-detection of kterm as a CJK terminal works if the environment variable TERM indicates "kterm"; otherwise mined has to be told that it runs in a CJK terminal and which encoding to use:
For kterm -km sjis, set LC_CTYPE=ja_JP.sjis (or invoke mined +ES).
For kterm -km euc, set LC_CTYPE=ja_JP.eucjp (or invoke mined +EJ).

Note: By default (i.e., without explicit -km option or corresponding *vt100.kanjiMode resource configured), kterm runs in ISO 2022 mode (yes, it does indeed) which is not supported by mined.

hanterm

CJK display is buggy at the line beginning or after a Tab, often only the second byte of the character code is displayed as an ASCII character instead of displaying the complete CJK character.

Character attributes in hanterm used to be all mapped to reverse, so there was a workaround to enable a visible position in the scrollbar which is displayed as blank space. The criteria for this workaround to apply are: CJK terminal (detected or configured), TERM=xterm, Korean encoding (UHC or Johab) configured with parameter or locale. Replaced to enable nicer colours in scrollbar. To reactive workaround for older hanterm, set environment variable MINEDSCROLLFG="0".

konsole

Due to the lack of decent Unicode font support in the default configuration of the KDE konsole terminal, menu appearance options -QQ and -Qr should not be used; rounded borders are disabled by default.

The Mined runtime support library includes a configuration file konsole/xterm-modified.keytab which defines enhanced escape sequences for function keys and other modified keys in order to enable the functionality described in this manual. Unfortunately, the qt framework used by konsole inhibits the use of some keys and many key combinations.

It is especially obnoxious that by ignorance of qt, konsole is not able to offer the middle keypad key ("5" in application mode) to applications so the HOP function has to be invoked by alternative means in mined.
As a remedy, the HOP function is assigned to the "Menu" key (next to the "Windows" key on PC keyboards) in the configuration sample file mentioned above.

DOS console

With the djgpp-compiled version apparently there is a Control-C problem on older Windows versions. Every first Control-C will display ^C on the screen at the current position without mined noticing it, while every second Control-C will be passed to mined. This problem does not occur on Windows XP. It does occur on Windows ME in a DOS box. It does not occur with the cygwin-compiled version.

PuTTY

This Windows terminal emulation for remote login provides various keyboard (esp. keypad and function key) assignment emulations. In SCO mode, shifted function keys are different from those of xterm SCO function key emulation; both are supported. In application keypad mode, the Control-HOP key (Control-keypad-5) emits the same escape sequence as the Page Down key and cannot be used thus.

Work-around support to enable 8-bit character set on weird terminals

There exist some exceptionally weird 7 bit terminals that have an alternative character set containing composed character which can be displayed simultaneously with the default character set. For those there is optional output translation which embeds non-ASCII characters into the respective code switching sequences. To enable output character transformation, set the environment variable MINEDOUT to contain the upper half (with respect to an 8 bit character set) of the translation table into the terminal's alternate character set. (Character set switching will be done as specified in the termcap (as/ae) or terminfo (smacs/rmacs) entry.) An example setting of MINEDOUT is included in the environment sample file profile.mined in the Mined runtime support library for Siemens 9780x terminals.

Concerning some especially stupid terminal drivers

There used to be terminal drivers which make use of the soft handshake mechanism by exchange of ^S and ^Q characters but yet pass them through to application programs which is quite stupid. If it is necessary to ignore such hazardous ^S and ^Q keys, the environment variable NoCtrlSQ or NoControlSQ must be set. Mined will then not disable the tty channel soft handshake setting either.

Keyboard mapping / Input method pre-selection

With the environment variable MINEDKEYMAP the active or standby mapping or both can be preselected. The value is a two-letter script tag to set the active mapping, or it is prepended with "-" to set the standby mapping, or a combination.
Example:

export MINEDKEYMAP=-gr will set Greek keyboard mapping standby.

export MINEDKEYMAP=py-rs will set Pinyin input method active and Radical/Stroke input method standby.
The respective tags attached to the keyboard mappings can be looked up in the Input Method flag menu; the HOP function toggles between display of the full input method name and its tag.

Smart Quotes style configuration

Smart quotes style can also be preselected with the environment variable MINEDQUOTES which should then contain the opening/closing quote pair or just the opening quote mark (double or single quotes).
Example:

export MINEDQUOTES="»" sets these »inward« quotes and corresponding single smart quotes.

export MINEDQUOTES="»»" sets these »Swedish» quotes and corresponding single smart quotes.
The value of the MINEDQUOTES variable must be encoded in UTF-8.

Han info configuration

With the environment variable MINEDHANINFO, the information shown for Han characters can be preselected. If the variable is defined, Han info mode is enabled. It may contain letters to select description, pronunciation information, and display mode to be used:

M	show Mandarin pronunciation
C	show Cantonese pronunciation
J	show Japanese pronunciation
S	show Sino-Japanese pronunciation
H	show Hangul pronunciation
K	show Korean pronunciation
V	show Vietnamese pronunciation
P	show Hanyu Pinlu pronunciation
T	show Tang pronunciation
D	show character description
F	display full information (in popup-menu form); without F, the information will be shown on the status line where it is subject to truncation

Common paste buffer configuration

The paste buffers, used for cut/copy/paste operations, as well as the inter-window paste buffer, are located in a temporary directory, using system conventions by default. To maintain the inter-window paste functionality even remotely, mined uses the environement variables MINEDTMP and MINEDUSER which, in combination, point to a user-defined temporary directory and file name pattern to be used for buffer files:

• Set MINEDTMP to refer to a common mounted network directory on all machines which means that the value of $MINEDTMP may have to be different to reflect different mount points across the network.
• Set MINEDUSER to the same name within the network even if using different user name accounts.

For details, see also the FILES section below.

Keypad configuration

In order to enable the Alt key modifier for quicker entry of "ESC" commands, or the keypad "5" key as a HOP key, or the Scroll Lock or Pause key as a HOP key (for convenience on notebooks), you may have to add some X configuration. Also you may want to distinguish "Home" and "End" keys on the two keypads of PC keyboards in order to enhance keypad value (double assignment of functions to keypads is a waste of resources). And you may want to enable control and shift modifiers for keypad and function keys. See the example file Xdefaults.mined in the Mined runtime support library for suggestions.

Printing configuration

Mined uses the script uprint from the Mined runtime support library to print the current contents of the text being edited in any selected encoding (unless the environment variable MINEDPRINT is set to direct mined to use a different print command).
If the support library is not installed in one of its standard locations (system-dependent), it should be made available in the usual command search path. The script uses either paps or uniprint for actual formatting (print preprocessing). paps is available at http://paps.sourceforge.net/ and uses the Pango layout engine for formatting. uniprint is part of the yudit distribution; if you don't have it installed on your system, there is another script makeprint in the support library which can be used to download and build the needed uniprint program. The mined print script (uprint) prefers paps if it is available as it has more capabilities for printing a wide range of Unicode characters, and it does right-to-left formatting.
The font to be used with uprint can be configured with the environment variables FONT, FONTPATH, FONTSIZE. Also the printer can be configured as usual with PRINTER. In addition, uprint checks an environment variable LPR for an alternative for the system printing command (lpr/lp) if that is needed.
Note: If printing with uprint fails for some reason, mined tries to print with either the print command configured in the environment variable LPR as a fallback, or with lp/lpr as a last resort. Working character encoding support cannot be expected in this case, however.
See Environment variables to configure Printing for further details.

Display layout

Some of the special indication characters (that substitute non-displayable contents) and some of the colours used by mined for special indications and interactive elements may be configured to the user's preference.
Note: For the configurable character indications, two environment variables exist each, to configure an 8 bit value (Latin-1 encoded) and to configure a Unicode value (UTF-8 encoded). The UTF-8 encoded values (e.g. MINEDUTFRET) take precedence in a UTF-8 terminal. In an 8 bit terminal, or if the respective UTF-8 variable is not configured, the Latin-1 encoded value applies. See the example script profile.mined in the Mined runtime support library for more details and for a number of suggestions of suitable values. Mined does not apply any default non-Latin-1 indications in order to avoid display problems with fonts that do not support them. Depending on your visual preference, there are a number of suitable Unicode characters for use as indications especially in the Unicode ranges of Arrows, Geometric Shapes and Symbols (U+2190-U+2BFF).
Note: For the Latin-1 encoded configured indication markers (variables MINEDRET etc, not MINEDUTFRET etc), if the configured character is in the small letters range (actually '`'...DEL) the alternate character set is used for display. This works also in a UTF-8 terminal, provided that the corresponding UTF-8-encoded indication configuration variable is not set, e.g. MINEDRET=j MINEDUTFRET= (or not defined) would indicate line-ends by displaying a graphic lower right corner.
Note: For the UTF-8-encoded configured indication markers (variables MINEDUTFRET etc), if the marker is a double-width character, a replacement will be displayed instead.

Line ends

Line ends are usually marked by a "«" double left angle character. This visual indication can be changed with the environment variable MINEDRET (8 bit terminals) or MINEDUTFRET (UTF-8 terminals). The default or configured marker is used as an indicator at the end of every text line on screen (so you can see how many blank spaces there are).
Multi-character markers: If a second character is configured, it is used to fill the rest of the screen line, a third configured character would terminate the indication at the end of the screen line. ("··«" is a nice setting for people who used to work at Siemens terminals.)

Pattern:

		MINEDRET=123	# line end displays as 122222223

Suggestion for a nice line end on UTF-8 mode terminals (check if character is included in your font, however!):

		MINEDUTFRET=⏎	# U+23CE

The indication of DOS line ends (CR/LF) may be configured with the variable MINEDDOSRET or MINEDUTFDOSRET.

Paragraph ends

With the option -p, mined displays distinct indicators for line ends and paragraph ends. A paragraph is defined to continue while lines end with white space (space or Tab character). The default paragraph marker is "¶" and is also used to indicate a line ending with a Unicode Paragraph Separator. It can be changed with the environment variable MINEDPARA or MINEDUTFPARA.

Tab characters

Tab characters are usually indicated by a sequence of '·' (middle dot) characters. This can be changed with the environment variable MINEDTAB (8 bit terminal) or MINEDUTFTAB (UTF-8 terminals).
Multi-character markers: If two characters are configured, the second is used to mark the middle of the Tab span. If three characters are configured, the first and last are used to mark the beginning and end of the Tab span.

Pattern:

		MINEDTAB=123	# Tab displays as 12222223
		MINEDTAB=12	# Tab displays as 11112111

Long lines

Lines which are too long for the screen are usually indicated by a '»' double right angle (guillemot) character. If the current position is behind the screen margin, the line is shifted out left which is indicated by a '«' double left angle. These markers can be changed with the environment variable MINEDSHIFT or MINEDUTFSHIFT. The first character is used to indicate a line continued to the left of the screen, the second character is used to indicate a line continued to the right of the screen.

Unicode characters

For a description of special display indications in UTF-8 text editing mode see "Unicode display" above. The indication and highlighting mode of a non-displayable Unicode character (typically a UTF-8 character in a Latin-1 terminal), as well as the highlighting mode (colour) of the indication of illegal UTF-8 sequences, can be configured with the variable MINEDUNI.

Display mode of indicators

It is recommended to display these indicator characters in a dim display mode to prevent distraction from the text contents. The default is a red colour which is a moderate dark red in xterm. The display mode can be used by placing the code part of an ANSI display control sequence in the environment variable MINEDDIM. E.g., MINEDDIM=31 would select the default mode, red foreground; MINEDDIM="32;40" would display indicators in green on black.

Scrollbar colour

The foreground and background colours of the scrollbar can be configured with MINEDSCROLLFG and MINEDSCROLLBG, respectively, using ANSI modes; if only the background is configured, the foreground is the reverse of it. In general, to support fine-grained scrollbar display in UTF-8 terminals, the foreground and background colour settings should be the reverse of each other. The default for the background is "46;34;48;5;45" if use of 256 colour mode is enabled, or "46;34" if it is disabled. The default for the foreground is "", meaning that the reverse background is used, with a workaround for hanterm (see above).

Menu colour and border style

The highlighting background colour of the selected menu item can be configured with MINEDSEL, using ANSI modes. The default is "34".
Menu border styles can be selected with the option -Q. For a nice selection bar that extends from left to right menu border, the setting -QQ is recommended; it is the default setting (since mined 2000.13) except if the terminal is assumed not to provide sufficient font configuration for this option; it depends on certain graphic Unicode characters being included in the terminal font and can be disabled with -Qq.

Online Help access

Mined looks for its online help file in a number of typical directories for installation of the Mined runtime support library. If it is placed in a non-standard location, the environment variable MINEDDIR should point to the directory. (Mined also tries to find the online help file in the directory where it was started from; this is especially useful for the DOS/Windows version.)

---

Mined configuration

Script highlighting

The the mined distribution contains a file src/colours.cfg; it contains entries with the script name (as listed in the Unicode data file Scripts.txt), blank space, and a colour index into the xterm 256-colour mode. (To make good use of 256 colour mode, the terminal program should be compiled with 256 colour support enabled. Configure xterm with configure --enable-256-color .)
Edit colours.cfg before building mined to adapt coloured script display to your preferences.

Encodings and Encoding menu

The mined distribution contains a file src/charmaps.cfg which defines the character encodings that mined knows and how they are presented in the Encoding menu, together with flags for indication in the Encoding flag and tags for use with the -E and +E options (and the MINEDDETECT environment variable).
The configuration file allows the definition of sub-menus in the Encoding menu.
Each character encoding entry charmap-name must correspond to an existing character mapping file charmaps/charmap-name.map. Additional character mappings can be generated with the script mkchrmap.

Encodings recognised by locale names

The mined distribution contains a file src/locales.cfg which maps locale names to associated character encodings. While this list contains mainly locale names without explicit encoding suffix, mined also checks generic locale name suffix values and assumes the corresponding terminal encoding. Thus the given names or suffixes can be used even on legacy systems without locale support to indicate the terminal environment and preferred text encoding properly to mined.

Keyboard mapping (Input method)

The mined distribution contains a file src/keymaps.cfg and a script mkkbmap; go into the src directory and use the script to generate additional keyboard mappings: The parameter to the mkkbmap script can be one of

path.../name.mim
a keyboard mapping file of the m17n-db multilingualization package
path.../name.kmap
a keyboard mapping file of the yudit text editor
path.../name.vim
a keyboard mapping file of the vim text editor
path.../name.cit
an input method mapping file of the cxterm terminal, binary form; only works if the cxterm binary/text conversion utility cit2tit is accessible
path.../name.tit
an input method mapping file of the cxterm terminal, text form; only works if the character set conversion utility iconv is accessible and works on the mapping file
path.../name.utf
an input method mapping file of the cxterm terminal, already converted to UTF-8 encoding (e.g. with iconv)
Cangjie [ < HKSCS Changjie table file name > ]
with this tag, a keyboard mapping for the Cangjie input method will be generated, taking information from the Unihan database (unicode.org);
with a second parameter, a Big5-encoded table of HKSCS Changjie input codes will be merged in, the parameter is either the file name or a + sign which is implicitly expanded to the relative path name etc/charmaps/hkscs/hkscs-2004-cj.txt; the HKSCS input codes file should be taken from http://info.gov.hk/digital21/eng/hkscs/
MainlandTelegraph , TaiwanTelegraph
with one of these tags, a keyboard mapping will be generated using one of these telegraph codes as an input method, taking information from the Unihan database (unicode.org)
Cantonese , HanyuPinlu , Mandarin , Tang
with one of these tags, a keyboard mapping will be generated using the according Chinese pronunciation as an input method, taking information from the Unihan database (unicode.org)
JapaneseKun , JapaneseOn
with one of these tags, a keyboard mapping will be generated using Japanese or Sino-Japanese pronunciation as an input method, taking information from the Unihan database (unicode.org)
Korean , Vietnamese
with one of these tags, a keyboard mapping will be generated using Korean or Vietnamese pronunciation as an input method, taking information from the Unihan database (unicode.org)
VIQR , VNI , Vtelex
with one of these tags, a keyboard mapping will be generated for the respective Vietnamese input methods, taking character information from the Unicode database (unicode.org)
script tag
for many scripts listed in the UnicodeData.txt database, character names listed there can build a useful keyboard mapping; mkkbmap will then generate an according keyboard mapping file, e.g. for Bopomofo

Each successful generation of a mapping table adds an entry to the configuration file keymaps.cfg; the entry is however initially disabled as it usually needs manual adjustment: edit the configuration file; enable the new entry by removing the leading '#' character, check the first element which will be the name of the mapping to appear in the Input Method menu, check the last element of the entry which is a two-letter shortcut and must be unique for all mappings, then move the entry to the position where you want it to appear in the menu. You can also group mappings by adding "-" lines in this configuration file.
For the Unicode data version used for included keyboard mappings, see the mined change log.
For the keyboard mappings generated from Unihan data, characters are sorted according to the priorities of their Unicode ranges (assigning lower priority to "Supplement" and "Extension" and "Compatibility" ranges). So for some input mnemos, the "pick list" for the Cangjie input method is displayed more in order of relevance (since 2000.10).
For keyboard mappings for CJK encodings, mkkbmap will add appropriate punctuation mapping entries for Chinese, Japanese, Korean, respectively, in addition to the entries derived from the respective data source.

---

MSDOS-only notes

DOS binaries: Two DOS-based versions, compiled with djgpp and with cygwin, are available for download from the mined web site http://towo.net/mined/ for users who want a quick binary on DOS-based systems. The djgpp binary is a "dual-mode" executable which runs on plain DOS and also supports long file names in a Windows 98/2000/XP/... DOS box (not NT4.0). It does not run in an xterm, however.

Highlight mode: The ANSI codes for selecting normal and exposed display can be chosen with the environment variable MINEDCOL. The two selections are separated by a space. Each selection is a semicolon-separated list of the code values. The default behaviour corresponds to the setting

set MINEDCOL=7 27

Example: Green on red text, red on green status:

set MINEDCOL=34;42 32;44

For command line options, "/" can be used instead of "-".

The "ESC -" command cannot go back within a group of files named by the same wildcard expression. It goes to the previous file name (or wildcard expression) instead.

Enabling the keypad HOP key: If you have a very old and crappy BIOS, you may have to enable use of the cursor block "5" key (for use as a HOP key) with a TSR driver (ENHKBD.COM) or an enhanced keyboard driver. (Older PC keyboard drivers were often so ignorant to forbid you to use that key.)

Immediate adjustment to changed window size does not work in the DOS version if the size change is caused by a TSR (e.g. VGAMAX using a hotkey); in that case, mined adjusts its screen display only after the next key is typed.

The cygwin terminal environment (cygwin in a DOS box window) provides an emulation of a Unix 8 bit character set so non-ASCII characters entered in this version are different from those entered in other DOS-based versions. Editing UTF-8 text, on the other hand, works transparently in all DOS-based versions. See PC terminals above for more details.

In order to enable mouse use in a DOS box, deactivate "QuickEdit mode" in the properties menu.

The following only applies if DOS ANSI driver output is used which is currently not the case in any configuration:

The default colour setting depends on an extended ANSI driver (like NNANSI) as does the scroll down function anyway. Unfortunately, there is no way to find out the current colour setting nor is there an inverse video mode in many ANSI drivers (only a fixed black on white mode) so that it is impossible to implement just inverse display for highlighting. Therefore, if mined thinks to see an ANSI driver of the simpler kind, it will change its colour setting defaults. In any case, these can be overridden with the MINEDCOL variable.

Recommended ANSI drivers:

NNANSI by Tom Almy (very capable, but needs some installation effort), or

ANSI.COM by Michael J. Mefford (small, works well at usual screen sizes).

Mined tries to analyse the ANSI drivers capabilities by checking some control sequences. This works, however, only if the ANSI driver is at least able to send cursor position reports. For primitive ANSI drivers that cannot even do that, mined's operation can be ensured with an emergency procedure: A faked pseudo-report should be stuffed into mined as its first input (with some key-stuffing program) and mined will use no further cursor position requests. It will also assume a simple ANSI driver then. The faked report should consist of the screen size in lines and columns, embedded at the positions of the ANSI cursor report sequence but with different surrounding characters. For an invocation of mined on a 25 lines and 80 columns screen a batch file for this would look like:

keypress xx25x80xx
mined %1 %2 %3 %4 %5 %6 %7 %8 %9

The remaining remarks apply to the Turbo-C version only which is no longer supported (use djgpp instead):

• The file size being edited is limited to 200KB to 500KB (depending on average line length and number of lines).
• Typing of Control-P while display output is active (i.e., during screen paging) can hang the system. Typing of Control-C or Control-Break while display output is active can at least leave some garbage on the screen. Control-S may stop screen output until Control-Q is typed. Typing of Control-P, Control-C, or Control-Break while a search operation is active can be desastrous. (Can anyone tell me how to disable BIOS/MSDOS interpretation of these characters from Turbo-C?)
• The Turbo-C version is configured to handle screen output using the "conio" module. (It used to use an ANSI driver.) The disadvantage of conio is that it doesn't handle arbitrary screen modes and sizes whereas good ANSI drivers support them all.

ENVIRONMENT VARIABLES

Environment variables for configuration of mined are listed in the script file profile.mined in the Mined runtime support library together with explanations and suggested values.
Further variables used by mined in the usual meaning are:

HOME

USER

SHELL

SYS$SCRATCH (VMS)

TMPDIR

TMP

TEMP (MSDOS)

TERM

Terminal type to be assumed.

ESCDELAY

Delay after an ESCAPE character that mined waits for recognition of a function key control sequence. Default is 450 ms.

MAPDELAY (non-standard)

Similar delay that mined applies to wait for subsequent input characters when applying keyboard mapping for an input method. Default is 900 ms.

LINES, COLUMNS (MSDOS ANSI mode only)

Line / column count of terminal to be assumed.

windir

Used to determine if it runs under MS Windows and set some defaults (screen output delay) accordingly.

Environment variables to configure Printing

MINEDPRINT

Print command to use instead of uprint; the value must contain an embedded "%s" which is replaced with the file name.

FONT

Name of a font file, e.g. LucidaBrightRegular or bodoni.ttf for use with uprint/uniprint (the file must reside in the configured font path), or name of a font as specified with fontconfig (in $HOME/.fonts.conf or /etc/fonts/fonts.conf) for use with uprint/paps.

FONTPATH

Directory search path (separate directory names with ":") for use with uprint/uniprint which uses Truetype fonts.

FONTSIZE

Font size to be used with uprint (paps or uniprint).

LPR

Print spooling command to be used by uprint (or mined itself if uprint does not work) instead of the system-specific print spooling command (e.g. lpr).

PRINTER

Name of printer to spool to.

FILES

Unix

$MINEDDIR

directory in which the Mined runtime support library is installed, including the online help file mined.hlp and the printing script uprint

$MINEDDIR/help/mined.hlp

online help file, first attempt to find it

$0/mined.hlp

online help file in mined program directory, next attempt

/usr/share/mined/help/mined.hlp

online help file, next attempt

/usr/local/share/mined/help/mined.hlp

online help file, next attempt

/usr/share/lib/mined/help/mined.hlp

online help file, next attempt

/opt/mined/share/help/mined.hlp

online help file, next attempt

/usr/share/doc/packages/mined/help/mined.hlp

online help file, next attempt

---

$MINEDTMP

directory for auxiliary files, first attempt

Using this variable and $MINEDUSER (see below), you can establish copy and paste among machines that share network directories but are normally configured to use separate (usually local) temporary directories.

$TMPDIR

directory for auxiliary files, next attempt

$TMP

directory for auxiliary files, next attempt

$TEMP

directory for auxiliary files, next attempt

/usr/tmp

directory for auxiliary files, next attempt

/tmp

directory for auxiliary files, next attempt

$MINEDUSER

user name assumed instead of $USER for building auxiliary file names; using this, common copy-and-paste buffers can be used on a network file system from different machines where the user possibly has different user names

---

$HOME/.fonts.conf

fonts configuration file for use with uprint/paps; for description, see http://fontconfig.org/fontconfig-user.html or man fonts.conf

---

minedbuf.< USER >.< PID >.< NN >

temporary file for paste buffer; USER is either $MINEDUSER or $USER

minedbuf.< USER >

file for inter-window paste buffer; USER is either $MINEDUSER or $USER; see descriptions of $MINEDTMP and $MINEDUSER above for how to set up a common inter-window paste buffer in a heterogeneous network

minedpanic.< USER >.< PID >

panic file to rescue text in case of crash or external signal caught

---

VMS

SYS$MINEDTMP:$MINEDBUF$user.pid.nn

paste buffer, first attempt

SYS$SCRATCH:$MINEDBUF$user.pid.nn

paste buffer, next attempt

SYS$SCRATCH:$MINEDPANIC$user.pid

panic file, first attempt

SYS$MINEDTMP:$MINEDBUF$user

inter-window paste buffer, first attempt

SYS$SCRATCH:$MINEDBUF$user

inter-window paste buffer, next attempt

SYS$SCRATCH:$MINEDPANIC$user.pid

panic file, next attempt

If SYS$SCRATCH is not available, SYS$LOGIN is used instead.

---

MSDOS / Windows

%MINEDDIR%\help\mined.hlp

online help file, first attempt (to find it)

mined.hlp (in mined program directory)

online help file, next attempt

%MINEDTMP%\minedbuf.nn

paste buffer

%MINEDTMP%\minedbuf

inter-window paste buffer

%MINEDTMP%\minedbuf.%MINEDUSER%

inter-window paste buffer, as configured to use the same file as other mined versions in a heterogeneous network; note, however, that %MINEDUSER% will be shortened to 3 characters in pure DOS

%MINEDTMP%\mined-pa.nic

panic file

If %MINEDTMP% is not available, %TEMP% or %TMP% or \ are used.

DIAGNOSTICS

In all cases where it is considered sensible, the appropriate message of a system error occurred is displayed (instead of printing numerical hieroglyphs or indistinguished commonplace messages as many other UNIX tools do).

BUGS

In an extremely narrow terminal window (less than 8 characters), if lines are shifted out of the display, moving the cursor around may cause positioning errors and display garbage.

(Unix:) Mined cannot edit a pipe or device file and hangs if you try to do so. (But it can insert from, or write to, a pipe.)
This restriction does not refer to editing from standard input in a piped command like cmd | mined which works of course.

(MSDOS:) Piped editing from standard input does not work for unknown reason. This restriction does not apply to the cygwin version.

(MSDOS, xterm:) Non-cygwin versions (djgpp etc.) do not work in an xterm for unknown reason. The cygwin version does of course work in xterm.

AUTHOR AND ACKNOWLEDGEMENTS

Long ago, the initial version of mined was written for the Minix educational operating system by Michiel Huisjes. It was adapted to Unix by Achim Müller who added termcap support. Mined was later debugged, partly rewritten and enhanced and is now maintained by Thomas Wolff.
Please send comments, suggestions, bug reports to mined@towo.net.

Mailing list

Mined is also hosted as a sourceforge project (sf.net/projects/mined) where a mailing list is available. To subscribe for information about updates, or discussion, error reports, and feature requests, or to send a mail, please go to the Mined mailing list page.

Acknowledgements

• Thanks to Nadim Shaikli < shaikli @ yahoo.com > for discussion of right-to-left issues and interworking with mlterm.
• Thanks to Mike Fabian < mfabian @ suse.de > for making the RPM package included in the SuSE distribution.
• Thanks to Ziying Sherwin < sherwin @ nlm.nih.gov > and R. P. Channing Rodgers < rodgers @ nlm.nih.gov > for suggestions and information about CJK input method support and multiple choice handling (pick lists).
• Thanks to Tobias Ernst < tobias_ernst @ eml.cc > for providing a Mac OS X makefile and suggestion and information to implement Emacs command mode.
• Thanks to 吴咏炜 (Wu Yongwei) < yongwei @ eastday.com > for suggestions and information about Pinyin input methods, for discussion about keyboard mappings for CJK punctuation, and for further maintaining the Pinyin input method.
• Thanks to Ramakrishnan Muthukrishnan < rkrishnan @ debian.org > for making the Debian package.
• Thanks to Thierry Thomas < thierry @ FreeBSD.org > for making the FreeBSD package.
• Thanks to Jim Breen for suggesting better overview of input methods and more language-specific advice for non-techy persons which led to the new chapter on Language support.