Term::VT102(3) a class to emulate a DEC VT102 terminal

SYNOPSIS


use Term::VT102;
my $vt = Term::VT102->new ('cols' => 80, 'rows' => 24);
while (<>) { $vt->process ($_); }

DESCRIPTION

The VT102 class provides emulation of most of the functions of a DEC VT102 terminal. Once initialised, data passed to a VT102 object is processed and the in-memory ``screen'' modified accordingly. This ``screen'' can be interrogated by the external program in a variety of ways.

This allows your program to interface with full-screen console programs by running them in a subprocess and passing their output to a VT102 class. You can then see what the application has written on the screen by querying the class appropriately.

OPTIONS

Setting cols or rows in the new() hash allows you to change the size of the terminal being emulated. If you do not specify a size, the default is 80 columns by 24 rows.

After initialisation, you can read and set the following terminal options using the option_read() and option_set() methods:

  LINEWRAP      line wrapping; 1=on, 0=off. Default is OFF.
  LFTOCRLF      treat LF (\n) as CRLF (\r\n); 1=on, 0=off. Default OFF.
  IGNOREXOFF    ignore XON/XOFF characters; 1=on (ignore). Default ON.

METHODS

The following methods are provided:
attr_pack ($fg,$bg,$bo,$fa,$st,$ul,$bl,$rv)
Returns the packed version of the given attribute settings, which are given in the same order as returned by attr_unpack. The packed version will be a binary string not longer than 2 bytes.
attr_unpack ($data)
Returns a list of the contents of the given packed attribute settings, of the form ($fg,$bg,$bo,$fa,$st,$ul,$bl,$rv).

$fg and $bg are the ANSI foreground and background text colours, and $bo, $fa, $st, $ul, $bl, and $rv are flags (1 = on, 0 = off) for bold, faint, standout, underline, blink and reverse respectively.

callback_call ($name, $par1, $par2)
Calls the callback $name (eg 'ROWCHANGE') with parameters $par1 and $par2, as if the VT102 module had called it. Does nothing if that callback has not been set with callback_set ().
callback_set ($callback, $ref, $private)
Sets the callback callback to function reference ref with private data $private.

See the section on CALLBACKS below.

new (%config)
Returns a new VT102 object with options specified in %config (see the OPTIONS section for details).
option_read ($option)
Returns the current value of terminal option $option (see OPTIONS for details), or undef if that option does not exist. Note that you cannot read the terminal size with this call; use size() for that.
option_set ($option, $value)
Sets the current value of terminal option $option to $value, returning the old value or undef if no such terminal option exists or you have specified an undefined $value. Note that you cannot resize the terminal with this call; use resize() for that.
process ($string)
Processes the string $string (which can be zero-length), updating the VT102 object accordingly and calling any necessary callbacks on the way.
resize ($cols, $rows)
Resizes the VT102 terminal to cols columns by rows rows, eg $vt->resize (80, 24). The virtual screen is cleared first.
reset ()
Resets the object to its ``power-on'' state.
row_attr ($row, [$startcol, $endcol])
Returns the attributes for row $row (or undef if out of range) as a string of packed attributes, each character cell's attributes being 2 bytes long. To unpack the attributes for a given cell, use substr(), eg $attr=substr($row,4,2) would set $attr to the attributes for cell 3 (steps of 2: 0 .. 2 .. 4, so 4 means the 3rd character). You would then use the attr_unpack() method to unpack that character cell's attributes.

If $startcol and $endcol are defined, only returns the part of the row between columns $startcol and $endcol inclusive instead of the whole row.

row_text ($row, [$startcol, $endcol])
Returns the textual contents of row $row (or undef if out of range), with totally unused characters being represented as NULL (\0). If $startcol and $endcol are defined, only returns the part of the row between columns $startcol and $endcol inclusive instead of the whole row.
row_plaintext ($row, [$startcol, $endcol])
Returns the textual contents of row $row (or undef if out of range), with unused characters being represented as spaces. If $startcol and $endcol are defined, only returns the part of the row between columns $startcol and $endcol inclusive instead of the whole row.
row_sgrtext ($row, [$startcol, $endcol])
Returns the textual contents of row $row (or undef if out of range), with unused characters being represented as spaces, and ANSI/ECMA-48 escape sequences (CSI SGR) used to set the colours and attributes as appropriate. If $startcol and $endcol are defined, only returns the part of the row between columns $startcol and $endcol inclusive instead of the whole row.

Use row_sgrtext to get a row if you want to output it to a real terminal and preserve all colours, bold, etc.

sgr_change ($source, $dest)
Returns a string containing ANSI/ECMA-48 escape sequences to change colours and attributes from $source to $dest, which are both packed attributes (see attr_pack). This is used internally by row_sgrtext.
cols ()
Return the number of columns in the VT102 object.
rows ()
Return the number of rows in the VT102 object.
size ()
Return a pair of values (columns,rows) denoting the size of the terminal in the VT102 object.
x ()
Return the current cursor X co-ordinate (1 being leftmost).

Note: It is possible for the current X co-ordinate to be 1 more than the number of columns. This happens when the end of a row is reached such that the next character would wrap on to the next row.

y ()
Return the current cursor Y co-ordinate (1 being topmost).
cursor ()
Return the current cursor state (1 being on, 0 being off).
xtitle ()
Return the current xterm window title.
xicon ()
Return the current xterm window icon name.
status ()
Return a list of values ($x,$y,$attr,$ti,$ic), where $x and $y are the cursor co-ordinates (1,1 = top left), $attr is a packed version of the current attributes (see attr_unpack), $ti is the xterm window title, and $ic is the xterm window icon name.
version ()
Return the version of the VT102 module being used.

CALLBACKS

Callbacks are the processing loop's way of letting your main program know that something has happened. They are called while in a process() loop.

To specify a callback, use the callback_set interface, giving a reference to the function to call. Your function should take five scalar arguments: the VT102 object being processed, the name of the callback, and two arguments whose value depends on the callback, as shown below. The final argument is the private data scalar you passed when you called callback_set.

The name of the callback is passed to the callback function so that you can have one function to handle all callbacks if you wish.

Available callback names are:

  BELL          BEL (beep, \007) character received
  CLEAR         screen about to be cleared
  OUTPUT        data (arg1) to be sent back to data source
  ROWCHANGE     screen row (row number is argument 1) content has changed
  SCROLL_DOWN   about to scroll down (arg1=top row, arg2=num to scroll)
  SCROLL_UP     about to scroll up (ditto)
  UNKNOWN       unknown/unsupported code (arg1=name, arg2=code/sequence)
  STRING        string received (arg1=source, eg PM, APC, arg2=string)
  XICONNAME     xterm icon name to be changed to arg1
  XWINTITLE     xterm title name to be changed to arg1
  LINEFEED      line feed about to be processed (arg1=row)
  GOTO          cursor about to be moved (args=new pos)

Note that the wording of the above is significant in terms of exactly when the callback is called. For instance, CLEAR is called just before the screen is cleared, whereas ROWCHANGE is called after the given row has been changed.

A good callback handler for OUTPUT is to simply syswrite() argument 1 to your data source - eg if you're reading from a telnet session, write that argument straight to it. It is used for cursor position request responses and suchlike.

Note that SCROLL_DOWN is called when scrolling down, so text is about to move UP the screen; arg1 will be the row number of the bottom of the scrolling region, and arg2 will be the number of rows to be scrolled. Likewise, SCROLL_UP is called when text is about to move down; arg1 will be the row number of the top of the scrolling region.

The STRING callback is called for escape sequences that contain a string that would otherwise be ignored, such as DSC, PM, and APC. The first argument is the escape sequence that contained the string, such as DSC, and the second argument is the string itself. This callback doesn't get called for OSC strings.

The LINEFEED callback can be thought of as ``line completed'', it's called when LF, NEL or IND are about to be processed or just before a line wraps, so it generally indicates that an application has finished updating a particular line on the screen. Handy for scrollback buffer processing.

The GOTO callback is only called just before the cursor is explicitly moved, by one of CUU, CUD, VPR, CUF, HPR, CUB, CNL, CPL, CHA, HPA, CUP, HVP. The parameters give the destination column and row, without taking scrolling and boundaries into account.

Finally, note that ROWCHANGE is only triggered when text is being entered; screen scrolling or screen clearance does not trigger it, that would trigger a SCROLL_DOWN or SCROLL_UP or CLEAR. Line or character insertion or deletion will cause one or more ROWCHANGE callbacks, however.

SUPPORTED CODES

The following sequences are supported:

   007 (BEL)   beep
   010 (BS)    backspace
   011 (HT)    horizontal tab to next tab stop
   012 (LF)    line feed
   013 (VT)    line feed
   014 (FF)    line feed
   015 (CR)    carriage return
   021 (XON)   resume transmission (only if option IGNOREXOFF is cleared)
   023 (XOFF)  stop transmission (only if option IGNOREXOFF is cleared)
   030 (CAN)   interrupt escape sequence
   032 (SUB)   interrupt escape sequence
   033 (ESC)   start escape sequence
   177 (DEL)   ignored
   233 (CSI)   same as ESC [
   ESC 7 (DECSC)   save state
   ESC 8 (DECRC)   restore most recently saved state
   ESC H (HTS)     set tab stop at current column
   ESC g           visual beep - treated as BEL
   ESC # 8 (DECALN)  DEC screen alignment test - fill screen with E's
   CSI @ (ICH)     insert blank characters
   CSI A (CUU)     move cursor up
   CSI B (CUD)     move cursor down
   CSI C (CUF)     move cursor right
   CSI D (CUB)     move cursor left
   CSI E (CNL)     move cursor down and to column 1
   CSI F (CPL)     move cursor up and to column 1
   CSI G (CHA)     move cursor to column in current row
   CSI H (CUP)     move cursor to row, column
   CSI J (ED)      erase display
   CSI K (EL)      erase line
   CSI L (IL)      insert blank lines
   CSI M (DL)      delete lines
   CSI P (DCH)     delete characters on current line
   CSI X (ECH)     erase characters on current line
   CSI a (HPR)     move cursor right
   CSI c (DA)      return ESC [ ? 6 c (VT102)
   CSI d (VPA)     move to row (current column)
   CSI e (VPR)     move cursor down
   CSI f (HVP)     move cursor to row, column
   CSI m (SGR)     set graphic rendition
   CSI n (DSR)     device status report
   CSI r (DECSTBM) set scrolling region to (top, bottom) rows
   CSI s (CUPSV)   save cursor position
   CSI u (CUPRS)   restore cursor position
   CSI ` (HPA)     move cursor to column in current row
   CSI g (TBC)     clear tab stop (CSI 3 g = clear all stops)

LIMITATIONS

Unknown escape sequences and control characters are ignored. All escape sequences pertaining to character sets are ignored.

The following known control characters / sequences are ignored:

   005 (ENQ)   trigger answerback message
   016 (SO)    activate G1 charset, carriage return
   017 (SI)    activate G0 charset

The following known escape sequences are ignored:

   ESC %@ (CSDFL)    select default charset (ISO646/8859-1)
   ESC %G (CSUTF8)   select UTF-8
   ESC %8 (CSUTF8)   select UTF-8 (obsolete)
   ESC (8 (G0DFL)    G0 charset = default mapping (ISO8859-1)
   ESC (0 (G0GFX)    G0 charset = VT100 graphics mapping
   ESC (U (G0ROM)    G0 charset = null mapping (straight to ROM)
   ESC (K (G0USR)    G0 charset = user defined mapping
   ESC (B (G0TXT)    G0 charset = ASCII mapping
   ESC )8 (G1DFL)    G1 charset = default mapping (ISO8859-1)
   ESC )0 (G1GFX)    G1 charset = VT100 graphics mapping
   ESC )U (G1ROM)    G1 charset = null mapping (straight to ROM)
   ESC )K (G1USR)    G1 charset = user defined mapping
   ESC )B (G1TXT)    G1 charset = ASCII mapping
   ESC *8 (G2DFL)    G2 charset = default mapping (ISO8859-1)
   ESC *0 (G2GFX)    G2 charset = VT100 graphics mapping
   ESC *U (G2ROM)    G2 charset = null mapping (straight to ROM)
   ESC *K (G2USR)    G2 charset = user defined mapping
   ESC +8 (G3DFL)    G3 charset = default mapping (ISO8859-1)
   ESC +0 (G3GFX)    G3 charset = VT100 graphics mapping
   ESC +U (G3ROM)    G3 charset = null mapping (straight to ROM)
   ESC +K (G3USR)    G3 charset = user defined mapping
   ESC >  (DECPNM)   set numeric keypad mode
   ESC =  (DECPAM)   set application keypad mode
   ESC N  (SS2)      select G2 charset for next char only
   ESC O  (SS3)      select G3 charset for next char only
   ESC P  (DCS)      device control string (ended by ST)
   ESC X  (SOS)      start of string
   ESC ^  (PM)       privacy message (ended by ST)
   ESC _  (APC)      application program command (ended by ST)
   ESC \  (ST)       string terminator
   ESC n  (LS2)      invoke G2 charset
   ESC o  (LS3)      invoke G3 charset
   ESC |  (LS3R)     invoke G3 charset as GR
   ESC }  (LS2R)     invoke G2 charset as GR
   ESC ~  (LS1R)     invoke G1 charset as GR

The following known CSI (ESC [) sequences are ignored:

   CSI q (DECLL)   set keyboard LEDs

The following known CSI (ESC [) sequences are only partially supported:

   CSI h (SM)      set mode (only support CSI ? 25 h, cursor on/off)
   CSI l (RM)      reset mode (as above)

EXAMPLES

For some examples, including how to interface Term::VT102 with Net::Telnet or a command such as SSH, please see the examples/ directory in the distribution.

AUTHORS

Copyright (C) 2003 Andrew Wood "<andrew dot wood at ivarch dot com>". Distributed under the terms of the Artistic License 2.0.

Credit is also due to:

  Charles Harker <CHarker at interland.com>
    - reported and helped to diagnose a bug in the handling of TABs
  Steve van der Burg <steve.vanderburg at lhsc.on.ca>
    - supplied basis for an example script using Net::Telnet
  Chris R. Donnelly <cdonnelly at digitalmotorworks.com>
    - added support for DECTCEM, partial support for SM/RM
  Paul L. Stoddard
    - reported a possible bug in cursor movement handling
  Joerg Walter
    - provided a patch for Unicode handling

THINGS TO WATCH OUT FOR

Make sure that your code understands NULL (\000) - you will get this in strings where nothing has been printed on the screen. For instance, the sequence ``12\e[C34'' (``12'', ``CUF (move right)'', ``34'') you might think would yield the string ``12 34'', but in fact it can also yield ``12\00034'' - that is, ``12'' followed by a zero byte followed by ``34''. This is because the screen's contents defaults to zeroes, not spaces.

To avoid that, use row_plaintext, which will convert NULLs to spaces, instead of row_text.

Different types of terminal disagree on certain corner cases. For example, gnome-terminal and screen handle TAB stops and TABbing past the end of the screen in slightly different ways. Term::VT102 is closer to screen in the way it handles this sort of thing.