mlterm-dev-en — mlterm mailing list(English)

Re: [Mlterm-dev-en] Re: [I18n]Bidi on mlterm

[Noah L. <nl...@co...>]

Hello,

I checked mlterm out of cvs a couple of hours ago, and I'm
having crash and other problems with bidi and with utf8
generally, even after applying your patch. I'm using linux
2.4.8 and xfree86 3.3.6. When I start mlterm
$ LANG=en_US.UTF-8 mlterm
I get a blank window. The blank window happens with or
without bidi. I open mlconfig and set the font size to 14
(it starts at 17), and then it crashes:

(gdb) bt
#0  0x4033bcc6 in free () from /lib/libc.so.6
#1  0x4033bc24 in free () from /lib/libc.so.6
#2  0x4014e4e6 in kik_mem_free (ptr=0x809b440, file=0x8069cbd "ml_char.c", 
    line=148, func=0x8069cc7 "ml_str_delete") at kik_mem.c:185
#3  0x0804fa80 in ml_str_delete (str=0x809b440, size=80) at ml_char.c:148
#4  0x0805dce2 in ml_imgline_final (line=0x80934a8) at ml_image_line.c:81
#5  0x08051b2e in ml_image_resize (image=0x8090680, num_of_cols=114, 
    num_of_rows=38) at ml_image.c:904
#6  0x08058b6e in window_resized (win=0x8090578) at ml_term_screen.c:1110
#7  0x0805a35d in font_size_changed (termscr=0x8090578)
    at ml_term_screen.c:2518
#8  0x0805a3af in change_font_size (p=0x8090578, font_size=14)
    at ml_term_screen.c:2539
#9  0x08063fda in sig_child (self=0x8090804, pid=15020) at ml_config_menu.c:267
#10 0x08068981 in sig_child (sig=17) at ml_sig_child.c:52
#11 0x402e97c8 in killpg () from /lib/libc.so.6
#12 0x080669c9 in ml_term_manager_event_loop (term_man=0xbffff670)
    at ml_term_manager.c:1523
#13 0x0805fcf9 in main (argc=1, argv=0xbffff9b4) at main.c:44
#14 0x402d75b0 in __libc_start_main () from /lib/libc.so.6

On the other hand, if I specify the font size when I start
mlterm:
$ LANG=en_US.UTF-8 mlterm --fsrange=14-14
then the window opens and I see a prompt. When I press a
key, it crashes:

(gdb) bt
#0  0x402e9971 in kill () from /lib/libc.so.6
#1  0x402e9775 in raise () from /lib/libc.so.6
#2  0x402eada3 in abort () from /lib/libc.so.6
#3  0x08068183 in ml_bidi (order=0x8090900, src=0x8087830, size=27)
    at ml_bidi.c:99
#4  0x0805ef1e in ml_imgline_render_bidi (line=0x8087540)
    at ml_image_line.c:910
#5  0x08051df4 in ml_image_render_bidi (image=0x8091630) at ml_image.c:1025
#6  0x0805b896 in ml_term_screen_stop_vt100_cmd (termscr=0x8091528)
    at ml_term_screen.c:3860
#7  0x08056354 in ml_parse_vt100_sequence (vt100_parser=0x80a1558)
    at ml_vt100_parser.c:2426
#8  0x0806510a in receive_next_event (term_man=0xbffff660)
    at ml_term_manager.c:619
#9  0x080669c9 in ml_term_manager_event_loop (term_man=0xbffff660)
    at ml_term_manager.c:1523
#10 0x0805fcf9 in main (argc=2, argv=0xbffff9a4) at main.c:44
#11 0x402d75b0 in __libc_start_main () from /lib/libc.so.6

This last one's less critical.

When I run mlterm with bidi disabled, it works, but it seems
to refuse to use the fonts I specify in ~/.mlterm/font. So,
I copied those fonts to their own directory and tried this:

$ ls -l
-rw-r--r--    1 nlevitt  nlevitt    784156 Dec  9 22:37 unifontd.pcf.gz
-rw-r--r--    1 nlevitt  nlevitt     74014 Dec  9 22:37 unifonts.pcf.gz
$ mkfontdir
$ cat fonts.dir 
2
unifontd.pcf.gz -gnu-fixed-medium-r-normal-f-16-160-75-75-c-80-iso10646-1
unifonts.pcf.gz -gnu-fixed-medium-r-normal-h-8-160-75-75-c-80-iso10646-1
$ xset fp $PWD
$ xset fp rehash
$ LANG=en_US.UTF-8 mlterm --bi=false
a font for displaying us ascii chars is not found.

This is dvdeug's version of unifont, which I have used
successfully with xterm-utf8. What's wrong here? 

Noah

Re: [Mlterm-dev-en] LOCALE problems

[Tomohiro K. <tk...@ri...>]

Hi,

At Sun, 06 Jan 2002 17:01:27 -0800,
Scott Jonathan wrote:

> I recently did an upgrade from Debain Potato 2.2 to Debian Woody 3.0. Now 
> my locale settings are all in disarray it seems. Could someone point me to 
> information on how to get it back to normal?

I think you have installed "locales" package.  When you installed it,
didn't you asked about the installing locales?  If you checked 
"ja_JP.EUC-JP EUC-JP", you will have ja_JP.eucJP locale.

If you have not installed the package, please install it.  Otherwise,
try "dpkg-reconfigure locales" and you will be asked which locales to
be installed.

---
久保田智広 Tomohiro KUBOTA <ku...@de...>
http://www.debian.or.jp/~kubota/
"Introduction to I18N"  http://www.debian.org/doc/manuals/intro-i18n/

Re: [Mlterm-dev-en] LOCALE problems

[Araki K. <j00...@ip...>]

	Hi,

	Subject: [Mlterm-dev-en] LOCALE problems
	From: "Scott Jonathan" <jo...@ho...>
	Message-ID: <F60...@ho...>
	Date: Sun, 06 Jan 2002 17:01:27 -0800

> I recently did an upgrade from Debain Potato 2.2 to Debian Woody 3.0. Now 
> my locale settings are all in disarray it seems. Could someone point me to 
> information on how to get it back to normal?
>
> "locale -a | grep ja" shows me:
> ja_JP
> ja_JP.ujis
> japanese
> japanese.euc
> japanese.sjis
>
> There used to be a locale "ja_JP.eucJP" but it is no longer showing up. And 
> as a consequence, I cannot input or output Japanese in mlterm. 

        I don't use Linux that I have no idea why ja_JP.eucJP locale lost ,
        but does
        
        $ LC_CTYPE=ja_JP.ujis mlterm

        or

        $ export LC_CTYPE=ja_JP.ujis
        $ mlterm

        enable you to input or output Japanese ?

        # Encoding autodetection may fail under ja_JP.ujis locale , so -km eucjp option
        # or ENCOIDNG = eucjp entry in ~/.mlterm/man is recommended.

        In addition , even if mlterm fails to setlocale(LC_CTYPE) , it
        can at least output japanese with -km eucjp option(or select EUC-JP
        encoding from mlconfig).
	
--
kiken
j00...@ip...

[Mlterm-dev-en] LOCALE problems

[Scott J. <jo...@ho...>]

I recently did an upgrade from Debain Potato 2.2 to Debian Woody 3.0. Now 
my locale settings are all in disarray it seems. Could someone point me to 
information on how to get it back to normal?

"locale -a | grep ja" shows me:
ja_JP
ja_JP.ujis
japanese
japanese.euc
japanese.sjis

There used to be a locale "ja_JP.eucJP" but it is no longer showing up. And 
as a consequence, I cannot input or output Japanese in mlterm. 

Any and all help is much appreciated!

Jonathan Scott

_________________________________________________________________
かわいい & 使えるブラウザで、インターネット生活がもっと楽しくなる！ 
http://explorer.msn.co.jp/

Re: [Mlterm-dev-en] Re: [I18n]Bidi on mlterm

[Araki K. <j00...@ip...>]

	Hi,

	Subject: Re: [Mlterm-dev-en] Re: [I18n]Bidi on mlterm
	From: Araki Ken <j00...@ip...>
	Message-ID: <200...@pd...>
	Date: Mon, 07 Jan 2002 08:33:47 +0900

>         Isn't this problem fixed by an attached patch ?

> Index: src/kik_list.h
> ===================================================================
> RCS file: /home/ken/cvsroot/kiklib/src/kik_list.h,v
> retrieving revision 1.4
> diff -u -r1.4 kik_list.h
> --- src/kik_list.h	2001/12/01 17:28:59	1.4
> +++ src/kik_list.h	2002/01/06 23:25:57

        src/kik_list.h is a typo of kiklib/src/kik_list.h , sorry.

        Best Regards.
	
--
kiken
j00...@ip...

Re: [Mlterm-dev-en] Re: [I18n]Bidi on mlterm

[Araki K. <j00...@ip...>]

	Hi,

	Subject: [Mlterm-dev-en] Re: [I18n]Bidi on mlterm
	From: Karl Koehler <ko...@or...>
	Message-ID: <Pin...@r5...>
	Date: Sun, 6 Jan 2002 20:53:39 +0100 (MEZ)

> I am trying to get mlterm to work here. 
> I configured  with --enable-fribidi --enable-debug .
> I am running on a i686 with linux-2.2.16,
> LANG=de_DE.utf8 .
> When I start mlterm-2.1.2 with the options below:
>> $ mlterm --km utf8 --bi
> It will crash quite fast:
>
> Program received signal SIGSEGV, Segmentation fault.
> search_mem_log (ptr=0x819cd98) at kik_mem.c:81
> 81                      else if( kik_iterator_indirect( iterator)->ptr == 
> ptr)
> (gdb) where
> #0  search_mem_log (ptr=0x819cd98) at kik_mem.c:81
> #1  0x401414fb in kik_mem_free (ptr=0x819cd98, file=0x8069e8b "ml_char.c", line=254, 
>     func=0x8069f98 "ml_char_final") at kik_mem.c:168
> #2  0x080503af in ml_char_set (ch=0x809fe00, bytes=0xbffff1a4 "a", size=1, font=0x8085df8, 
>     font_decor=0, fg_color=MLC_FG_COLOR, bg_color=MLC_BG_COLOR) at ml_char.c:254
> #3  0x08055048 in put_char (vt100_parser=0x809ede8, ch=0xbffff1a4 "a", len=1, cs=US_ASCII, 
>     prop=0) at ml_vt100_parser.c:334
> #4  0x08056f2a in ml_parse_vt100_sequence (vt100_parser=0x809ede8) at ml_vt100_parser.c:2363
> #5  0x08066132 in receive_next_event (term_man=0xbffff2b4) at ml_term_manager.c:617
> #6  0x08067a69 in ml_term_manager_event_loop (term_man=0xbffff2b4) at ml_term_manager.c:1473
> #7  0x08060cff in main (argc=4, argv=0xbffff5e4) at main.c:44
> #8  0x402cb7ee in __libc_start_main () from /lib/libc.so.6
> (gdb) 
>
> As others seem to have used it already, I am wandering what I'm 
> doing wrong.

        Thanks a lot for you report.
        
        You didn't do anything wrong , I think , and I found a suspicious and very
        critical bug around kik_list.h

        Isn't this problem fixed by an attached patch ?
	
--
kiken
j00...@ip...

Index: src/kik_list.h
===================================================================
RCS file: /home/ken/cvsroot/kiklib/src/kik_list.h,v
retrieving revision 1.4
diff -u -r1.4 kik_list.h
--- src/kik_list.h	2001/12/01 17:28:59	1.4
+++ src/kik_list.h	2002/01/06 23:25:57
@@ -77,7 +77,7 @@
 { \
 	KIK_ITERATOR( type)   new_iterator = NULL ; \
 	\
-	if( ( new_iterator = malloc( sizeof( KIK_ITERATOR( type)))) == NULL) \
+	if( ( new_iterator = malloc( sizeof( *new_iterator))) == NULL) \
 	{ \
 		kik_error_printf( "malloc() failed in kik_list_insert_after().\n") ; \
 		abort() ; \
@@ -101,7 +101,7 @@
 { \
 	KIK_ITERATOR( type)   new_iterator = NULL ; \
 	\
-	if( ( new_iterator = malloc( sizeof( KIK_ITERATOR( type)))) == NULL) \
+	if( ( new_iterator = malloc( sizeof( *new_iterator))) == NULL) \
 	{ \
 		kik_error_printf( "malloc() failed in kik_list_insert_before().\n") ; \
 		abort() ; \
@@ -125,7 +125,7 @@
 { \
 	KIK_ITERATOR( type)   new_iterator = NULL ; \
 	\
-	if( ( new_iterator = malloc( sizeof( KIK_ITERATOR( type)))) == NULL) \
+	if( ( new_iterator = malloc( sizeof( *new_iterator))) == NULL) \
 	{ \
 		kik_error_printf( "malloc() failed in kik_list_insert_head().\n") ; \
 		abort() ; \
@@ -148,7 +148,7 @@
 { \
 	KIK_ITERATOR( type)   new_iterator = NULL ; \
 	\
-	if( ( new_iterator = malloc( sizeof( KIK_ITERATOR( type)))) == NULL) \
+	if( ( new_iterator = malloc( sizeof( *new_iterator))) == NULL) \
 	{ \
 		kik_error_printf( "malloc() failed in kik_list_insert_tail().\n") ; \
 		abort() ; \

[Mlterm-dev-en] Re: [I18n]Bidi on mlterm

[Karl K. <ko...@or...>]

Hi,

I am trying to get mlterm to work here. 
I configured  with --enable-fribidi --enable-debug .
I am running on a i686 with linux-2.2.16,
LANG=de_DE.utf8 .
When I start mlterm-2.1.2 with the options below:
> $ mlterm --km utf8 --bi
It will crash quite fast:

Program received signal SIGSEGV, Segmentation fault.
search_mem_log (ptr=0x819cd98) at kik_mem.c:81
81                      else if( kik_iterator_indirect( iterator)->ptr == 
ptr)
(gdb) where
#0  search_mem_log (ptr=0x819cd98) at kik_mem.c:81
#1  0x401414fb in kik_mem_free (ptr=0x819cd98, file=0x8069e8b "ml_char.c", line=254, 
    func=0x8069f98 "ml_char_final") at kik_mem.c:168
#2  0x080503af in ml_char_set (ch=0x809fe00, bytes=0xbffff1a4 "a", size=1, font=0x8085df8, 
    font_decor=0, fg_color=MLC_FG_COLOR, bg_color=MLC_BG_COLOR) at ml_char.c:254
#3  0x08055048 in put_char (vt100_parser=0x809ede8, ch=0xbffff1a4 "a", len=1, cs=US_ASCII, 
    prop=0) at ml_vt100_parser.c:334
#4  0x08056f2a in ml_parse_vt100_sequence (vt100_parser=0x809ede8) at ml_vt100_parser.c:2363
#5  0x08066132 in receive_next_event (term_man=0xbffff2b4) at ml_term_manager.c:617
#6  0x08067a69 in ml_term_manager_event_loop (term_man=0xbffff2b4) at ml_term_manager.c:1473
#7  0x08060cff in main (argc=4, argv=0xbffff5e4) at main.c:44
#8  0x402cb7ee in __libc_start_main () from /lib/libc.so.6
(gdb) 

As others seem to have used it already, I am wandering what I'm 
doing wrong.
Anyhow, It whould not crash, but rather warn me ..

	Karl

Re: [Mlterm-dev-en] mlterm ISCII support 2002/01/06

[Araki K. <j00...@ip...>]

	Hi,

	Subject: Re: [Mlterm-dev-en] mlterm ISCII support 2002/01/06
	From: Jyotirmoy Saikia <jm...@cs...>
	Message-ID: <Pin...@cs...>
	Date: Sun, 6 Jan 2002 15:26:40 -0500 (EST)

> I've downloaded it and also installed it. I've run the configure script
> using --enable-iscii option.

        Oh , I'm very sorry.
        
        ./configure --help shows
        
--enable-iscii          iscii[disabled]

        but --enable-ind option , which I wrote in <200...@pd...> ,
        is correct.

        I forgot to fix it.

> How do you I use mlconfig?
> Again, Gtk is not installed in my machine? Is Gtk is necessary in order to
> install mlterm?

        gtk+ is not necessary to build mlterm , but mlconfig , a configuration menu ,
        depends on gtk+.

        mlconfig is started if you press control key button 3.

        Best Regards.
	
--
kiken
j00...@ip...

Re: [Mlterm-dev-en] mlterm ISCII support 2002/01/06

[Jyotirmoy S. <jm...@cs...>]

Hi,

On Sun, 6 Jan 2002, Araki Ken wrote:

:        [tar archive]
:        http://www.geocities.co.jp/SiliconValley-Cupertino/6461/mlterm-2.2.0-pre20020106.tar.gz

I've downloaded it and also installed it. I've run the configure script
using --enable-iscii option.

After installing it, i've run it using the following command:
  mlterm -X=false -V -E iscii -w 16 -C Malayalam

Before installing mlterm, I did install libind.tar.gz, truetype.tar.gz and
xfstt  and also installing indian.h in /usr/lib.

Then I pressed Alt_R for ISCII inputting. But I was not able to see any
Indian language glyphs! Did I skip any step in installing it?

Again, I've seen that -C option for choosing the language doesn't verify
an invalid parameter to it. I've given any arbitray string (like XYZ, abc
etc) to it and saw it working without showing any error message. I think
it should give an error message in case of an invalid parameter; isn't it?

:   3. ISCII languages can be dynamically changed from mlconfig.

How do you I use mlconfig?

Again, Gtk is not installed in my machine? Is Gtk is necessary in order to
install mlterm?

: BTW , I sincerely would like native people's comments whether the design ,
: the implementation or the behavior of mlterm ISCII support is ok or not.

Surely I'd like to do that. :-)

With best regards.

     Jyotirmoy


---------------------------------------------------------------
 Jyotirmoy Saikia | Phone: +91-512-597114 (Hall)
 D312/Hall-IV     |                597608 (CSE lab)
 IIT Kanpur       | E-mail: jm...@cs...
 PIN-208016       |         js...@ya...
 U.P., India      | Homepage: http://www.cse.iitk.ac.in/users/jms
-----------------------------------------------------------------

[Mlterm-dev-en] commit log [2002/01/06]

[Araki K. <j00...@ip...>]

       Hi,

       I committed changes below
       
* ISCII support codes are added. (still unstable)
* INIT_PARSER short cut key is removed and "Full Rest" button is added to mlconfig. 
* when color rgb is explicitly assigned in ~/.mlterm/color , reversed color is not
  drawn correctly under Anti Alias mode. fixed. (thanks to nekoie san)
* ISO-8859-11 can be used as a -E option value.
* ~/.mlterm/vlfont is renamed to ~/.mlterm/vfont
* -V/--varlen/use_variable_length_column option is renamed to 
  -V/--varwidth/use_variable_column_width
* Kochi Mincho sample is added to etc/aafont.

       I committed  ISCII support codes , which are still unstable.
       If you don't need it , do ./configure without --enable-ind option.

       Best Regards.

--
kiken
j00...@ip...

[Mlterm-dev-en] mlterm ISCII support 2002/01/06

[Araki K. <j00...@ip...>]

	Hi,

        Next test release with ISCII support is here =>

        [tar archive]
        http://www.geocities.co.jp/SiliconValley-Cupertino/6461/mlterm-2.2.0-pre20020106.tar.gz
        [patch to cvs current]
        http://www.geocities.co.jp/SiliconValley-Cupertino/6461/mlterm-2.2.0-pre20020102-pre20020106.patch.gz

        COMPLETE LIST
        1. Many minor bugs are fixed , which results in less core dump.
        2. Design is cleaned up. More ISCII related codes are shared with
           Bidi and Arabic shaping codes.
        3. ISCII languages can be dynamically changed from mlconfig.
        4. Each pty window has its own ISCII state.
           This enables you to use different ISCII languages among multiple pty
           windows.

        TODO and BUG LIST
        1. sometimes core dumps around ml_convert_ascii_to_iscii() when multiple
           pty windows are opened.
        2. lines which contains ISCII characters cannot do auto wraparound.
           (this is quite difficult to solve...)

        BTW , I sincerely would like native people's comments whether the design ,
        the implementation or the behavior of mlterm ISCII support is ok or not.
        
        I'm still fumbling my way now...

        Best Regards.
	
--
kiken
j00...@ip...

Re: [Mlterm-dev-en] mlterm ISCII support patch

[Araki K. <j00...@ip...>]

	Hi,

	Subject: [Mlterm-dev-en] mlterm ISCII support patch (was: Re: mlterm ISCII support)
	From: Araki Ken <j00...@ip...>
	Message-ID: <200...@pd...>
	Date: Fri, 04 Jan 2002 16:37:20 +0900

>        Then , I made an ISCII support patch for mlterm.
>        It is still a prototype implementation(unstable!) , but I can see ,
>        edit in command line and copy&paste indian characters like iitm-term.
>        Please test and give me comments.
>
>        http://www.geocities.co.jp/SiliconValley-Cupertino/6461/mlterm-2.2.0-pre20020102-pre20020104.patch.gz
>        http://www.geocities.co.jp/SiliconValley-Cupertino/6461/mlterm-2.2.0-pre20020104.tar.gz

         Attached patch is for bug fix of mlterm-2.2.0-pre20020104.

         1. Roman lang ISCII (-C Roman) is not shown. fixed.
         2. In some cases glyph rendering is corrupt. fixed.

         Regards.
	
--
kiken
j00...@ip...

Index: src/ml_image_line.c
===================================================================
RCS file: /home/ken/cvsroot/mlterm/src/ml_image_line.c,v
retrieving revision 1.50
diff -u -r1.50 ml_image_line.c
--- src/ml_image_line.c	2002/01/04 06:51:44	1.50
+++ src/ml_image_line.c	2002/01/04 22:05:58
@@ -1209,7 +1209,7 @@
 				ml_char_t *  c ;
 				ml_char_t *  comb ;
 				u_int  comb_size ;
-				
+
 				if( prev_font_filled - font_filled > dst_pos)
 				{
 					font_filled = prev_font_filled - dst_pos ;
@@ -1233,10 +1233,25 @@
 					comb_pos = 0 ;
 					while( 1)
 					{
-						ml_char_combine( &dst[dst_pos] ,
+						if( ml_char_is_null( &dst[dst_pos]))
+						{
+							/*
+							 * combining is forbidden if base character is null
+							 */
+							ml_char_copy( &dst[dst_pos] , c) ;
+						}
+						else if( ! ml_char_combine( &dst[dst_pos] ,
 							ml_char_bytes( c) , ml_char_size( c) ,
 							ml_char_font( c) , ml_char_font_decor( c) ,
-							ml_char_fg_color( c) , ml_char_bg_color( c)) ;
+							ml_char_fg_color( c) , ml_char_bg_color( c)))
+						{
+						#ifdef  DEBUG
+							kik_warn_printf( KIK_DEBUG_TAG
+								" combining failed.\n") ;
+						#endif
+						
+							break ;
+						}
 
 						if( comb_pos >= comb_size)
 						{
@@ -1252,10 +1267,23 @@
 
 			if( dst_pos >= 0 && font_filled == prev_font_filled)
 			{
-				ml_char_combine( &dst[dst_pos] ,
+				if( ml_char_is_null( &dst[dst_pos]))
+				{
+					/*
+					 * combining is forbidden if base character is null
+					 */
+					ml_char_copy( &dst[dst_pos] , ch) ;
+				}
+				else if( ! ml_char_combine( &dst[dst_pos] ,
 					ml_char_bytes( ch) , ml_char_size( ch) ,
 					ml_char_font( ch) , ml_char_font_decor( ch) ,
-					ml_char_fg_color( ch) , ml_char_bg_color( ch)) ;
+					ml_char_fg_color( ch) , ml_char_bg_color( ch)))
+				{
+				#ifdef  DEBUG
+					kik_warn_printf( KIK_DEBUG_TAG
+						" combining failed.\n") ;
+				#endif
+				}
 			}
 			else
 			{
@@ -1331,6 +1359,11 @@
 	int  comb_pos ;
 	ml_char_t *  ch ;
 
+	if( ! ml_is_char_combining())
+	{
+		return  0 ;
+	}
+
 	src_len = line->num_of_filled_chars ;
 	if( ( src = ml_str_alloca( src_len)) == NULL)
 	{
@@ -1420,6 +1453,9 @@
 	return  logical_char_index ;
 }
 
+/*
+ * this should be called before ml_imgline_stop_visual_indian()
+ */
 int
 ml_iscii_convert_logical_char_index_to_visual(
 	ml_image_line_t *  line ,
Index: src/ml_iscii.c
===================================================================
RCS file: /home/ken/cvsroot/mlterm/src/ml_iscii.c,v
retrieving revision 1.2
diff -u -r1.2 ml_iscii.c
--- src/ml_iscii.c	2002/01/04 04:02:06	1.2
+++ src/ml_iscii.c	2002/01/04 21:10:34
@@ -20,7 +20,7 @@
 #ifdef  USE_IND
 
 
-/* --- static functions --- */
+/* --- static variables --- */
 
 static char  font_name[256] ;
 static char  font_name_prefix[256] ;
@@ -32,17 +32,31 @@
 static struct a2i_tabl a2i_map[A2IMAXLEN] ;
 static int  a2i_map_size ;
 
+/* used for iitkeyb */
 static char  prev_key[512] ;
 
 
 /* --- global functions --- */
 
 int
-ml_iscii_init(
+ml_iscii_init(void)
+{
+	return  1 ;
+}
+
+int
+ml_iscii_final(void)
+{
+	return  1 ;
+}
+
+int
+ml_iscii_set_lang(
 	char *  default_lang
 	)
 {
 	char *  p ;
+	int  counter ;
 
 	if( ( glyph_map_size = indian_init( glyph_map , default_lang , font_name_prefix , ":")) == -1)
 	{
@@ -54,12 +68,17 @@
 	 * font name is assumed not to be abbriviated.
 	 */
 	p = font_name_prefix ;
-	
-	while( ! isdigit( *p))
+
+	for( counter = 0 ; counter < 7 ; counter ++)
 	{
+		if( ( p = strchr( p , '-')) == NULL)
+		{
+			return  0 ;
+		}
+
 		p ++ ;
 	}
-
+	
 	*(p ++) = '\0' ;
 
 	while( isdigit( *p))
@@ -76,12 +95,6 @@
 	return  1 ;
 }
 
-int
-ml_iscii_final(void)
-{
-	return  1 ;
-}
-
 char *
 ml_iscii_get_font_name(
 	u_int  font_size
@@ -121,12 +134,16 @@
 				keymap = ISCIIKEY_NONE ;
 			}
 		}
-		else if( keymap == ISCIIKEY_IITKEYB)
+		else /* if( keymap == ISCIIKEY_IITKEYB) */
 		{
 			if( ( a2i_map_size = readkeymap( a2i_map , "iitkeyb" , ":")) == -1)
 			{
 				keymap = ISCIIKEY_NONE ;
 			}
+			else
+			{
+				memset( prev_key , 0 , sizeof( prev_key)) ;
+			}
 		}
 	}
 
@@ -166,7 +183,7 @@
 			snprintf( iscii , ascii_len + 1 , "%s" ,
 				ins2iscii( a2i_map , dup , a2i_map_size)) ;
 		}
-		else
+		else /* if( keymap == ISCIIKEY_IITKEYB) */
 		{
 			iitk2iscii( a2i_map , dup , prev_key , a2i_map_size) ;
 			snprintf( iscii , strlen( dup) + 1 , "%s" , dup) ;
@@ -184,19 +201,26 @@
 	return  ascii_len ;
 }
 
+
 #else  /* USE_IND */
 
 
 int
-ml_iscii_init(
-	char *  default_lang
-	)
+ml_iscii_init(void)
 {
 	return  0 ;
 }
 
 int
 ml_iscii_final(void)
+{
+	return  0 ;
+}
+
+int
+ml_iscii_set_lang(
+	char *  default_lang
+	)
 {
 	return  0 ;
 }
Index: src/ml_iscii.h
===================================================================
RCS file: /home/ken/cvsroot/mlterm/src/ml_iscii.h,v
retrieving revision 1.1
diff -u -r1.1 ml_iscii.h
--- src/ml_iscii.h	2002/01/04 03:49:42	1.1
+++ src/ml_iscii.h	2002/01/04 22:08:53
@@ -9,6 +9,13 @@
 #include  <kiklib/kik_types.h>
 
 
+/*
+ * !! Notice !!
+ * ISCII state is shared globally , so you cannot use multiple ISCII langs
+ * at the same time.
+ */
+
+
 typedef enum  ml_iscii_keymap
 {
 	ISCIIKEY_NONE ,
@@ -18,9 +25,11 @@
 } ml_iscii_keymap_t ;
 
 
-int  ml_iscii_init( char *  default_lang) ;
+int  ml_iscii_init(void) ;
 
 int  ml_iscii_final(void) ;
+
+int  ml_iscii_set_lang( char *  default_lang) ;
 
 char *  ml_iscii_get_font_name( u_int  font_size) ;
 
Index: src/ml_shaping.c
===================================================================
RCS file: /home/ken/cvsroot/mlterm/src/ml_shaping.c,v
retrieving revision 1.7
diff -u -r1.7 ml_shaping.c
--- src/ml_shaping.c	2002/01/04 06:51:44	1.7
+++ src/ml_shaping.c	2002/01/04 21:31:42
@@ -4,6 +4,7 @@
 
 #include  "ml_shaping.h"
 
+#include  <stdio.h>		/* fprintf */
 #include  <string.h>		/* strncpy */
 #include  <kiklib/kik_mem.h>	/* alloca */
 
@@ -331,6 +332,24 @@
 				font_filled = ml_iscii_convert_to_font_index(
 						font_buf , dst_len , iscii_buf) ;
 
+			#ifdef  __DEBUG
+				{
+					int  i ;
+					
+					for( i = 0 ; i < iscii_filled ; i ++)
+					{
+						fprintf( stderr , "%.2x " , iscii_buf[i]) ;
+					}
+					fprintf( stderr , "=>\n") ;
+					
+					for( i = 0 ; i < font_filled ; i ++)
+					{
+						fprintf( stderr , "%.2x " , font_buf[i]) ;
+					}
+					fprintf( stderr , "\n") ;
+				}
+			#endif
+				
 				for( counter = 0 ; counter < font_filled ; counter ++)
 				{
 					ml_char_set_bytes( &dst_shaped[counter] , &font_buf[counter] , 1) ;
Index: src/ml_term_manager.c
===================================================================
RCS file: /home/ken/cvsroot/mlterm/src/ml_term_manager.c,v
retrieving revision 1.76
diff -u -r1.76 ml_term_manager.c
--- src/ml_term_manager.c	2002/01/04 03:49:42	1.76
+++ src/ml_term_manager.c	2002/01/04 20:53:47
@@ -1398,16 +1398,18 @@
 		ml_set_word_separators( value) ;
 	}
 
+	ml_iscii_init() ;
+
 	if( ( value = kik_conf_get_value( conf , "iscii_type")))
 	{
-		if( ! ml_iscii_init( value))
+		if( ! ml_iscii_set_lang( value))
 		{
-			ml_iscii_init( "Malayalam") ;
+			ml_iscii_set_lang( "Malayalam") ;
 		}
 	}
 	else
 	{
-		ml_iscii_init( "Malayalam") ;
+		ml_iscii_set_lang( "Malayalam") ;
 	}
 
 	term_man->cmd_path = NULL ;
@@ -1489,6 +1491,7 @@
 #endif
 	ml_keymap_final( &term_man->keymap) ;
 	ml_termcap_final( &term_man->termcap) ;
+	ml_iscii_final() ;
 
 	ml_sig_child_final() ;
 	ml_remove_sig_child_listener( term_man) ;

[Mlterm-dev-en] mlterm ISCII support patch (was: Re: mlterm ISCII support)

[Araki K. <j00...@ip...>]

	Hi,

	Subject: Re: mlterm ISCII support
	From: Jyotirmoy Saikia <jm...@cs...>
	Message-ID: <Pin...@cs...>
	Date: Fri, 4 Jan 2002 01:48:01 -0500 (EST)

> Wish you both a very happy, bright and prosperous new year.

       Happy New Year !

> Now let me tell you some important characterstics of Indian languages.

       Thanks a lot!
       I don't still understand all about ISCII , which is very complicated and
       I began study only a few days ago... , but I know it more:)

       Then , I made an ISCII support patch for mlterm.
       It is still a prototype implementation(unstable!) , but I can see ,
       edit in command line and copy&paste indian characters like iitm-term.
       Please test and give me comments.

       http://www.geocities.co.jp/SiliconValley-Cupertino/6461/mlterm-2.2.0-pre20020102-pre20020104.patch.gz
       http://www.geocities.co.jp/SiliconValley-Cupertino/6461/mlterm-2.2.0-pre20020104.tar.gz

       I prepared screen shots of mlterm and iitm-term , too.

       http://www.geocities.co.jp/SiliconValley-Cupertino/6461/mlterm.png
       http://www.geocities.co.jp/SiliconValley-Cupertino/6461/iitm-term.png
       
       mlterm depends on libind.tar.gz and truetype.tar.gz at
       http://www.tenet.res.in/Donlab/Indlinux/X-Windows/download.html , and xfstt.
       so please install and setup them in advance.

       # Please don't forget to install indian.h to where cc can find(e.g. /usr/local/include).

       If you don't use a tar archive but a gzipped patch , you have to cvs checkout
       cvs-current mlterm as follows.
       
       cvs -d :pserver:ano...@cv...:/cvsroot/mlterm login
       (hit Enter)
       cvs -d :pserver:ano...@cv...:/cvsroot/mlterm checkout mlterm

       after you get it , please do as follows.

       cd mlterm
       gunzip -c mlterm-2.2.0-pre20020102-pre20020104.patch.gz | patch -p1
       ./configure --enable-ind
       make
       make install
       export LIBIND_PATH=/usr/local/libind
       mlterm -X=false -V -E iscii -w 16 -m -C Malayalam

       o -X=false is necessary if you use X Input Method(e.g. kinput2) which conflicts with
         xfstt.(recommended)
       o -V option enables variable column width.(must)
       o -m option enables character combining , which is necessary since mlterm diverts characters
         combining framework for rendering ISCII glyph internally.(must)
       o -C [lang] enables [lang] glyph. Without -C option , Malayalam is used by default.
         mlterm cannot change lang dynamically at present.(optional)
       o -w [font size] assigns font size you want to use.(optional)

       pressing Alt_R lets you use ISCII inputting.
       F1 key swithes Inscript <=> Phonetic keyboard(?).

       BTW , Jyotirmoy , are you involved in devevlopping libind ?
       I have some questions about it , so if I finishes implementing this ,
       I'll send them to ind...@la... , but if you
       are a person concerned , I want to ask you now.

       Best Regards.

--
kiken
j00...@ip...

[Mlterm-dev-en] mlterm 2.1.2 (was: mlterm 2.1.1)

[Araki K. <j00...@ip...>]

	Hi,

	Subject: mlterm 2.1.1 (was: Re: [Mlterm-dev-en] 2.1.0 released)
	From: Araki Ken <j00...@ip...>
	Message-ID: <200...@pd...>
	Date: 30 Dec 2001 08:24:00 +0900

>         I believe there are no more critical bugs...

          Sorry , there is still a bug...

          Please use an attached patch.

          mlterm 2.1.2 is also released at sourceforge web site.

ChangeLog
* selected region is reset in scrolling with dragging mouse. fixed.
          
--
kiken
j00...@ip...

Index: src/ml_term_screen.c
===================================================================
RCS file: /home/ken/cvsroot/mlterm/src/ml_term_screen.c,v
retrieving revision 1.247
diff -u -r1.247 ml_term_screen.c
--- src/ml_term_screen.c	2001/12/29 22:37:17	1.247
+++ src/ml_term_screen.c	2002/01/01 20:00:55
@@ -2420,10 +2420,6 @@
 }
 
 
-/*
- * callbacks of ml_image_scroll events
- */
- 
 static int
 flush_scroll_cache(
 	ml_term_screen_t *  termscr ,
@@ -2535,6 +2531,10 @@
 	termscr->scroll_cache_boundary_end = boundary_end ;
 }
 
+/*
+ * callbacks of ml_bs_event_listener_t events.
+ */
+ 
 static int
 window_scroll_upward(
 	void *  p ,
@@ -2550,8 +2550,6 @@
 		return  0 ;
 	}
 
-	ml_restore_selected_region_color( &termscr->sel) ;
-
 	set_scroll_boundary( termscr , 0 , ml_image_get_rows( termscr->image) - 1) ;
 	
 	termscr->scroll_cache_rows += size ;
@@ -2574,8 +2572,6 @@
 		return  0 ;
 	}
 
-	ml_restore_selected_region_color( &termscr->sel) ;
-	
 	set_scroll_boundary( termscr , 0 , ml_image_get_rows( termscr->image) - 1) ;
 	
 	termscr->scroll_cache_rows -= size ;
@@ -2583,6 +2579,11 @@
 	return  1 ;
 }
 
+
+/*
+ * callbacks of ml_image_scroll events
+ */
+ 
 static int
 window_scroll_upward_region(
 	void *  p ,

mlterm 2.1.1 (was: Re: [Mlterm-dev-en] 2.1.0 released)

[Araki K. <j00...@ip...>]

	Hi,

	Subject: [Mlterm-dev-en] 2.1.0 released
	From: Araki Ken <j00...@ip...>
	Message-ID: <200...@pd...>
	Date: 29 Dec 2001 17:09:57 +0900

>         I released mlterm-2.1.0.

        mlterm-2.1.0 has a critical bug around cursor movement.
        I decided that mlterm-2.1.1 should be released.
        I'm VERY SORRY.

        Please do cvs checkout -r rel-2_1_1 or download
        http://prdownloads.sourceforge.net/mlterm/mlterm-2.1.1.tar.gz

        I believe there are no more critical bugs...

ChangeLog
* cursor doesn't move correctly on full width characters(critical bug!). fixed.
* If scroll commands appear with scroll region changing one by one , the number
  of redrawing screen becomes too much. fixed.
* mistake of mlterm.spec is fixed.
* 2.1.1 is released.

--
kiken
j00...@ip...

[Mlterm-dev-en] 2.1.0 released

[Araki K. <j00...@ip...>]

        Hi,

        I released mlterm-2.1.0.
        
        Please do cvs checkout with "-r rel-0_2_1" option.
        http://prdownloads.sourceforge.net/mlterm/mlterm-2.1.0.tar.gz
        will be also available soon.

mlterm-2.1.0 new features(doc/en/History)
* Big5 and koi8 copy&paste problem is fixed.
* Bidi rendering and Arabic shaping are supported.
* BIG5HKSCS is supported.
* most of vttest items are passed.
* processing encodings of copy&paste text is improved.
* mlconfig is improved.(e.g. it works asynchronous with mlterm itself)
* mlterm icons(doc/icon/) are contributed by Kubota Tomohiro san:)
* XIM processing is improved.
* and many minor fixes/improvement.

--
kiken
j00...@ip...

[Mlterm-dev-en] commit log [2001/12/29] and 2.1.0 release

[Araki K. <j00...@ip...>]

	Hi,

        I committed changes below.
        
* compilation of src/ml_image.c fails in Solaris,HP-UX,EWS4800. fixed.
  (thanks to Sakamoto Hironori san)
* support for the sequences invoking dec-special characters is improved.
* memory leaks when combining chars are selected twice or more. fixed.

        if no problems are reported about this CVS current version ,
        I'll do cvs tag rel-2_1_0 as it is and release mlterm 2.1.0.

        Best Regards.

--
kiken
j00...@ip...

[Mlterm-dev-en] commit log [2001/12/28]

[Araki K. <j00...@ip...>]

	Hi,

        I committed changes below.
	
[20011228]
* INIT_PARSER short cut key (which initializes status of pty encoding parser)
  is added.
* invoking graphic characters by SI/SO is supported.(non-ISO2022 encoding only)

[20011227]
* if pty encoding is stateless , state of pty encoding parser is not
  reset when key is pressed.
* "ESC [ ? 25 h" , "ESC [ ? 25 l" sequences are supported.


--
kiken
j00...@ip...

[Mlterm-dev-en] commit log [2001/12/26]

[Araki K. <j00...@ip...>]

        Hi,

        I committed changes below.

* typos in mlterm.1 are fixed.(thanks to Kubota Tomohiro san)
* something like "ESC [ 34 ; 0 ; m" sequence is parsed incorrectly.fixed.
  (thanks to Sakamoto Hironori san)
* cursor is not repainted under transparent mode. fixed.
  (thanks to Uebayashi Masao san)
  
--
kiken
j00...@ip...

[Mlterm-dev-en] commit log [2001/12/25-2]

[Araki K. <j00...@ip...>]

	Hi,

        I committed changes below.
        
* memory leaks if you use both -m and -bi options under utf8 encoding. fixed.
* mkf/mkf_ucs4_xxx.c are thread-safe.
* Shift + Mouse Button operations under console apps useing mouse tracking
  are supported.

        I'll do cvs tag rel-2_1_0 around December 28th or 29th(JST).
        I would like information of platforms where mlterm can be compiled or not.

        Best Regards.

--
kiken
j00...@ip...

[Mlterm-dev-en] commit log [2001/12/25]

[Araki K. <j00...@ip...>]

	Hi,

        I committed changes below.
        
* wraparound word can be selected.
* mlterm core dumps if XMODIFIERS variable is empty or illegal value. fixed.

--
kiken
j00...@ip...

[Mlterm-dev-en] commit log [2001/12/24]

[Araki K. <j00...@ip...>]

	Hi,

        I committed changes below.

* US-ASCII chars may be shown incorrectly under EUC-CN/GBK/GB18030. fixed.
* mlterm.1 document is updated a little more.(thanks to Kubota Tomohiro san)
* Unicode <=> CNS11643-1992 plane 3 convertion table is added.
* EUC-TW is excluded from Unicode subset group.
* the default value of use_bidi and use_combining is true.
* if mlterm is resized under bidi-mode , you cannot copy&paste RTL chars. fixed.

        I have one question.
        mlterm is now using "*-gbk-0" font for 2byte chars of GB18030.
        
        # FYI , Even if you don't have "*-gbk-0" font , you can show 2byte chars
        # of GB18030 by starting mlterm with -u(--onlyucsfont) option.
        
        But which is more preferable , processing 2byte chars of GB18030 as GBK by
        default or converting them to Unicode by default ?


        BTW , I plan to release 2.1.0 soon(maybe in a week).
        
        Would you please test CVS-current mlterm to find unknown bugs ?
        In addition , if you have anything unknown about mlterm(e.g. options) ,
        please tell it to me. I also want to improve documentations.

        Best Regards,
        
--
kiken
j00...@ip...

[Mlterm-dev-en] commig log[2001/12/22]

[Araki K. <j00...@ip...>]

	Hi,

        First of all , I committed changes below.

* Hangul chars are not shown correctly in --noucsfont. fixed.
* manual page(mlterm.1) is updated. (thanks to Kubota Tomohiro san)
* the same CTEXT generator engine is used whether --big5bug option is set or not. fixed.

	Subject: Re: [Mlterm-dev-en] commig log , BIG5 copy&paste
	From: "Edward G.J. Lee" <ed...@ms...>
	Message-ID: <200...@da...>
	Date: Fri, 21 Dec 2001 23:12:26 +0800

> >         Would you please start mlterm as follows
> > 
> >         $ LC_CTYPE=zh_TW.Big5 mlterm -km big5 --big5bug=false
> > 
> >         and copy and paste from mlterm to rxvt.
> > 
> >         if it is ok , I'll remove --big5bug option.
> 
>   Yes, it's ok! Seems need not the --big5bug option(XFree-4.1.0).

        Thanks a lot for testing:) , but I'm VERY SORRY. This was my SERIOUS MISTAKE.
        I used the same CTEXT generator engine whether --big5bug option or --big5bug=false
        option is used.
        --big5bug option is still necessary ...

--
kiken
j00...@ip...

Re: [Mlterm-dev-en] commig log , BIG5 copy&paste

[Edward G.J. L. <ed...@ms...>]

On Thu, Dec 20, 2001, Araki Ken wrote:
[...]
>         Would you please start mlterm as follows
> 
>         $ LC_CTYPE=zh_TW.Big5 mlterm -km big5 --big5bug=false
> 
>         and copy and paste from mlterm to rxvt.
> 
>         if it is ok , I'll remove --big5bug option.

  Yes, it's ok! Seems need not the --big5bug option(XFree-4.1.0).

  Merry Christmas and happy new year. :-)


-- 
Warm Regards,
Edward G.J. Lee

[Mlterm-dev-en] commit log

[Araki K. <j00...@ip...>]

	Hi,
        
        I committed changes below.
        
* option names are changed.
  --ucs2other => --noucsfont
  unicode_to_other_cs => not_use_unicode_font
  --all2ucs => --onlyucsfont
  all_cs_to_unicode => only_use_unicode_font
* full width chars are not be shown correctly in --onlyucsfont. fixed.
* -8/--88591 (iso88591_font_for_usascii) option is added.
* -R /usr/local/lib -R /usr/X11R6/lib options are added to LIBS macro in
  src/Makefile.in tool/mlconfig/Makefile.in
  (thanks to Uebayashi Masao san)
* illegal chars which is not based on ISO2022 (e.g. KOI8-R) are also accepted if
  at all possible in ISO2022-based encodings.

        I changed quite a few things these days around X Selection processing etc
        according to the discussion in mlterm-dev-ja.
        Kubota Tomohiro san will soon update mlterm.1 document and detail it.

--
kiken
j00...@ip...