Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
Le mardi 3 octobre 2006 07:09, Kapil Hari Paranjape a écrit : However, I am still unable to get *any* reasonable output from this input for your version of par. There may be something wrong in my settings so I am sending you my environment variables and the output of par jez.out kapil.out. Your variable PARINIT contains 'rTbqgR B=.,?_A_a Q=_s|'. If you remove option 'g', it work well. Nevertheless, option 'g' shouldn't produce this. It is a bug. I try to correct it today. -- Jérôme Pouiller (Jezz) jezz AT sysmic DOT org
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
Hello, On Tue, 03 Oct 2006, Jérôme Pouiller wrote: Le mardi 3 octobre 2006 07:09, Kapil Hari Paranjape a écrit : However, I am still unable to get *any* reasonable output from this input for your version of par. There may be something wrong in my settings so I am sending you my environment variables and the output of par jez.out kapil.out. Your variable PARINIT contains 'rTbqgR B=.,?_A_a Q=_s|'. If you remove option 'g', it work well. Nevertheless, option 'g' shouldn't produce this. It is a bug. I try to correct it today. Thanks for this. Yes. It works now. As you say the problem is with the character tests in reformat.c. These are the character tests (like isalnum) for single-width characters. We need to include the character tests for wide characters (iswalnum). Regards, Kapil. -- signature.asc Description: Digital signature
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
Le mardi 3 octobre 2006 11:02, Kapil Hari Paranjape a écrit : On Tue, 03 Oct 2006, Jérôme Pouiller wrote: [...] Nevertheless, option 'g' shouldn't produce this. It is a bug. I try to correct it today. Thanks for this. Yes. It works now. As you say the problem is with the character tests in reformat.c. These are the character tests (like isalnum) for single-width characters. We need to include the character tests for wide characters (iswalnum). I havn't notice any use of functions for single-width characters. But, during processing of option 'g', I didn't updade correctly width of words. It is now fixed. I have fiwed warning during compilation too. You can find all files par_1.52-i18n.3* on http://sysmic.org/par/debian. bye, -- Jérôme Pouiller (Jezz) jezz AT sysmic DOT org
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
Le samedi 16 septembre 2006 03:22, Kapil Hari Paranjape a écrit : On Fri, 15 Sep 2006, Jérôme Pouiller wrote: Le Ven 15 septembre 2006 9:14, Kapil Hari Paranjape a écrit : [...] 2. From the 'par' built by you as well as the one built by me I am unable to get the 'correct' output for the sample file that was submit by the original submitter of the bug report #310495. Is there some specific environment variable that needs to be set or unset? Yes, your current charset ñust be identic to charset of input text. Something like export LANG=en_IE.UTF-8 should make output correct. My current LANG environment variable is en_IN.UTF-8. I would be grateful if you would download the text given by the original submitter in the bug report and check that your version of par manages to format it correctly. I haven't been able to do that here so far ... Actually, my result is different from result submitted in bug report. I send in attachement my result (mine.out), text given in bug report (your.out) and diff between them (diff). Have you same results than me? Results are different but, I don't think your.out is better than mine.out. What do you think about? -- Jérôme Pouiller (Jezz) jezz AT sysmic DOT org http://sysmic.org 1,2c1,2 Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen --- Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen 5c5 kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä --- kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä 9,10c9,10 par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä --- par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä 15c15 kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä --- kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä 17c17 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, --- -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
Hello, On Mon, 02 Oct 2006, Jérôme Pouiller wrote: Le samedi 16 septembre 2006 03:22, Kapil Hari Paranjape a écrit : I would be grateful if you would download the text given by the original submitter in the bug report and check that your version of par manages to format it correctly. I haven't been able to do that here so far ... Actually, my result is different from result submitted in bug report. I send in attachement my result (mine.out), text given in bug report (your.out) and diff between them (diff). Have you same results than me? Results are different but, I don't think your.out is better than mine.out. What do you think about? I agree that your output and the submitters do not seem to differ significantly. However, I am still unable to get *any* reasonable output from this input for your version of par. There may be something wrong in my settings so I am sending you my environment variables and the output of par jez.out kapil.out. Any help will be greatly appreciated. Thanks and regards, Kapil. -- Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan,ü¨ük ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäYü¯ü³¸ü¸ü¼¸ü°üü°ü´üü°ü¸üü°ü¼ü ü °ü virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan,üük ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäY?ü¸ inen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan,üük ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäYü´ ü¼ inen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan,ü?k ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäYü¸ ?inen virke, jonka avulla testaan, kuinka ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. Tässä tyhjänpäiväinen virke, jonka avulla testaan, ääkköset ja UTF-8 -koodaus toimivat par-työkalun kanssa. BASH=/bin/bash COLUMNS=113 IFS=$' \t\n' LANG=en_GB.UTF-8 LANGUAGE=en_IN:en_US:en_GB:en LC_ADDRESS=en_GB.UTF-8 LC_ALL=en_GB.UTF-8 LC_COLLATE=en_GB.UTF-8 LC_CTYPE=en_GB.UTF-8 LC_IDENTIFICATION=en_GB.UTF-8 LC_MEASUREMENT=en_GB.UTF-8 LC_MESSAGES=en_GB.UTF-8 LC_MONETARY=en_GB.UTF-8 LC_NAME=en_GB.UTF-8 LC_NUMERIC=en_GB.UTF-8 LC_PAPER=en_GB.UTF-8 LC_TELEPHONE=en_GB.UTF-8 LC_TIME=en_GB.UTF-8 LINES=43 PARINIT='rTbgqR B=.,?_A_a Q=_s|' SHELL=/bin/bash SHELLOPTS=braceexpand:emacs:hashall:histexpand:history:interactive-comments:monitor linux-gate.so.1 = (0xe000) libc.so.6 = /lib/tls/i686/cmov/libc.so.6 (0xa7e2b000) /lib/ld-linux.so.2 (0xa7f66000) signature.asc Description: Digital signature
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
On Fri, 15 Sep 2006, Jérôme Pouiller wrote: Le Ven 15 septembre 2006 9:14, Kapil Hari Paranjape a écrit : [...] 2. From the 'par' built by you as well as the one built by me I am unable to get the 'correct' output for the sample file that was submit by the original submitter of the bug report #310495. Is there some specific environment variable that needs to be set or unset? Yes, your current charset ñust be identic to charset of input text. Something like export LANG=en_IE.UTF-8 should make output correct. My current LANG environment variable is en_IN.UTF-8. I would be grateful if you would download the text given by the original submitter in the bug report and check that your version of par manages to format it correctly. I haven't been able to do that here so far ... Thanks and best regards, Kapil. -- signature.asc Description: Digital signature
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
Le Sam 16 septembre 2006 3:22, Kapil Hari Paranjape a écrit : On Fri, 15 Sep 2006, Jérôme Pouiller wrote: Le Ven 15 septembre 2006 9:14, Kapil Hari Paranjape a écrit : [...] 2. From the 'par' built by you as well as the one built by me I am unable to get the 'correct' output for the sample file that was submit by the original submitter of the bug report #310495. Is there some specific environment variable that needs to be set or unset? Yes, your current charset ñust be identic to charset of input text. Something like export LANG=en_IE.UTF-8 should make output correct. My current LANG environment variable is en_IN.UTF-8. I would be grateful if you would download the text given by the original submitter in the bug report and check that your version of par manages to format it correctly. I haven't been able to do that here so far ... Strange... I am currently in holidays. I will test this behaviour as soon as I will come back at home (1st October). best regards, -- Jérôme Pouiller (Jezz) jerome DOT pouiller AT sysmic DOT org
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
Dear Jérôme, Thanks for your effort and sorry for the long silence. On Thu, 03 Aug 2006, Jérôme Pouiller wrote: Le mardi 1 août 2006 15:30, Jérôme Pouiller a écrit : ok, it was a stupid bug: In par.c:719: wchar_t ch; should be char ch; this bug not appear in -O0. I have updated patch (http://sysmic.org/par/par-1.52-i18n.2.diff.gz), package (http://sysmic.org/par/debian/par_1.52-i18n.2_i386.deb), and all other stuffs (http://sysmic.org/par) PS : As suggested by some Debian mainteners, I have changed structure of my archive: * par_1.52.orig.tar.gz is original from Adam M. Costello * par-1.52-i18n.2.diff.gz is my patch * par-1.52-i18n.2.tgz is par_1.52.orig.tar.gz patched with par-1.52-i18n.2.diff.gz * debian/par_1.52.orig.tar.gz is a symbolic link on par_1.52.orig.tgz * debian/par_1.52-i18n.2.diff.gz is patch to apply to build debian package. This patch contains all debian stuff and my patch in debian/patches * debian/par_1.52-i18n.2_i386.deb is built package I downloaded this package. I also downloaded the patch applied and built the package. I am planning to upload the package with this fix. 1. I get some warnings about 'swprintf' not defined while building the package. I could not fix these warnings. 2. From the 'par' built by you as well as the one built by me I am unable to get the 'correct' output for the sample file that was submit by the original submitter of the bug report #310495. Is there some specific environment variable that needs to be set or unset? Thanks and best regards, Kapil. -- signature.asc Description: Digital signature
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
Le mardi 1 août 2006 15:30, Jérôme Pouiller a écrit : Le vendredi 28 juillet 2006 13:04, Mohammed Adnène Trojette a écrit : On Mon, Jul 24, 2006, Mohammed Adnène Trojette wrote: It works well for me too. Great! I don't know what was the problem, but this version seems OK to me. It seems I was too fast. It still freeze when I launch it in vim. It works well on command line. I confirm this bug: $ par 50 EOF bla bla EOF bla bla ^C $ par version par 1.52-i18n.1 I thinks it is an error with use of strtok(). It shouldn't be very hard to resolve. I think, I can solve it before end of week. ok, it was a stupid bug: In par.c:719: wchar_t ch; should be char ch; this bug not appear in -O0. I have updated patch (http://sysmic.org/par/par-1.52-i18n.2.diff.gz), package (http://sysmic.org/par/debian/par_1.52-i18n.2_i386.deb), and all other stuffs (http://sysmic.org/par) PS : As suggested by some Debian mainteners, I have changed structure of my archive: * par_1.52.orig.tar.gz is original from Adam M. Costello * par-1.52-i18n.2.diff.gz is my patch * par-1.52-i18n.2.tgz is par_1.52.orig.tar.gz patched with par-1.52-i18n.2.diff.gz * debian/par_1.52.orig.tar.gz is a symbolic link on par_1.52.orig.tgz * debian/par_1.52-i18n.2.diff.gz is patch to apply to build debian package. This patch contains all debian stuff and my patch in debian/patches * debian/par_1.52-i18n.2_i386.deb is built package -- Jérôme Pouiller (jerome DOT pouiller AT sysmic DOT org)
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
Le vendredi 28 juillet 2006 13:04, Mohammed Adnène Trojette a écrit : On Mon, Jul 24, 2006, Mohammed Adnène Trojette wrote: It works well for me too. Great! I don't know what was the problem, but this version seems OK to me. It seems I was too fast. It still freeze when I launch it in vim. It works well on command line. I confirm this bug: $ par 50 EOF bla bla EOF bla bla ^C $ par version par 1.52-i18n.1 I thinks it is an error with use of strtok(). It shouldn't be very hard to resolve. I think, I can solve it before end of week. -- Jérôme Pouiller (Jezz) jerome DOT pouiller AT sysmic DOT org
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
On Mon, Jul 24, 2006, Mohammed Adnène Trojette wrote: It works well for me too. Great! I don't know what was the problem, but this version seems OK to me. It seems I was too fast. It still freeze when I launch it in vim. It works well on command line. -- adn Mohammed Adnène Trojette
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
Le mardi 11 juillet 2006 17:57, Mohammed Adnène Trojette a écrit : On Tue, Jul 11, 2006, Mohammed Adnène Trojette wrote: Please could upload a patched version of par in experimental, even if upstream has not integrated the patch yet? Just for information, with par rebuilt with your patch, par did not work any longer. Sorry for my belated answer, I was a little busy these days. How did you build it? Can you send me an exemple? What is your $LANG value? What is the behaviour? Is it crash? Version I have packaged (http://www.sysmic.org/par/par_1.52-i18n.1_i386.deb) seems work well. -- Jérôme Pouiller (jerome DOT pouiller AT sysmic DOT org)
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
On Mon, Jul 24, 2006, Jérôme Pouiller wrote: How did you build it? wget, apt-get source, patch, pdebuild, the classics ;) Can you send me an exemple? Well, I don't have the package anymore. What is your $LANG value? % echo $LANG fr_FR.UTF-8 What is the behaviour? Is it crash? It froze. Version I have packaged (http://www.sysmic.org/par/par_1.52-i18n.1_i386.deb) seems work well. It works well for me too. Great! I don't know what was the problem, but this version seems OK to me. Thanks, -- adn Mohammed Adnène Trojette
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
On Fri, Jun 23, 2006, Kapil Hari Paranjape wrote: - Forwarded message from Adam M. Costello [EMAIL PROTECTED] - Date: Fri, 23 Jun 2006 05:37:59 + From: Adam M. Costello [EMAIL PROTECTED] To: Jezz [EMAIL PROTECTED] Subject: Re: [PATCH] Add multibyte support Jezz [EMAIL PROTECTED] wrote: I hope this patch could be added to par debian package. If it works and has no effect on the observable behavior for ASCII input, then I have no objection, just a reminder: You need to change the version numbers in the code and documentation (preferably to something containing at least one letter), and add an entry to the releasenotes file. As for incorporating the patch upstream, I think I first want to fix a minor problem with 8-bit support. (Par 1.51 and earlier was written for 7-bit ASCII. Par 1.52 claimed to add support for 8-bit charsets, but didn't get it completely right.) Also, when I add multibyte support, I might want to add awareness of double-width characters at the same time (here I mean the width on the display, not in memory). AMC - End forwarded message - Hi, and first thanks for your work on par! Please could upload a patched version of par in experimental, even if upstream has not integrated the patch yet? That would be better to report bugs against it than external packages. Thanks for considering, -- adn Mohammed Adnène Trojette
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
On Tue, Jul 11, 2006, Mohammed Adnène Trojette wrote: Please could upload a patched version of par in experimental, even if upstream has not integrated the patch yet? Just for information, with par rebuilt with your patch, par did not work any longer. -- adn Mohammed Adnène Trojette
Bug#310495: [EMAIL PROTECTED]: Re: [PATCH] Add multibyte support]
- Forwarded message from Adam M. Costello [EMAIL PROTECTED] - Date: Fri, 23 Jun 2006 05:37:59 + From: Adam M. Costello [EMAIL PROTECTED] To: Jezz [EMAIL PROTECTED] Subject: Re: [PATCH] Add multibyte support Jezz [EMAIL PROTECTED] wrote: I hope this patch could be added to par debian package. If it works and has no effect on the observable behavior for ASCII input, then I have no objection, just a reminder: You need to change the version numbers in the code and documentation (preferably to something containing at least one letter), and add an entry to the releasenotes file. As for incorporating the patch upstream, I think I first want to fix a minor problem with 8-bit support. (Par 1.51 and earlier was written for 7-bit ASCII. Par 1.52 claimed to add support for 8-bit charsets, but didn't get it completely right.) Also, when I add multibyte support, I might want to add awareness of double-width characters at the same time (here I mean the width on the display, not in memory). AMC - End forwarded message - signature.asc Description: Digital signature