You can subscribe to this list here.
2001 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
(1) |
Dec
|
---|---|---|---|---|---|---|---|---|---|---|---|---|
2002 |
Jan
(1) |
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(4) |
Sep
|
Oct
|
Nov
(2) |
Dec
(2) |
2003 |
Jan
(25) |
Feb
(5) |
Mar
(12) |
Apr
(46) |
May
(47) |
Jun
|
Jul
(2) |
Aug
|
Sep
(15) |
Oct
(8) |
Nov
(11) |
Dec
|
2004 |
Jan
(25) |
Feb
(24) |
Mar
(13) |
Apr
(59) |
May
(52) |
Jun
(6) |
Jul
(3) |
Aug
(7) |
Sep
(33) |
Oct
(17) |
Nov
(16) |
Dec
(1) |
2005 |
Jan
(4) |
Feb
|
Mar
(2) |
Apr
(5) |
May
(50) |
Jun
(3) |
Jul
(1) |
Aug
(1) |
Sep
(7) |
Oct
(1) |
Nov
(2) |
Dec
(9) |
2006 |
Jan
(10) |
Feb
(6) |
Mar
(2) |
Apr
(24) |
May
(32) |
Jun
(53) |
Jul
(26) |
Aug
(28) |
Sep
(59) |
Oct
(72) |
Nov
(85) |
Dec
(57) |
2007 |
Jan
(43) |
Feb
(26) |
Mar
(25) |
Apr
(36) |
May
(13) |
Jun
(14) |
Jul
(53) |
Aug
(68) |
Sep
(46) |
Oct
(62) |
Nov
(15) |
Dec
(4) |
2008 |
Jan
(4) |
Feb
(5) |
Mar
(7) |
Apr
(5) |
May
|
Jun
|
Jul
|
Aug
|
Sep
(3) |
Oct
(5) |
Nov
|
Dec
(3) |
2009 |
Jan
(2) |
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
2010 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
(1) |
Sep
|
Oct
|
Nov
(3) |
Dec
|
2011 |
Jan
(1) |
Feb
|
Mar
|
Apr
|
May
|
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: Romain <ro...@li...> - 2004-04-28 06:33:22
|
Hi, a new web-site is being written. You can get a preview at http://tilp.info/test/ The goal of the new web site is to be easier and clearer. Comments, suggestions are welcome... thx, roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Romain <ro...@li...> - 2004-04-27 21:00:12
|
Hi, > Can you put the source tarballs on sf.net too? That's more interesting I put them ... BTW, could you tell me whether SilverLink support is working fine for you= ? roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Romain <ro...@li...> - 2004-04-27 11:44:15
|
Hi, > Can you put the source tarballs on sf.net too? That's more interesting > for package management. Ticalc.org doesn't keep older versions so the > ports (can) brake each time there's a new version. ok. >> This release should be the last one for now. The next one will be >> done to add support of new TI calcs (TI84+, ...). > > Do you happen to know when it will be available in Europe? I discussed this topic whith Kevin (Kofler)... It should be availaible at the summer (August / September). > >> In the meanwhile, I'm gonna to finish TiEmu. > > Maybe you also want to take a look at GnuTI. I don't if it's very > active, but anyway http://gnuti.sf.net/ Thanks ! It seems to be dead... There was also a project to goes on XTige= r but it deads, too. My goal is to improve/finish TiEmu by targetting m68k calc only as the first step. The second one is to add debugging capabilities (maybe for better use wit= h TiGCC). > > > ------------------------------------------------------- > This SF.net email is sponsored by: The Robotic Monkeys at ThinkGeek > For a limited time only, get FREE Ground shipping on all orders of $35 > or more. Hurry up and shop folks, this offer expires April 30th! > http://www.thinkgeek.com/freeshipping/?cpg=12297 > _______________________________________________ > TiLP-devel mailing list > TiL...@li... > https://lists.sourceforge.net/lists/listinfo/tilp-devel > roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Tijl C. <ti...@ul...> - 2004-04-27 10:13:02
|
On Tuesday 27 April 2004 11:32, Romain Li=E9vin wrote: > TiLP v6.72 has been upload today in the ticalc.org archives. Can you put the source tarballs on sf.net too? That's more interesting=20 for package management. Ticalc.org doesn't keep older versions so the=20 ports (can) brake each time there's a new version. > This release should be the last one for now. The next one will be > done to add support of new TI calcs (TI84+, ...). Do you happen to know when it will be available in Europe? > In the meanwhile, I'm gonna to finish TiEmu. Maybe you also want to take a look at GnuTI. I don't if it's very=20 active, but anyway http://gnuti.sf.net/ |
From: Romain <ro...@li...> - 2004-04-27 09:32:07
|
Hi all users ! TiLP v6.72 has been upload today in the ticalc.org archives. --------------[ Win32 ]------------- This release includes one important bug fix/improvement on the SilverLink support: transfers are reliable and fast, now. SilverLink users will have to install the TiglUsb driver v3.0. Windows 98 users: the GTK+ Wimp (Windows IMPersonator) theme engine is disabled to avoid a lot of warnings in the console. --------------[ Linux ]---------------- This release includes one important bug fix/improvement on the SilverLink support: transfers are reliable and fast, now. SilverLink users: you can use built-in support or the kernel module (v1.07b or kernel >=3D 2.6.7 (not released yet)). Parallel/serial link: if you are using a kernel module, upgrade to tipar v1.19 / tiser v1.20. It fixes a timeout issue (especially when doing FLAS= H transfers). ------------------------------------------- This release should be the last one for now. The next one will be done to add support of new TI calcs (TI84+, ...). In the meanwhile, I'm gonna to finish TiEmu. Have fun! roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Romain <ro...@li...> - 2004-04-26 10:05:55
|
Hi, >> It has been replaced by GtkToolbarItem. It's strange because it has no= t GtkToolItem, sorry. >> been moved into the deprecated section in the the online API. > > The online API doc isn't up to date yet, it's still at GTK 2.2. No, it isn't. GTK+ 2.4 stuffs are highlighted by a 'Since 2.4'. In fact, GtkToolbarChild has been removed because this structure is opaqu= e and should not be used by developers. TiLP uses it in an ugly fashion. Now, elements are accesible with gtk_toolbar_get_nth_item () which return= s an GtkToolItem. It should be a couple of minutes to fix this. > > JB. > > -- > Julien BLACHE <http://www.jblache.org= > > <jb...@jb...> GPG KeyID 0xF5D65169 > roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Julien B. <jb...@jb...> - 2004-04-26 09:17:04
|
Romain Li=E9vin <ro...@li...> wrote: > It has been replaced by GtkToolbarItem. It's strange because it has not > been moved into the deprecated section in the the online API. The online API doc isn't up to date yet, it's still at GTK 2.2. JB. --=20 Julien BLACHE <http://www.jblache.org>=20 <jb...@jb...> GPG KeyID 0xF5D65169 |
From: Julien B. <jb...@jb...> - 2004-04-26 08:51:37
|
Tijl Coosemans <ti...@ul...> wrote: > I think this was a problem in 6.67 where a config file was put > into /etc. As far as I can see this has been fixed in 6.68 where the > config files are placed in ${prefix}/etc. Which is still incorrect, as config files should go to $DESTDIR/$sysconfdir. $DESTDIR should be used when building packages, messing with $prefix is just asking for problems. The autotools have had full support for $DESTDIR (including libtool) for the last couple of releases already. JB. -- Julien BLACHE <http://www.jblache.org> <jb...@jb...> GPG KeyID 0xF5D65169 |
From: Romain <ro...@li...> - 2004-04-26 08:49:36
|
Hi, > Just removing the -D worked for me and gtk+2.4. > > The problem is in src/toolbar.c. GtkToolbarChild has been deprecated. It has been replaced by GtkToolbarItem. It's strange because it has not been moved into the deprecated section in the the online API. But, I don't know whether I could rewrite it because I believe that Glade does not support GTK+ 2.4 yet. roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Romain <ro...@li...> - 2004-04-26 08:42:10
|
Hi Tijl, Tijl Coosemans a dit : > On Sunday 25 April 2004 21:35, Julien BLACHE wrote: >> Brent Redeker <bre...@cc...> wrote: >> >>> against Gtk 2.4, which the Gentoo ebuild fixes with this line >>> (which happens right before the ebuild runs the configure command): >>> sed -e 's/GTK_DISABLE_DEPRECATED/GTK_DEPRECATED/g' -i >>> src/Makefile.in >> >> This is the wrong fix for this issue. I guess this does precisely the >> contrary, ie it probably enables deprecated widgets, which is not >> what we want. Better drop the -D entirely, or patch TiLP to not use >> the deprecated widget, but TiLP may not build with GTK 2.2 after >> that. > > Just removing the -D worked for me and gtk+2.4. > > The problem is in src/toolbar.c. GtkToolbarChild has been deprecated. ok, thanks. roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Tijl C. <ti...@ul...> - 2004-04-26 08:29:06
|
On Monday 26 April 2004 01:28, Brent Redeker wrote: > Apparently, the TiLP build process does something with > a 'registry' and in doing so, modifies files outside the build > root. I'm not sure what "registry" the Gentoo people meant - has > something like this been added to TiLP recently? I think this was a problem in 6.67 where a config file was put into /etc. As far as I can see this has been fixed in 6.68 where the config files are placed in ${prefix}/etc. |
From: Tijl C. <ti...@ul...> - 2004-04-26 08:12:08
|
On Sunday 25 April 2004 21:35, Julien BLACHE wrote: > Brent Redeker <bre...@cc...> wrote: > >> against Gtk 2.4, which the Gentoo ebuild fixes with this line >> (which happens right before the ebuild runs the configure command): >> sed -e 's/GTK_DISABLE_DEPRECATED/GTK_DEPRECATED/g' -i >> src/Makefile.in > > This is the wrong fix for this issue. I guess this does precisely the > contrary, ie it probably enables deprecated widgets, which is not > what we want. Better drop the -D entirely, or patch TiLP to not use > the deprecated widget, but TiLP may not build with GTK 2.2 after > that. Just removing the -D worked for me and gtk+2.4. The problem is in src/toolbar.c. GtkToolbarChild has been deprecated. |
From: Romain <ro...@li...> - 2004-04-26 07:18:24
|
HI, TI has just released their new calculator and created a specific site : http://www.84silver.com/ It seems this calc uses USB OTG (On The Go) specification. Will TiLP be able to support it ? Great question ! roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Romain <ro...@li...> - 2004-04-26 07:09:56
|
Hi, >>> against Gtk 2.4, which the Gentoo ebuild fixes with this line (which >>> happens right before the ebuild runs the configure command): >>> sed -e 's/GTK_DISABLE_DEPRECATED/GTK_DEPRECATED/g' -i src/Makefile.in >> >> This is the wrong fix for this issue. I guess this does precisely the >> contrary, ie it probably enables deprecated widgets, which is not what >> we want. Better drop the -D entirely, or patch TiLP to not use the >> deprecated widget, but TiLP may not build with GTK 2.2 after that. at the time being, the -D_DEPRECATED needs to be used with GTK+ 2.4. The best way is to fix that. I will do it asap. >> I fixed this bug (in labels.c) ages ago already. It took me two hours >> to track it down. Moreover, the Gentoo patch is simply completely >> wrong and doesn't do the right thing. Trash it, the author of this I agree. If path is shorter, the label is not refreshed any more. Anyways, I rewritten this part, the code was too buggy and unclear. > Hopefully, I'll be able to get 6.71 ebuild working. If your patch has > been applied to the main TiLP branch, then we can safely drop the > Gentoo "fix". (I'm sorry, I haven't been closely following TiLP bug > fixes for quite some time, so I can only assume a real fix was checked > in since this patch no longer applies cleanly to the 6.71 source) I > will also disable the sed kludge and see what happens - i.e. how badly > Gtk-2.4 breaks things :-) ok >> What's this ".keep" directory that is special-cased in the patch ? >> This patch seems bogus to me. > > AFAIK, this is to work around a Gentoo Portage sandbox violation when > building TiLP. Portage sandboxes the builds for all software, so that > it is only allowed to compile and install to a particular directory. (I > am mostly unfamiliar with FreeBSD ports, so I don't know if ports > enforces a sandbox, or just relies on portfiles being "correct".) > Apparently, the TiLP build process does something with a 'registry' and > in doing so, modifies files outside the build root. I'm not sure what > "registry" the Gentoo people meant - has something like this been added > to TiLP recently? The registry file is read only. The plugin dir (/usr/lib/tilp) is ro, too= . There is just one file TiLP is modifying: the ~/.tilp config file. TiLP should not modify anything else. thx, roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Romain <ro...@li...> - 2004-04-26 06:55:34
|
Hi Brent, > First of all, I'll show you which patches are being applied. I have > included them as attachments (tell me if you prefer me to just put the Like lkml, I will prefer to have them in-line (just quoted by a =3D=3D=3D= =3D=3D[ cut here ]=3D=3D=3D=3D=3D=3D=3D=3D=3D). > text 'in-line' in my email). The formatbug patch does not apply > cleanly to tilp-6.71 (I do not know if this was a fixed bug in 6.71, or > if the source just changed enough so that the patch does not work). This patch should not be needed any more. I completely rewritten and tested this section. There were some bug. I believe you need v6.72 for that. > However, the keepdir patch still works, but I am unsure of its > usefulness. I don't remember: why do you need to check .keep files ? Isn't there any other way to do that ? > issues. There was also a problem related to compiling TiLP against Gtk > 2.4, which the Gentoo ebuild fixes with this line (which happens right > before the ebuild runs the configure command): > sed -e 's/GTK_DISABLE_DEPRECATED/GTK_DEPRECATED/g' -i src/Makefile.in ok, I will take a look on GTK+ 2.4 compat issues on this evening. Tijl an= d you may need it. > > For reference, here are the recent Bugzilla entries for TiLP: > http://bugs.gentoo.org/show_bug.cgi?id=3D40263 "Error opening directory '/usr/lib/tilp': No such file or directory" -> fixed. > http://bugs.gentoo.org/show_bug.cgi?id=3D45599 "upgrading to GTK+2.4. " -> ASAP. > Let me know what you think of this so far, and the I can tell you what > problems are happening with the new 6.71 version of TiLP on Gentoo. > roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Romain <ro...@li...> - 2004-04-26 06:32:02
|
Hi, > It's the registry for the tiffep server, which is still an > experimental thing, and shouldn't be installed. The TiFFEP server is not used any longer from TiLP v4.x. This registry file is used by the right popup menu to match a file extension with an application or a shared object. This is the plugin database. roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Romain <ro...@li...> - 2004-04-26 06:27:12
|
Hi, >> Gentoo "fix". (I'm sorry, I haven't been closely following TiLP bug >> fixes for quite some time, so I can only assume a real fix was checked >> in since this patch no longer applies cleanly to the 6.71 source) I > > Romain modified the function again, I just hope he didn't manage to > break it again... I hope, too. I recently rewritten it and do deep tests. It should be working fine. roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Julien B. <jb...@jb...> - 2004-04-26 06:22:58
|
Brent Redeker <bre...@cc...> wrote: Hi, >> I fixed this bug (in labels.c) ages ago already. It took me two hours >> to track it down. Moreover, the Gentoo patch is simply completely >> wrong and doesn't do the right thing. Trash it, the author of this >> patch didn't bother enough to dig into the code and understand where >> the bug came from, and why it was a bug in the first place. > > Hopefully, I'll be able to get 6.71 ebuild working. If your patch has > been applied to the main TiLP branch, then we can safely drop the I committed it into the trunk (this was in revision 357 and 358). > Gentoo "fix". (I'm sorry, I haven't been closely following TiLP bug > fixes for quite some time, so I can only assume a real fix was checked > in since this patch no longer applies cleanly to the 6.71 source) I Romain modified the function again, I just hope he didn't manage to break it again... > will also disable the sed kludge and see what happens - i.e. how badly > Gtk-2.4 breaks things :-) That would be interesting. >> What's this ".keep" directory that is special-cased in the patch ? >> This patch seems bogus to me. > > AFAIK, this is to work around a Gentoo Portage sandbox violation when > building TiLP. Portage sandboxes the builds for all software, so that I seriously doubt it is a build-time problem, as the patch applies to the code that loads the plugins... > it is only allowed to compile and install to a particular > directory. (I am mostly unfamiliar with FreeBSD ports, so I don't know > if ports enforces a sandbox, or just relies on portfiles being > "correct".) Apparently, the TiLP build process does something with a Don't tell me they're spreading ".keep" directory accross the filesystem to keep some portage-related information in it ? Anyway, if this is the case, the correct fix for this is to have the plugin loading code ignore hidden files and directory, and it is certainly not to add a special case for gentoo stuff. Once again, someone didn't bother enough to do things the Right Way (tm). > 'registry' and in doing so, modifies files outside the build root. I'm > not sure what "registry" the Gentoo people meant - has something like > this been added to TiLP recently? It's the registry for the tiffep server, which is still an experimental thing, and shouldn't be installed. > I'm sorry, I don't know much more. I haven't had time to really try > and look at what these patches are doing. Thanks for your help, though. You're welcome. JB. -- Julien BLACHE <http://www.jblache.org> <jb...@jb...> GPG KeyID 0xF5D65169 |
From: Romain <ro...@li...> - 2004-04-26 06:21:28
|
Hi, Jack Curry a dit : > Whenever I run tilp, and try to transfer something, I get a timeout err= or, > (errno=3D110). I am running Debian sid, kernel 2.6.5, and the latest > version > of Tilp. What is wrong? > Could you give me: - calc model, - link cable, - eventually the log file (you can get it by copying console content). roms. --=20 Romain Li=E9vin : <ro...@li...> Web site : http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Brent R. <bre...@cc...> - 2004-04-25 23:28:09
|
Hi Julien, thanks for looking at this stuff. On Apr 25, 2004, at 2:35 PM, Julien BLACHE wrote: > It wasn't related to a locale issue, but some locales helped trigger > the bug, which was a basic coding mistake because someone forgot THAT > YOU MUST NEVER MODIFY A STRING RETURNED BY A GTK FUNCTION UNLESS THE > DOCUMENTATION EXPLICITLY TELLS THAT YOU CAN. And I bet there a couple > of other similar issues at some other places in TiLP, but I do not > have the time for a complete audit... > >> against Gtk 2.4, which the Gentoo ebuild fixes with this line (which >> happens right before the ebuild runs the configure command): >> sed -e 's/GTK_DISABLE_DEPRECATED/GTK_DEPRECATED/g' -i src/Makefile.in > > This is the wrong fix for this issue. I guess this does precisely the > contrary, ie it probably enables deprecated widgets, which is not what > we want. Better drop the -D entirely, or patch TiLP to not use the > deprecated widget, but TiLP may not build with GTK 2.2 after that. > > Without a build log, I can't say much more. > >> *tilp-6.68-r2 (13 Apr 2004) >> >> 13 Apr 2004; Patrick Kursawe <pho...@ge...> >> tilp-6.68-r1.ebuild, >> tilp-6.68-r2.ebuild, files/formatbug.patch: >> Traced bug #40263 to a buffer overflow, provided patch. > > I fixed this bug (in labels.c) ages ago already. It took me two hours > to track it down. Moreover, the Gentoo patch is simply completely > wrong and doesn't do the right thing. Trash it, the author of this > patch didn't bother enough to dig into the code and understand where > the bug came from, and why it was a bug in the first place. Hopefully, I'll be able to get 6.71 ebuild working. If your patch has been applied to the main TiLP branch, then we can safely drop the Gentoo "fix". (I'm sorry, I haven't been closely following TiLP bug fixes for quite some time, so I can only assume a real fix was checked in since this patch no longer applies cleanly to the 6.71 source) I will also disable the sed kludge and see what happens - i.e. how badly Gtk-2.4 breaks things :-) > >> *tilp-6.68-r1 (29 Mar 2004) >> >> 29 Mar 2004; Patrick Kursawe <pho...@ge...> >> tilp-6.68-r1.ebuild, >> files/6.68-keepdir.patch: >> Some changes because of bug #40263 > > What's this ".keep" directory that is special-cased in the patch ? > This patch seems bogus to me. AFAIK, this is to work around a Gentoo Portage sandbox violation when building TiLP. Portage sandboxes the builds for all software, so that it is only allowed to compile and install to a particular directory. (I am mostly unfamiliar with FreeBSD ports, so I don't know if ports enforces a sandbox, or just relies on portfiles being "correct".) Apparently, the TiLP build process does something with a 'registry' and in doing so, modifies files outside the build root. I'm not sure what "registry" the Gentoo people meant - has something like this been added to TiLP recently? I'm sorry, I don't know much more. I haven't had time to really try and look at what these patches are doing. Thanks for your help, though. Brent |
From: Julien B. <jb...@jb...> - 2004-04-25 19:36:13
|
Brent Redeker <bre...@cc...> wrote: Hi, > It seems both patches were a result of one bug in Gentoo's Bugzilla. I > think the keepdir patch is to fix problems with the Makefile > attempting to modify stuff outside of the PREFIX that was set for > it. And I think that the other patch was to fix some segfaults related > to locale issues. There was also a problem related to compiling TiLP It wasn't related to a locale issue, but some locales helped trigger the bug, which was a basic coding mistake because someone forgot THAT YOU MUST NEVER MODIFY A STRING RETURNED BY A GTK FUNCTION UNLESS THE DOCUMENTATION EXPLICITLY TELLS THAT YOU CAN. And I bet there a couple of other similar issues at some other places in TiLP, but I do not have the time for a complete audit... > against Gtk 2.4, which the Gentoo ebuild fixes with this line (which > happens right before the ebuild runs the configure command): > sed -e 's/GTK_DISABLE_DEPRECATED/GTK_DEPRECATED/g' -i src/Makefile.in This is the wrong fix for this issue. I guess this does precisely the contrary, ie it probably enables deprecated widgets, which is not what we want. Better drop the -D entirely, or patch TiLP to not use the deprecated widget, but TiLP may not build with GTK 2.2 after that. Without a build log, I can't say much more. > *tilp-6.68-r2 (13 Apr 2004) > > 13 Apr 2004; Patrick Kursawe <pho...@ge...> > tilp-6.68-r1.ebuild, > tilp-6.68-r2.ebuild, files/formatbug.patch: > Traced bug #40263 to a buffer overflow, provided patch. I fixed this bug (in labels.c) ages ago already. It took me two hours to track it down. Moreover, the Gentoo patch is simply completely wrong and doesn't do the right thing. Trash it, the author of this patch didn't bother enough to dig into the code and understand where the bug came from, and why it was a bug in the first place. > *tilp-6.68-r1 (29 Mar 2004) > > 29 Mar 2004; Patrick Kursawe <pho...@ge...> > tilp-6.68-r1.ebuild, > files/6.68-keepdir.patch: > Some changes because of bug #40263 What's this ".keep" directory that is special-cased in the patch ? This patch seems bogus to me. JB. -- Julien BLACHE <http://www.jblache.org> <jb...@jb...> GPG KeyID 0xF5D65169 |
From: Brent R. <bre...@cc...> - 2004-04-25 18:18:44
|
Hi roms, I'm starting a new thread to address some of the stuff related to Gentoo Linux that I mentioned in my earlier email. First of all, I'll show you which patches are being applied. I have included them as attachments (tell me if you prefer me to just put the text 'in-line' in my email). The formatbug patch does not apply cleanly to tilp-6.71 (I do not know if this was a fixed bug in 6.71, or if the source just changed enough so that the patch does not work). However, the keepdir patch still works, but I am unsure of its usefulness. It seems both patches were a result of one bug in Gentoo's Bugzilla. I think the keepdir patch is to fix problems with the Makefile attempting to modify stuff outside of the PREFIX that was set for it. And I think that the other patch was to fix some segfaults related to locale issues. There was also a problem related to compiling TiLP against Gtk 2.4, which the Gentoo ebuild fixes with this line (which happens right before the ebuild runs the configure command): sed -e 's/GTK_DISABLE_DEPRECATED/GTK_DEPRECATED/g' -i src/Makefile.in For reference, here are the recent Bugzilla entries for TiLP: http://bugs.gentoo.org/show_bug.cgi?id=40263 http://bugs.gentoo.org/show_bug.cgi?id=45599 And here is the appropriate part of the ebuild Changelog: *tilp-6.68-r2 (13 Apr 2004) 13 Apr 2004; Patrick Kursawe <pho...@ge...> tilp-6.68-r1.ebuild, tilp-6.68-r2.ebuild, files/formatbug.patch: Traced bug #40263 to a buffer overflow, provided patch. 31 Mar 2004; Patrick Kursawe <pho...@ge...> tilp-6.06.ebuild, tilp-6.68.ebuild: Cleanup 30 Mar 2004; Patrick Kursawe <pho...@ge...> tilp-6.68-r1.ebuild: Fixing for gtk 2.4, see bug #45599 *tilp-6.68-r1 (29 Mar 2004) 29 Mar 2004; Patrick Kursawe <pho...@ge...> tilp-6.68-r1.ebuild, files/6.68-keepdir.patch: Some changes because of bug #40263 Let me know what you think of this so far, and the I can tell you what problems are happening with the new 6.71 version of TiLP on Gentoo. Thanks for your time, Brent Redeker |
From: <ro...@li...> - 2004-04-25 13:14:27
|
Hi Ben, Benjamin Moody a =E9crit : > The problem occurs when a file both has no extension and has a valid > TI header. The file I had was a fragment of an OS upgrade, starting > with **TIFL**, but it could have been any TI file and have the same > effect. >=20 > tifiles_is_a_tib_file() calls tifiles_is_a_ti_file() to see if there > is a header, finds it present, then checks the extension, at which > point it crashes since there was no extension. Yes, you're right. Patch merged ! >=20 > Thanks again, > Benjamin Moody >=20 >=20 Thanks a _lot_, Romain. --=20 Romain Li=E9vin (roms): <ro...@ti...> Web site: http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |
From: Benjamin M. <ben...@AR...> - 2004-04-25 00:59:53
|
The problem occurs when a file both has no extension and has a valid TI header. The file I had was a fragment of an OS upgrade, starting with **TIFL**, but it could have been any TI file and have the same effect. tifiles_is_a_tib_file() calls tifiles_is_a_ti_file() to see if there is a header, finds it present, then checks the extension, at which point it crashes since there was no extension. Thanks again, Benjamin Moody |
From: <ro...@li...> - 2004-04-24 13:23:24
|
Hi, > Here's the problem: TiLP 6.71 segfaults whenever it encounters a file > *that might be a valid TI file, but has no extension.* Could you tell me how to reproduce the problem ? I attempted to create a=20 'foo' file but TiLP did not crash. >=20 > In this case, the function tifiles_get_extension() returns NULL, and > tifiles_is_a_tib_file() then proceeds to compare NULL with "tib." I > guess this counts as a bug in libtifiles, but in fact it was there in > previous versions of the library. >=20 > The difference is that in previous versions of TiLP, > tifiles_is_a_xxx_file (for xxx =3D group, backup, flash, tib) was only > ever called after the extension had already been checked. In version > 6.71, tifiles_is_a_tib_file is called for every file in the directory, > as part of clist_refresh(). >=20 > So my solution is to add the check to each of the four > tifiles_is_a_xxx_file functions: if there is no extension, the file > isn't valid no matter what it looks like. >=20 > Here's a patch, hoping it works... Thanks. I will apply it as soon as I would have able to reproduce bug.=20 To be sure ! >=20 > diff -u libtifiles-0.5.8/src/typesxx.c libtifiles-0.5.8-patch/src/types= xx.c > --- libtifiles-0.5.8/src/typesxx.c 2004-04-17 06:55:45.000000000 -0400 > +++ libtifiles-0.5.8-patch/src/typesxx.c 2004-04-22 20:16:23.000000000 = -0400 > @@ -788,6 +788,9 @@ > int i; > char *e =3D tifiles_get_extension(filename); > =20 > + if (e=3D=3DNULL) > + return 0; > + > if (!tifiles_is_a_ti_file(filename)) > return 0; > =20 > @@ -823,6 +826,9 @@ > int i; > char *e =3D tifiles_get_extension(filename); > =20 > + if (e=3D=3DNULL) > + return 0; > + > if (!tifiles_is_a_ti_file(filename)) > return 0; > =20 > @@ -844,6 +850,9 @@ > int i; > char *e =3D tifiles_get_extension(filename); > =20 > + if (e=3D=3DNULL) > + return 0; > + > if (!tifiles_is_a_ti_file(filename)) > return 0; > =20 > @@ -865,6 +874,9 @@ > { > char *e =3D tifiles_get_extension(filename); > =20 > + if (e=3D=3DNULL) > + return 0; > + > if (!tifiles_is_a_ti_file(filename)) > return 0; > =20 >=20 > Benjamin Moody >=20 >=20 thx, roms. --=20 Romain Li=E9vin (roms): <ro...@ti...> Web site: http://www.lievin.net "Linux, y'a moins bien mais c'est plus cher !" |