Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Control: tags -1 + buster-ignore Hello, Holger Wansing, le jeu. 20 juin 2019 10:31:39 +0200, a ecrit: > ButterflyOfFire wrote: > > I think it is okay for those diacritics. > > I have uploaded that fix now. Thanks! > Leaving this bug open as a reminder for the "real" problem inside of gtk (?) Tagging accordingly. To gtk+2.0 maintainers: the debian installer is getting hung inside the graphical debconf (and not in the textual debconf) when trying to make gtk+2.0 display "لاحقاً" and we have replaced it with "لاحقا" as a workaround for Buster, but we'll have to fix this bug eventually. Samuel
Processed: Re: Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Processing control commands: > tags -1 + buster-ignore Bug #929877 [installation-reports] installation-reports: Buster installer hangs at hard disk step with arabic language Added tag(s) buster-ignore. -- 929877: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=929877 Debian Bug Tracking System Contact ow...@bugs.debian.org with problems
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Hi, ButterflyOfFire wrote: > Hi, > I think it is okay for those diacritics. > Regards, I have uploaded that fix now. Leaving this bug open as a reminder for the "real" problem inside of gtk (?) Holger > ‐‐‐ Original Message ‐‐‐ > Le mardi 18 juin 2019 22:26, Holger Wansing a écrit : > > > Hi, > > > > Samuel Thibault sthiba...@debian.org wrote: > > > > > Hello, > > > Holger Wansing, le mar. 04 juin 2019 20:59:12 +, a ecrit: > > > > > > > Am Sonntag, 2. Juni 2019 schrieb Holger Wansing: > > > > > > > > > Am Sonntag, 2. Juni 2019 schrieb Samuel Thibault: > > > > > > > > > > > ButterflyOfFire, le dim. 02 juin 2019 15:43:41 +, a ecrit: > > > > > > > > > > > > > > > "لاحقاً" > > > > > > > > > > > > > > This word means "later" and can be replaced by "لاحقا". > > > > > > > > > > > > That avoids the issue here indeed. I however see it used in various > > > > > > > > > > Hmm. > > > > > There has been no changing in the ar.po of partman-auto > > > > > for 3 years now (JFTR), thus the real problem seems to be > > > > > sitting somewhere else ... > > > > > > > > How should we proceed here? > > > > Since we are late in the freeze, and it's most probably not that > > > > easy to find out what happened and what the real issue is: > > > > should we go with the "work-around" proposed by > > > > Butterflyoffire and confirmed to fix the issue by Samuel? > > > > > > Personally, I don't think I'll have the time to debug what is actually > > > happening inside gtk until the release, so even if it's ugly, I guess > > > the browntape fix is the least worst I can come up with. > > > > That looks fair. > > I have committed the proposed fix now. > > > > @butterflyoffire: could you please take a look at > > https://salsa.debian.org/installer-team/d-i/commit/fa17b7afffb2ee4888a371efa4153c5c075915e7 > > to double-check if the fix is fine by you? > > Thanks! > > > > Holger > > > > > > - > > > > Holger Wansing hwans...@mailbox.org > > PGP-Fingerprint: 496A C6E8 1442 4B34 8508 3529 59F1 87CA 156E B076 > > -- Holger Wansing PGP-Fingerprint: 496A C6E8 1442 4B34 8508 3529 59F1 87CA 156E B076
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Hi, I think it is okay for those diacritics. Regards, ‐‐‐ Original Message ‐‐‐ Le mardi 18 juin 2019 22:26, Holger Wansing a écrit : > Hi, > > Samuel Thibault sthiba...@debian.org wrote: > > > Hello, > > Holger Wansing, le mar. 04 juin 2019 20:59:12 +, a ecrit: > > > > > Am Sonntag, 2. Juni 2019 schrieb Holger Wansing: > > > > > > > Am Sonntag, 2. Juni 2019 schrieb Samuel Thibault: > > > > > > > > > ButterflyOfFire, le dim. 02 juin 2019 15:43:41 +, a ecrit: > > > > > > > > > > > > > "لاحقاً" > > > > > > > > > > > > This word means "later" and can be replaced by "لاحقا". > > > > > > > > > > That avoids the issue here indeed. I however see it used in various > > > > > > > > Hmm. > > > > There has been no changing in the ar.po of partman-auto > > > > for 3 years now (JFTR), thus the real problem seems to be > > > > sitting somewhere else ... > > > > > > How should we proceed here? > > > Since we are late in the freeze, and it's most probably not that > > > easy to find out what happened and what the real issue is: > > > should we go with the "work-around" proposed by > > > Butterflyoffire and confirmed to fix the issue by Samuel? > > > > Personally, I don't think I'll have the time to debug what is actually > > happening inside gtk until the release, so even if it's ugly, I guess > > the browntape fix is the least worst I can come up with. > > That looks fair. > I have committed the proposed fix now. > > @butterflyoffire: could you please take a look at > https://salsa.debian.org/installer-team/d-i/commit/fa17b7afffb2ee4888a371efa4153c5c075915e7 > to double-check if the fix is fine by you? > Thanks! > > Holger > > > - > > Holger Wansing hwans...@mailbox.org > PGP-Fingerprint: 496A C6E8 1442 4B34 8508 3529 59F1 87CA 156E B076
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Hi, Samuel Thibault wrote: > Hello, > > Holger Wansing, le mar. 04 juin 2019 20:59:12 +, a ecrit: > > Am Sonntag, 2. Juni 2019 schrieb Holger Wansing: > > > Am Sonntag, 2. Juni 2019 schrieb Samuel Thibault: > > > > ButterflyOfFire, le dim. 02 juin 2019 15:43:41 +, a ecrit: > > > > > >> "لاحقاً" > > > > > > > > > > This word means "later" and can be replaced by "لاحقا". > > > > > > > > That avoids the issue here indeed. I however see it used in various > > > > > > Hmm. > > > There has been no changing in the ar.po of partman-auto > > > for 3 years now (JFTR), thus the real problem seems to be > > > sitting somewhere else ... > > > > How should we proceed here? > > Since we are late in the freeze, and it's most probably not that > > easy to find out what happened and what the real issue is: > > should we go with the "work-around" proposed by > > Butterflyoffire and confirmed to fix the issue by Samuel? > > Personally, I don't think I'll have the time to debug what is actually > happening inside gtk until the release, so even if it's ugly, I guess > the browntape fix is the least worst I can come up with. That looks fair. I have committed the proposed fix now. @butterflyoffire: could you please take a look at https://salsa.debian.org/installer-team/d-i/commit/fa17b7afffb2ee4888a371efa4153c5c075915e7 to double-check if the fix is fine by you? Thanks! Holger -- Holger Wansing PGP-Fingerprint: 496A C6E8 1442 4B34 8508 3529 59F1 87CA 156E B076
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Hello, Holger Wansing, le mar. 04 juin 2019 20:59:12 +, a ecrit: > Am Sonntag, 2. Juni 2019 schrieb Holger Wansing: > > Am Sonntag, 2. Juni 2019 schrieb Samuel Thibault: > > > ButterflyOfFire, le dim. 02 juin 2019 15:43:41 +, a ecrit: > > > > >> "لاحقاً" > > > > > > > > This word means "later" and can be replaced by "لاحقا". > > > > > > That avoids the issue here indeed. I however see it used in various > > > > Hmm. > > There has been no changing in the ar.po of partman-auto > > for 3 years now (JFTR), thus the real problem seems to be > > sitting somewhere else ... > > How should we proceed here? > Since we are late in the freeze, and it's most probably not that > easy to find out what happened and what the real issue is: > should we go with the "work-around" proposed by > Butterflyoffire and confirmed to fix the issue by Samuel? Personally, I don't think I'll have the time to debug what is actually happening inside gtk until the release, so even if it's ugly, I guess the browntape fix is the least worst I can come up with. Samuel
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Am Sonntag, 2. Juni 2019 schrieb Holger Wansing: > > Hi, > > Am Sonntag, 2. Juni 2019 schrieb Samuel Thibault: > > ButterflyOfFire, le dim. 02 juin 2019 15:43:41 +, a ecrit: > > > >> "لاحقاً" > > > > > > This word means "later" and can be replaced by "لاحقا". > > > > That avoids the issue here indeed. I however see it used in various > > Hmm. > There has been no changing in the ar.po of partman-auto > for 3 years now (JFTR), thus the real problem seems to be > sitting somewhere else ... How should we proceed here? Since we are late in the freeze, and it's most probably not that easy to find out what happened and what the real issue is: should we go with the "work-around" proposed by Butterflyoffire and confirmed to fix the issue by Samuel? Holger -- Sent from my Jolla phone http://www.jolla.com/
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Hi, Am Sonntag, 2. Juni 2019 schrieb Samuel Thibault: > ButterflyOfFire, le dim. 02 juin 2019 15:43:41 +, a ecrit: > > >> "لاحقاً" > > > > This word means "later" and can be replaced by "لاحقا". > > That avoids the issue here indeed. I however see it used in various Hmm. There has been no changing in the ar.po of partman-auto for 3 years now (JFTR), thus the real problem seems to be sitting somewhere else ... Holger -- Sent from my Jolla phone http://www.jolla.com/
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
ButterflyOfFire, le dim. 02 juin 2019 15:43:41 +, a ecrit: > >> "لاحقاً" > > This word means "later" and can be replaced by "لاحقا". That avoids the issue here indeed. I however see it used in various places, notably in the grub installer: msgid "" "Warning: If the installer failed to detect another operating system that is " "present on your computer, modifying the master boot record will make that " "operating system temporarily unbootable, though GRUB can be manually " "configured later to boot it." msgstr "" "تحذير: إن كان برنامج التثبيت قد فشل باكتشاف نظام تشغيلٍ آخر موجودٍ على حاسوبك، " "فإن تعديل سجلّ الإقلاع الرّئيسي سوف يجعل نظام التّشغيل ذاك غير قابلٍ للإقلاع " "مؤقّتاً، إلّا أنّه من الممكن تهيئة GRUB يدويّاً لإقلاعه لاحقاً." That one however shows up without any concern, that's odd... Samuel
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Hi, >> "لاحقاً" This word means "later" and can be replaced by "لاحقا".
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Samuel Thibault, le dim. 02 juin 2019 17:07:04 +0200, a ecrit: > Samuel Thibault, le dim. 02 juin 2019 16:47:49 +0200, a ecrit: > > But it's translated along the two other pieces "The installer can guide > > you..." and "If you choose guided..." so the issue might be in those. > > It's when I add this piece that gtk starts choking. Any idea? More precisely, it's this bit: > "لاحقاً" Samuel
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Samuel Thibault, le dim. 02 juin 2019 16:47:49 +0200, a ecrit: > But it's translated along the two other pieces "The installer can guide > you..." and "If you choose guided..." so the issue might be in those. It's when I add this piece that gtk starts choking. Any idea? "الفرصة لاحقاً لرؤية النتائج و تخصيصها." Samuel
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Samuel Thibault, le dim. 02 juin 2019 16:44:45 +0200, a ecrit: > Samuel Thibault, le dim. 02 juin 2019 16:33:03 +0200, a ecrit: > > Samuel Thibault, le dim. 02 juin 2019 16:06:50 +0200, a ecrit: > > > Samuel Thibault, le dim. 02 juin 2019 15:56:31 +0200, a ecrit: > > > > Samuel Thibault, le dim. 02 juin 2019 15:47:28 +0200, a ecrit: > > > > > Holger Wansing, le dim. 02 juin 2019 15:00:09 +0200, a ecrit: > > > > > > So, that would come down to this two commits > > > > > > (if it's a matter of arabic translations): > > > > > > https://salsa.debian.org/installer-team/d-i/commit/1e16f41d79578a47a370c7a47ec49abdcfc7fb52 > > > > > > https://salsa.debian.org/installer-team/d-i/commit/5654defa9a4b75196e83e905465cd3aa8704c093 > > > > > > > > > > I tried to revert these two, but that didn't fix the issue. > > > > > > > > The bug could be coming from another string, not showing up in beta5, > > > > but showing up in rc1. > > > > > > Dropping the translations of partman-auto does fix the issue for me. > > > > It's this translation which breaks things up: > > > > msgid "Partitioning method:" > > msgstr "طريقة التجزئة:" > > > > Any idea? :) > > Mmm, putting > > msgid "Partitioning method:" > msgstr "foobar:" > > there does get the issue as well... But it's translated along the two other pieces "The installer can guide you..." and "If you choose guided..." so the issue might be in those. Samuel
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Samuel Thibault, le dim. 02 juin 2019 16:33:03 +0200, a ecrit: > Samuel Thibault, le dim. 02 juin 2019 16:06:50 +0200, a ecrit: > > Samuel Thibault, le dim. 02 juin 2019 15:56:31 +0200, a ecrit: > > > Samuel Thibault, le dim. 02 juin 2019 15:47:28 +0200, a ecrit: > > > > Holger Wansing, le dim. 02 juin 2019 15:00:09 +0200, a ecrit: > > > > > So, that would come down to this two commits > > > > > (if it's a matter of arabic translations): > > > > > https://salsa.debian.org/installer-team/d-i/commit/1e16f41d79578a47a370c7a47ec49abdcfc7fb52 > > > > > https://salsa.debian.org/installer-team/d-i/commit/5654defa9a4b75196e83e905465cd3aa8704c093 > > > > > > > > I tried to revert these two, but that didn't fix the issue. > > > > > > The bug could be coming from another string, not showing up in beta5, > > > but showing up in rc1. > > > > Dropping the translations of partman-auto does fix the issue for me. > > It's this translation which breaks things up: > > msgid "Partitioning method:" > msgstr "طريقة التجزئة:" > > Any idea? :) Mmm, putting msgid "Partitioning method:" msgstr "foobar:" there does get the issue as well... Samuel
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Samuel Thibault, le dim. 02 juin 2019 16:06:50 +0200, a ecrit: > Samuel Thibault, le dim. 02 juin 2019 15:56:31 +0200, a ecrit: > > Samuel Thibault, le dim. 02 juin 2019 15:47:28 +0200, a ecrit: > > > Holger Wansing, le dim. 02 juin 2019 15:00:09 +0200, a ecrit: > > > > So, that would come down to this two commits > > > > (if it's a matter of arabic translations): > > > > https://salsa.debian.org/installer-team/d-i/commit/1e16f41d79578a47a370c7a47ec49abdcfc7fb52 > > > > https://salsa.debian.org/installer-team/d-i/commit/5654defa9a4b75196e83e905465cd3aa8704c093 > > > > > > I tried to revert these two, but that didn't fix the issue. > > > > The bug could be coming from another string, not showing up in beta5, > > but showing up in rc1. > > Dropping the translations of partman-auto does fix the issue for me. It's this translation which breaks things up: msgid "Partitioning method:" msgstr "طريقة التجزئة:" Any idea? :) Samuel
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Samuel Thibault, le dim. 02 juin 2019 15:56:31 +0200, a ecrit: > Samuel Thibault, le dim. 02 juin 2019 15:47:28 +0200, a ecrit: > > Holger Wansing, le dim. 02 juin 2019 15:00:09 +0200, a ecrit: > > > So, that would come down to this two commits > > > (if it's a matter of arabic translations): > > > https://salsa.debian.org/installer-team/d-i/commit/1e16f41d79578a47a370c7a47ec49abdcfc7fb52 > > > https://salsa.debian.org/installer-team/d-i/commit/5654defa9a4b75196e83e905465cd3aa8704c093 > > > > I tried to revert these two, but that didn't fix the issue. > > The bug could be coming from another string, not showing up in beta5, > but showing up in rc1. Dropping the translations of partman-auto does fix the issue for me. Samuel
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Samuel Thibault, le dim. 02 juin 2019 15:47:28 +0200, a ecrit: > Holger Wansing, le dim. 02 juin 2019 15:00:09 +0200, a ecrit: > > So, that would come down to this two commits > > (if it's a matter of arabic translations): > > https://salsa.debian.org/installer-team/d-i/commit/1e16f41d79578a47a370c7a47ec49abdcfc7fb52 > > https://salsa.debian.org/installer-team/d-i/commit/5654defa9a4b75196e83e905465cd3aa8704c093 > > I tried to revert these two, but that didn't fix the issue. The bug could be coming from another string, not showing up in beta5, but showing up in rc1. Samuel
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Holger Wansing, le dim. 02 juin 2019 15:00:09 +0200, a ecrit: > So, that would come down to this two commits > (if it's a matter of arabic translations): > https://salsa.debian.org/installer-team/d-i/commit/1e16f41d79578a47a370c7a47ec49abdcfc7fb52 > https://salsa.debian.org/installer-team/d-i/commit/5654defa9a4b75196e83e905465cd3aa8704c093 I tried to revert these two, but that didn't fix the issue. Samuel
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
ButterflyOfFire, le dim. 02 juin 2019 12:12:04 +, a ecrit: > Tested right now on my laptop, I cannot reproduce the bug. It seems working. > > Tested with : > kvm -cdrom /var/tmp/d-i/debian-buster-DI-rc1-amd64-netinst.iso -drive > file=blip -m 1G Hum :/ I tried to rebuild partman-base without the arabic translation, and I'm still getting the issue, so it's not something in there but somewhere else. Just to be sure, I put 4G of memory, with the same result. I also tried two dozens other languages and scripts, only Arabic gave me the issue. Samuel
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
While ButterflyOfFire reported to be unable to reproduce this, I can indeed reproduce this with qemu here. Looking more closely, that bug is there for me with the rc1 installer, however not with the alpha5 one. So, that would come down to this two commits (if it's a matter of arabic translations): https://salsa.debian.org/installer-team/d-i/commit/1e16f41d79578a47a370c7a47ec49abdcfc7fb52 https://salsa.debian.org/installer-team/d-i/commit/5654defa9a4b75196e83e905465cd3aa8704c093 @ButterflyOfFire: is there something eye-catching problematic with one of these changings? Holger -- Holger Wansing PGP-Fingerprint: 496A C6E8 1442 4B34 8508 3529 59F1 87CA 156E B076
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Samuel Thibault, le dim. 02 juin 2019 13:00:51 +0200, a ecrit: > Cc-ing debian-l10n-arabic because they can perhaps look at the partman > translation files and check that there is nothing odd in the file (e.g. > odd ligatures or such odd thing that might disturb gtk). I forgot to mention that I tried in Persian and didn't have the issue, so that confirms that there is probably something in the Arabic translation which triggers a bug in gtk. > While doing some tests, I noticed that when running the installer in > arabic language, the hard disk detection step hangs. To reproduce: > > - dd < /dev/zero > blip bs=1M count=1 seek=1 > - kvm -cdrom /var/tmp/d-i/debian-buster-DI-rc1-amd64-netinst.iso -drive > file=blip -m 2G > - Select the graphical installer > - Select the arabic language > - Select that you don't care that not everything is translated > - Select everything by default except typing non-empty passwords and > user name > > The partman packages get loaded but the installer remains stuck at the > attached screenshot. /var/log/syslog does not provide more information > than what we get when doing all the same in English. The ongoing > processes showing up in ps are > > udpkg --configure --force-configure partman-base > /var/lib/dpkg/info/partman-base.postinst configure > /bin/partman > parted_server > /lib/partman/display.d/10initial_auto > > Switching to VT4 to see the log then going back to VT5 shows a fully > gray screen, so I guess it's the gtk frontend which is somehow stuck due > to something in the arabic translation? > > I tried today's weekly build, with the same result (the only difference > being that it doesn't warn that not everything is translated). > > Samuel > > -- Package-specific info: > > Boot method: CD > Image version: debian-buster-DI-rc1-amd64-netinst.iso > > Machine: kvm > > > Base System Installation Checklist: > [O] = OK, [E] = Error (please elaborate below), [ ] = didn't try it > > Initial boot: [O] > Detect network card:[O] > Configure network: [O] > Detect CD: [O] > Load installer modules: [O] > Clock/timezone setup: [O] > User/password setup:[O] > Detect hard drives: [E] > Partition hard drives: [ ] > Install base system:[ ] > Install tasks: [ ] > Install boot loader:[ ] > Overall install:[ ] > > > -- > > Please make sure that the hardware-summary log file, and any other > installation logs that you think would be useful are attached to this > report. Please compress large files using gzip. > > Once you have filled out this report, mail it to sub...@bugs.debian.org.
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Hi, Tested right now on my laptop, I cannot reproduce the bug. It seems working. Tested with : kvm -cdrom /var/tmp/d-i/debian-buster-DI-rc1-amd64-netinst.iso -drive file=blip -m 1G Salutations,
Bug#929877: installation-reports: Buster installer hangs at hard disk step with arabic language
Package: installation-reports Severity: serious Justification: prevents from installing at all in arabic language Hello, Cc-ing debian-l10n-arabic because they can perhaps look at the partman translation files and check that there is nothing odd in the file (e.g. odd ligatures or such odd thing that might disturb gtk). While doing some tests, I noticed that when running the installer in arabic language, the hard disk detection step hangs. To reproduce: - dd < /dev/zero > blip bs=1M count=1 seek=1 - kvm -cdrom /var/tmp/d-i/debian-buster-DI-rc1-amd64-netinst.iso -drive file=blip -m 2G - Select the graphical installer - Select the arabic language - Select that you don't care that not everything is translated - Select everything by default except typing non-empty passwords and user name The partman packages get loaded but the installer remains stuck at the attached screenshot. /var/log/syslog does not provide more information than what we get when doing all the same in English. The ongoing processes showing up in ps are udpkg --configure --force-configure partman-base /var/lib/dpkg/info/partman-base.postinst configure /bin/partman parted_server /lib/partman/display.d/10initial_auto Switching to VT4 to see the log then going back to VT5 shows a fully gray screen, so I guess it's the gtk frontend which is somehow stuck due to something in the arabic translation? I tried today's weekly build, with the same result (the only difference being that it doesn't warn that not everything is translated). Samuel -- Package-specific info: Boot method: CD Image version: debian-buster-DI-rc1-amd64-netinst.iso Machine: kvm Base System Installation Checklist: [O] = OK, [E] = Error (please elaborate below), [ ] = didn't try it Initial boot: [O] Detect network card:[O] Configure network: [O] Detect CD: [O] Load installer modules: [O] Clock/timezone setup: [O] User/password setup:[O] Detect hard drives: [E] Partition hard drives: [ ] Install base system:[ ] Install tasks: [ ] Install boot loader:[ ] Overall install:[ ] -- Please make sure that the hardware-summary log file, and any other installation logs that you think would be useful are attached to this report. Please compress large files using gzip. Once you have filled out this report, mail it to sub...@bugs.debian.org.