Bug#987449: [rc-1 graphical d-i] graphical installer hangs at language chooser step for several languages

[Cyril Brulebois]

Control: block -1 by 987587

Hi,

Holger Wansing  (2021-04-24):
> While trying to debug another problem yesterday, I found that the RC1 
> graphical
> installer fails to go forward, when selecting one of these languages:
>  - Kannada
>  - Marathi
>  - Persian
>  - Sinhala
> 
> (tested with rc1-netinst and -netboot image, on an amd64 qemu machine
> with 1G RAM)
> 
> Installer hangs, with only displaying a headline and nothing more,
> eating up all CPU time.

Alright, finally, I'm able to reproduce the issue with locally built
netboot-gtk mini.iso but also a repacked RC 1 netinst with just the
cdrom/gtk/initrd.img file updated! (I decided to give up spending time
trying to get debian-cd to build an image really similar to the released
one, that's a nice to have for me but not needed to debug this issue,
and followed .)

I've been stuck for a while but diffoscope delivered the truth!

> Interestingly, yesterday's daily does not have this problem, while RC1
> has!!! And sadly it has not been discovered, that alpha3 is also
> affected, while alpha2 is fine!

Most of my testing (no pun intended) usually comes after one of those:

make -C build/ build_netboot-gtk USE_UDEBS_FROM=testing
make -C build/ build_netboot-gtk USE_UDEBS_FROM=unstable

For a netinst, one would need to use the build_cdrom_gtk target instead,
and deploy build/dest/cdrom/gtk under install.amd64/ in the ISO.

**BUT** (and that's a big but and I cannot lie ♫)

This minimal `make` call is lacking a variable that makes a huge
difference for the languages above. Their translations are incomplete
and trigger a warning after language selection, but that can only happen
if there's a translation-status file!

Quoting `debian/rules` selectively:

# Daily builds vs. uploads to unstable:
ifeq (${SUITE},UNRELEASED)
TRANSSTATUS=
else
TRANSSTATUS=translation-status
endif

Once I figured that out, reproducing the issue is just a matter of
adding this on the `make` line:

make -C build/ build_netboot-gtk USE_UDEBS_FROM=testing 
TRANSSTATUS=translation-status

> Yet another problem which started to appear with alpha3, just like 
> "Bug#987377: rescue-mode: when in graphical mode, locks up one prompt before 
> the shell"
> !!! 

Yes, that looks like the very same issue, as reverting to the old
libpango udeb means we no longer get stuck with those languages. For
more details about the revert, see this message:
  https://bugs.debian.org/987377#91

In case things change in the archive in the meanwhile (we should get at
least a new linux upload soon-ish, one can still stick to archive
contents from when debian-installer 20210415 (for RC 1) was built:

make -C build/ build_netboot-gtk USE_UDEBS_FROM=testing 
TRANSSTATUS=translation-status SNAPSHOT_TS=20210415T151642Z

> The commit 
> https://salsa.debian.org/installer-team/debian-installer/-/commit/95fdc73ca8cde8d7a360fd3d742fc947a045ec0f
> "Drop fontconfig tweaks introduced in version 20170828 (See: #873462)" 
> comes to my mind, also included in alpha3 ... ???

Irrelevant in the end.

> The text installer seems to be not affected (well, only Persian from
> the langs above is available in text installer, and Persian works
> there).

Not surprising in hindsight since that's libpango. :)

> We already have 
> "Bug#983897: installation-reports: Installation in Uyghur hangs"
> where the installer hangs at some later steps - noticed for Uyghur and
> Kannada so far...

Yes, this is likely the kind of “bad luck somewhen” similar to what
happens with rescue mode for some languages (e.g. English) but not all
(e.g. neither French nor German).

> Uyghur is however not affected by this language chooser fail bug.

This part is likely explained by the fact it's at `2 M` while affected
languages are at `2 P` in `build/translation-status`. Note that all
languages at `2 P` don't trigger this bug though (e.g. Telugu doesn't).
Full list at `2 P`:
  am bn bo bs dz ka km ku lo mk ne nn si tl


Back to the “bad luck somewhen” aspect for Uyghur, it would be nice to
check with an unpatched and with a patched image. I might try and do
that at some point, but since both #987377 and this bug are being
confirmed as originating from the same issue (#987587), it looks to me
like I should move on to trying to find a solution (sticking to an old
version of the udeb isn't really a suitable one…).

> As already mentioned by Samuel, this may be another resurgence - in a
> tightened form - of
> "#929877: installation-reports: Buster installer hangs at hard disk step with 
> arabic language"
> from 2019, where the real reason has not been found...

From a quick read, I would think that's a rather different issue, due to
difficulties with some specific glyphs or glyph combinations.


Cheers,
-- 
Cyril Brulebois (k...@debian.org)
D-I release manager -- Release team member -- Freelance Consultant


signature.asc
Description: PGP signature

Bug#987377: Bug#987449: [rc-1 graphical d-i] graphical installer hangs at language chooser step for several languages

[Cyril Brulebois]

Hi,

Holger Wansing  (2021-04-25):
> Chris Hofstaedtler :
> > I spent a few hours today on this issue, but did not get very far.
> > Not totally unexpected if you know nothing about the installer.

And yet!

> > I have, however, a perf top output to share. I believe this more or
> > less means the font rendering (pango, fontconfig) is stuck somehow?

I believe the root cause of this issue is the same as for #987377, even
if I can't immediately confirm it for the reason mentioned below.

> That might support my wild guess, that "Drop fontconfig tweaks" changings 
> could be related.
> https://salsa.debian.org/installer-team/debian-installer/-/commit/95fdc73ca8cde8d7a360fd3d742fc947a045ec0f

I would have probably had the same initial intuition as you, but that's
very likely irrelevant.

> However, I don't get it, why recent dailies are not affected, while
> RC1 (and alpha3) is ???
…
> Are there differences during build for the "release builds" (alpha,
> rc) compared to the dailies?

I think you are testing different things:
 - For Alpha 2, Alpha 3, RC 1, you're likely using the Netinst ISO,
   which was fine up to Alpha 2, and started breaking with Alpha 3?
 - For daily builds, you're using the netboot/gtk/mini.iso ISO?

Locally, I can confirm that a locally built netboot-gtk ISO does *not*
exhibit the obvious issue on regular installation in e.g. Kannada while
the Netinst ISO for say RC 1 triggers the issue immediately.


There might be some fun thing going on, possibly memory corruption? That
might explain why we run into problems or we don't, depending on which
ISO is used, what the initramfs exactly contains (Netinst ISO don't use
the same initramfs as netboot-gtk), how files are ordered, or whether an
English or a French locale is used (English breaks, French or German is
fine, for #987377), what bootloader was used, etc.


I would need to restore a working debian-cd setup locally to have higher
chances of triggering the issue with a locally-produced d-i + Netinst
ISO, then proceeding to confirm it goes away with a package revert.

To avoid more speculations in #987449 (I have no conclusions yet, just a
*very strong* suspicion), I'll keep my findings to #987377 in a separate
reply.


Thanks everyone involved so far, tracking down the Alpha 2 vs. Alpha 3
issue (rather than Alpha 3 vs. RC 1 as one might have thought) was very
neat, it got me started with the idea we could actually have a common
issue that expresses itself in rather different and convoluted ways…


Cheers,
-- 
Cyril Brulebois (k...@debian.org)
D-I release manager -- Release team member -- Freelance Consultant


signature.asc
Description: PGP signature

Bug#987449: [rc-1 graphical d-i] graphical installer hangs at language chooser step for several languages

[Holger Wansing]

Hi,

Chris Hofstaedtler  wrote (Sun, 25 Apr 2021 14:13:05 +0200):
> While the -exact- same bug does not appear on the
> 20210424-7/amd64/iso-cd/debian-testing-amd64-netinst.iso image, I
> think the bug is still there, just better hidden.
> I managed to make debconf hang by:
>   - select Marathi
>   - on the next screen, click the Back button
>   - in the now appearing list, select something else (I think two
> above from what was selected)
>   - click the Continue button
>   - hangs

Yes, that way it is still reproducible with recent dailies or self-built
images.

Based on this, I have patched my build tree according
https://people.debian.org/~holgerw/d-i/proposal-to-switch-g-i-to-fonts-noto/20200402/
to use noto fonts for all languages except CJK, and with the resulting
netboot gtk mini.iso I am unable to make debconf hang, whatever I do!
(including the above path)

This might be a pointer, that this problem is indeed a font issue...


Holger



-- 
Holger Wansing 
PGP-Fingerprint: 496A C6E8 1442 4B34 8508  3529 59F1 87CA 156E B076

Bug#987449: [rc-1 graphical d-i] graphical installer hangs at language chooser step for several languages

[Chris Hofstaedtler]

* Holger Wansing  [210425 16:53]:
> Chris Hofstaedtler  wrote (Sun, 25 Apr 2021 14:13:05 +0200):
> > While the -exact- same bug does not appear on the
> > 20210424-7/amd64/iso-cd/debian-testing-amd64-netinst.iso image, I
> > think the bug is still there, just better hidden.
> > I managed to make debconf hang by:
> >   - select Marathi
> >   - on the next screen, click the Back button
> >   - in the now appearing list, select something else (I think two
> > above from what was selected)
> >   - click the Continue button
> >   - hangs
> 
> Yes, that way it is still reproducible with recent dailies or self-built
> images.

It might be helpful to try older images, and see if they also show
the hang if triggered like this. Then finding the first one that
doesnt ... maybe that yields a super[cg]lue.

Chris

Bug#987449: [rc-1 graphical d-i] graphical installer hangs at language chooser step for several languages

[Holger Wansing]

Hi,

Chris Hofstaedtler  wrote (Sat, 24 Apr 2021 23:45:13 +0200):
> Hello,
> 
> * Holger Wansing  [210424 21:42]:
> > While trying to debug another problem yesterday, I found that the RC1 
> > graphical
> > installer fails to go forward, when selecting one of these languages:
> > - Kannada
> > - Marathi
> > - Persian
> > - Sinhala
> 
> I spent a few hours today on this issue, but did not get very far.
> Not totally unexpected if you know nothing about the installer.
> 
> I have, however, a perf top output to share. I believe this more or less means
> the font rendering (pango, fontconfig) is stuck somehow?

That might support my wild guess, that "Drop fontconfig tweaks" changings 
could be related.
https://salsa.debian.org/installer-team/debian-installer/-/commit/95fdc73ca8cde8d7a360fd3d742fc947a045ec0f

However, I don't get it, why recent dailies are not affected, while RC1 (and
alpha3) is ???

(That might be the reason, why this issue slipped through our attention until
now...)

Are there differences during build for the "release builds" (alpha, rc) 
compared to the dailies?

CC'ing Steve for debian-cd side...

Holger



> Good luck,
> Chris
> 
> Samples: 82K of event 'cpu-clock:pppH', 4000 Hz, Event count (approx.): 
> 4370334125 lost: 0/0 drop: 0/0
> Overhead  Shared ObjectSymbol
>5.29%  /usr/lib/x86_64-linux-gnu/libpango-1.0.so.0.4600.2   0x13010
> d [.] pango_default_break 
>1.44%  /lib/libharfbuzz.so.0.20704.00x769b0
> ! [.] 0x000769b0  
>1.17%  /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6600.80x8487e
> B [.] g_utf8_get_char 
>1.05%  /lib/libharfbuzz.so.0.20704.00x57448
> ! [.] 0x00057448  
>1.04%  /usr/lib/x86_64-linux-gnu/libpango-1.0.so.0.4600.2   0x1fff2
> d [.] 0x0001fff2  
>1.02%  /lib/libharfbuzz.so.0.20704.00x8eff5
> ! [.] 0x0008eff5  
>0.88%  /lib/libharfbuzz.so.0.20704.00x611d0
> ! [.] 0x000611d0  
>0.83%  /lib/libharfbuzz.so.0.20704.00x5741e
> ! [.] 0x0005741e  
>0.81%  /lib/x86_64-linux-gnu/libc-2.31.so   0x88bbd
> D [.] _int_malloc 
>0.80%  /lib/libharfbuzz.so.0.20704.00x5a9f7
> ! [.] 0x0005a9f7  
>0.79%  /lib/libharfbuzz.so.0.20704.00x573d9
> ! [.] 0x000573d9  
>0.71%  /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6600.80x3edd8
> B [.] g_hash_table_lookup 
>0.70%  /usr/lib/x86_64-linux-gnu/libpango-1.0.so.0.4600.2   0xf5a0 
> d [.] g_utf8_get_char@plt 
>0.68%  /lib/libharfbuzz.so.0.20704.00x573ee
> ! [.] 0x000573ee  
>0.60%  /lib/libharfbuzz.so.0.20704.00x5a830
> ! [.] 0x0005a830  
>0.57%  /lib/libharfbuzz.so.0.20704.00x71730
> ! [.] 0x00071730  
>0.57%  /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6600.80x70590
> B [.] g_slice_free1   
>0.56%  /usr/lib/x86_64-linux-gnu/libgtk-x11-2.0.so.0.2400.330x1e7883   
> d [.] gtk_text_layout_get_line_display
>0.55%  /lib/libharfbuzz.so.0.20704.00x6a190
> ! [.] 0x0006a190  
>0.52%  /usr/lib/x86_64-linux-gnu/libpango-1.0.so.0.4600.2   0x34977
> d [.] pango_shape_with_flags  
>0.51%  /lib/libharfbuzz.so.0.20704.00xc210 
> ! [.] 0xc210  
>0.51%  /usr/lib/x86_64-linux-gnu/libgobject-2.0.so.0.6600.8 0x18fff
> B [.] g_object_unref  
>0.46%  /lib/libharfbuzz.so.0.20704.00x578b2
> ! [.] 0x000578b2  
>0.44%  /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6600.80x6ff70
> B [.] g_slice_alloc   
>0.43%  /lib/x86_64-linux-gnu/libc-2.31.so   0x86b1d
> D [.] _int_free   
>0.41%  /lib/libharfbuzz.so.0.20704.00x5a8a4
> ! [.] 0x0005a8a4  
>0.40%  /lib/libharfbuzz.so.0.20704.0 

Bug#987449: [rc-1 graphical d-i] graphical installer hangs at language chooser step for several languages

[Chris Hofstaedtler]

Hi,

* Holger Wansing  [210425 11:34]:
> Chris Hofstaedtler  wrote (Sat, 24 Apr 2021 23:45:13 +0200):
> > * Holger Wansing  [210424 21:42]:
> > > While trying to debug another problem yesterday, I found that the RC1 
> > > graphical
> > > installer fails to go forward, when selecting one of these languages:
> > >   - Kannada
> > >   - Marathi
> > >   - Persian
> > >   - Sinhala
> > 
> > I spent a few hours today on this issue, but did not get very far.
> > Not totally unexpected if you know nothing about the installer.
> > 
> > I have, however, a perf top output to share. I believe this more or less 
> > means
> > the font rendering (pango, fontconfig) is stuck somehow?
> 
> That might support my wild guess, that "Drop fontconfig tweaks" changings 
> could be related.
> https://salsa.debian.org/installer-team/debian-installer/-/commit/95fdc73ca8cde8d7a360fd3d742fc947a045ec0f
> 
> However, I don't get it, why recent dailies are not affected, while RC1 (and
> alpha3) is ???

While the -exact- same bug does not appear on the
20210424-7/amd64/iso-cd/debian-testing-amd64-netinst.iso image, I
think the bug is still there, just better hidden.
I managed to make debconf hang by:
  - select Marathi
  - on the next screen, click the Back button
  - in the now appearing list, select something else (I think two
above from what was selected)
  - click the Continue button
  - hangs

Chris

Bug#987449: [rc-1 graphical d-i] graphical installer hangs at language chooser step for several languages

[Cyril Brulebois]

Holger Wansing  (2021-04-25):
> Cyril Brulebois  wrote (Sun, 25 Apr 2021 11:53:22 +0200):
> > Given the question below, I suppose you could revert this change
> > manually and restart the installer (after all, it's about a symlink and
> > a line in a config file).
> 
> Sure. I did that already.
> But this does not gain any value, because building an image myself now without
> doing any changings leads to a working d-i (on i386, if arch matters) !!!

I understood that part. :)

> I don't need to do anything to get a working d-i, just build myself.
> 
> But why does the RC1 fail then?

I meant that you could modify a released image (Alpha 3 or RC 1),
early after you have booted it; this way you would see whether
reverting the change manually would change anything.


Cheers,
-- 
Cyril Brulebois (k...@debian.org)
D-I release manager -- Release team member -- Freelance Consultant


signature.asc
Description: PGP signature

Bug#987449: [rc-1 graphical d-i] graphical installer hangs at language chooser step for several languages

[Holger Wansing]

Hi,

Cyril Brulebois  wrote (Sun, 25 Apr 2021 11:53:22 +0200):
> Given the question below, I suppose you could revert this change
> manually and restart the installer (after all, it's about a symlink and
> a line in a config file).

Sure. I did that already.
But this does not gain any value, because building an image myself now without
doing any changings leads to a working d-i (on i386, if arch matters) !!!

I don't need to do anything to get a working d-i, just build myself.

But why does the RC1 fail then?

Holger



> > However, I don't get it, why recent dailies are not affected, while
> > RC1 (and alpha3) is ???
> > 
> > (That might be the reason, why this issue slipped through our
> > attention until now...)
> > 
> > Are there differences during build for the "release builds" (alpha,
> > rc) compared to the dailies?
> 
> You can check debian/rules for USE_UDEBS_FROM:
>  - releases fetch their udebs from testing;
>  - dailies fetch their udebs from unstable (to shorten the feedback
>loop, and to inform our decision to let some packages migrate or not
>when a freeze — udebs or otherwise — is in effect)
> 
> You can switch to/from UNRELEASED in the first line of the changelog to
> get the desired results. Or you can set the variable yourself if you run
> something like this:
> 
> make -C build/ rebuild_netboot-gtk USE_UDEBS_FROM=unstable
> 
> 
> Cheers,
> -- 
> Cyril Brulebois (k...@debian.org)
> D-I release manager -- Release team member -- Freelance Consultant


-- 
Holger Wansing 
PGP-Fingerprint: 496A C6E8 1442 4B34 8508  3529 59F1 87CA 156E B076

Bug#987449: [rc-1 graphical d-i] graphical installer hangs at language chooser step for several languages

[Cyril Brulebois]

Holger Wansing  (2021-04-25):
> Chris Hofstaedtler  wrote (Sat, 24 Apr 2021 23:45:13 +0200):
> > Hello,
> > 
> > * Holger Wansing  [210424 21:42]:
> > > While trying to debug another problem yesterday, I found that the RC1 
> > > graphical
> > > installer fails to go forward, when selecting one of these languages:
> > >   - Kannada
> > >   - Marathi
> > >   - Persian
> > >   - Sinhala
> > 
> > I spent a few hours today on this issue, but did not get very far.
> > Not totally unexpected if you know nothing about the installer.
> > 
> > I have, however, a perf top output to share. I believe this more or
> > less means the font rendering (pango, fontconfig) is stuck somehow?

Only quickly looked at pango1.0, freetype, fontconfig, those have last
changed in 2020, so before alpha3.

> That might support my wild guess, that "Drop fontconfig tweaks"
> changings could be related.
> https://salsa.debian.org/installer-team/debian-installer/-/commit/95fdc73ca8cde8d7a360fd3d742fc947a045ec0f

Given the question below, I suppose you could revert this change
manually and restart the installer (after all, it's about a symlink and
a line in a config file).

> However, I don't get it, why recent dailies are not affected, while
> RC1 (and alpha3) is ???
> 
> (That might be the reason, why this issue slipped through our
> attention until now...)
> 
> Are there differences during build for the "release builds" (alpha,
> rc) compared to the dailies?

You can check debian/rules for USE_UDEBS_FROM:
 - releases fetch their udebs from testing;
 - dailies fetch their udebs from unstable (to shorten the feedback
   loop, and to inform our decision to let some packages migrate or not
   when a freeze — udebs or otherwise — is in effect)

You can switch to/from UNRELEASED in the first line of the changelog to
get the desired results. Or you can set the variable yourself if you run
something like this:

make -C build/ rebuild_netboot-gtk USE_UDEBS_FROM=unstable


Cheers,
-- 
Cyril Brulebois (k...@debian.org)
D-I release manager -- Release team member -- Freelance Consultant


signature.asc
Description: PGP signature

Bug#987449: [rc-1 graphical d-i] graphical installer hangs at language chooser step for several languages

[Chris Hofstaedtler]

Hello,

* Holger Wansing  [210424 21:42]:
> While trying to debug another problem yesterday, I found that the RC1 
> graphical
> installer fails to go forward, when selecting one of these languages:
>   - Kannada
>   - Marathi
>   - Persian
>   - Sinhala

I spent a few hours today on this issue, but did not get very far.
Not totally unexpected if you know nothing about the installer.

I have, however, a perf top output to share. I believe this more or less means
the font rendering (pango, fontconfig) is stuck somehow?

Good luck,
Chris

Samples: 82K of event 'cpu-clock:pppH', 4000 Hz, Event count (approx.): 
4370334125 lost: 0/0 drop: 0/0
Overhead  Shared ObjectSymbol
   5.29%  /usr/lib/x86_64-linux-gnu/libpango-1.0.so.0.4600.2   0x13010  
  d [.] pango_default_break 
   1.44%  /lib/libharfbuzz.so.0.20704.00x769b0  
  ! [.] 0x000769b0  
   1.17%  /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6600.80x8487e  
  B [.] g_utf8_get_char 
   1.05%  /lib/libharfbuzz.so.0.20704.00x57448  
  ! [.] 0x00057448  
   1.04%  /usr/lib/x86_64-linux-gnu/libpango-1.0.so.0.4600.2   0x1fff2  
  d [.] 0x0001fff2  
   1.02%  /lib/libharfbuzz.so.0.20704.00x8eff5  
  ! [.] 0x0008eff5  
   0.88%  /lib/libharfbuzz.so.0.20704.00x611d0  
  ! [.] 0x000611d0  
   0.83%  /lib/libharfbuzz.so.0.20704.00x5741e  
  ! [.] 0x0005741e  
   0.81%  /lib/x86_64-linux-gnu/libc-2.31.so   0x88bbd  
  D [.] _int_malloc 
   0.80%  /lib/libharfbuzz.so.0.20704.00x5a9f7  
  ! [.] 0x0005a9f7  
   0.79%  /lib/libharfbuzz.so.0.20704.00x573d9  
  ! [.] 0x000573d9  
   0.71%  /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6600.80x3edd8  
  B [.] g_hash_table_lookup 
   0.70%  /usr/lib/x86_64-linux-gnu/libpango-1.0.so.0.4600.2   0xf5a0   
  d [.] g_utf8_get_char@plt 
   0.68%  /lib/libharfbuzz.so.0.20704.00x573ee  
  ! [.] 0x000573ee  
   0.60%  /lib/libharfbuzz.so.0.20704.00x5a830  
  ! [.] 0x0005a830  
   0.57%  /lib/libharfbuzz.so.0.20704.00x71730  
  ! [.] 0x00071730  
   0.57%  /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6600.80x70590  
  B [.] g_slice_free1   
   0.56%  /usr/lib/x86_64-linux-gnu/libgtk-x11-2.0.so.0.2400.330x1e7883 
  d [.] gtk_text_layout_get_line_display
   0.55%  /lib/libharfbuzz.so.0.20704.00x6a190  
  ! [.] 0x0006a190  
   0.52%  /usr/lib/x86_64-linux-gnu/libpango-1.0.so.0.4600.2   0x34977  
  d [.] pango_shape_with_flags  
   0.51%  /lib/libharfbuzz.so.0.20704.00xc210   
  ! [.] 0xc210  
   0.51%  /usr/lib/x86_64-linux-gnu/libgobject-2.0.so.0.6600.8 0x18fff  
  B [.] g_object_unref  
   0.46%  /lib/libharfbuzz.so.0.20704.00x578b2  
  ! [.] 0x000578b2  
   0.44%  /usr/lib/x86_64-linux-gnu/libglib-2.0.so.0.6600.80x6ff70  
  B [.] g_slice_alloc   
   0.43%  /lib/x86_64-linux-gnu/libc-2.31.so   0x86b1d  
  D [.] _int_free   
   0.41%  /lib/libharfbuzz.so.0.20704.00x5a8a4  
  ! [.] 0x0005a8a4  
   0.40%  /lib/libharfbuzz.so.0.20704.00x612b0  
  ! [.] 0x000612b0  
   0.39%  /usr/lib/x86_64-linux-gnu/libgobject-2.0.so.0.6600.8 0x2bbe8  
  B [.] g_signal_emit_valist
   0.39%  /lib/libharfbuzz.so.0.20704.00x717e0  
  ! [.] 0x000717e0  
   0.39%  /lib/libharfbuzz.so.0.20704.00x5a86b  
  ! [.] 0x0005a86b  
   0.38%  /lib/x86_64-linux-gnu/libc-2.31.so   0x95760  
  D [.] __strcmp_sse2   
   0.38%  /lib/libharfbuzz.so.0.20704.00x612d2  
  ! [.] 

Bug#987449: [rc-1 graphical d-i] graphical installer hangs at language chooser step for several languages

[Holger Wansing]

Package: debian-installer
Version: bullseye-rc1
Severity: grave


While trying to debug another problem yesterday, I found that the RC1 graphical
installer fails to go forward, when selecting one of these languages:
- Kannada
- Marathi
- Persian
- Sinhala

(tested with rc1-netinst and -netboot image, on an amd64 qemu machine with 1G 
RAM)

Installer hangs, with only displaying a headline and nothing more, eating up
all CPU time.

Interestingly, yesterday's daily does not have this problem, while RC1 has!!!
And sadly it has not been discovered, that alpha3 is also affected, while 
alpha2 is fine!

Yet another problem which started to appear with alpha3, just like 
"Bug#987377: rescue-mode: when in graphical mode, locks up one prompt before 
the shell"
!!! 

The commit 
https://salsa.debian.org/installer-team/debian-installer/-/commit/95fdc73ca8cde8d7a360fd3d742fc947a045ec0f
"Drop fontconfig tweaks introduced in version 20170828 (See: #873462)" 
comes to my mind, also included in alpha3 ... ???

The text installer seems to be not affected (well, only Persian from the langs
above is available in text installer, and Persian works there).


We already have 
"Bug#983897: installation-reports: Installation in Uyghur hangs"
where the installer hangs at some later steps - noticed for Uyghur and Kannada
so far...
Uyghur is however not affected by this language chooser fail bug.

As already mentioned by Samuel, this may be another resurgence - in a tightened
form - of
"#929877: installation-reports: Buster installer hangs at hard disk step with 
arabic language"
from 2019, where the real reason has not been found...


-- 
Holger Wansing 
PGP-Finterprint: 496A C6E8 1442 4B34 8508  3529 59F1 87CA 156E B076