Re: Have any difficulties arisen with the Lua scripting languaje into Hurd?

2026-02-11 Thread Jose Luis Alarcon Sanchez
El Wed, 11 Feb 2026 16:19:40 +0100
Samuel Thibault  escribió:
> Diego Nieto Cid, le mer. 11 févr. 2026 11:12:34 -0300, a ecrit:
> > This is the line with the assert:
> > 
> >     
> > [3]https://sources.debian.org/src/webkit2gtk/2.50.5-1/Source/JavaScriptCore
> > /heap/StructureAlignedMemoryAllocator.cpp#L106
> > 
> > And structureIDMask is defined  as 4GB - 1  here:
> > 
> >     
> > [4]https://sources.debian.org/src/webkit2gtk/2.50.5-1/Source/JavaScriptCore
> > /runtime/StructureID.h?hl=73#L56
> > 
> > I'm not sure, but it seems the code at
> > StructureAlignedMemoryAllocator.cpp is trying to allocate 4GB and,
> > since gnumach is in the dpkg log, I have to ask how much memory
> > does the hurd have?  
> 

I usually run Debian GNU/Hurd with 3 Gb. RAM + 128 Mb. Swap (swapfile)

> Also, checking the output of
> 
> ulimit -v
> 

$ ulimit -v
3145532
$

> Possibly, that would be a case for implementing MAP_NORESERVE.
> 
> Samuel
> 
> > 
> > THX.
> > Jose Luis.
> > 
> > References:
> > 
> > [1] mailto:[email protected]
> > [2] mailto:[email protected]
> > [3]
> > https://sources.debian.org/src/webkit2gtk/2.50.5-1/Source/JavaScriptCore/heap/
> > [4]
> > https://sources.debian.org/src/webkit2gtk/2.50.5-1/Source/JavaScriptCore/runtime/StructureID.h?hl=73#L56
> > [5] mailto:[email protected] [6] mailto:[email protected]
> > [7] mailto:[email protected] [8]
> > mailto:[email protected] [9] https://lordofunix.org/  
> 

Thanks!.
Jose Luis.

-- 
https://lordofunix.org/ 

Not Registered GNU/Hurd User.
Registered BSD User 51101.
Registered Linux User #213309.
Memories. You are talking about memories.
Rick Deckard. Blade Runner.


pgpNzFHOW5XNN.pgp
Description: Firma digital OpenPGP


Re: Have any difficulties arisen with the Lua scripting languaje into Hurd?

2026-02-11 Thread Jose Luis Alarcon Sanchez
El Wed, 11 Feb 2026 12:55:07 +0100
Samuel Thibault  escribió:
> Jose Luis Alarcon Sanchez, le mer. 11 févr. 2026 11:00:27 +0100, a
> ecrit:
> > El Tue, 10 Feb 2026 22:53:39 +0100
> > Samuel Thibault  escribió:  
> > > Jose Luis Alarcon Sanchez, le mar. 10 févr. 2026 20:41:38 +0100, a
> > > ecrit:  
> > > > El Tue, 10 Feb 2026 19:23:17 +0100
> > > > Samuel Thibault  escribió:
> > > > > Hello,
> > > > > 
> > > > > Jose Luis Alarcon Sanchez, le mar. 10 févr. 2026 15:37:08
> > > > > +0100, a ecrit:
> > > > > > How is it possible that at the end of last year, from one
> > > > > > day to the next, the two best web browsers available on
> > > > > > GNU/Hurd, Luakit and Surf, which are very similar and based
> > > > > > on the same scripting language, Lua, went from working very
> > > > > > satisfactorily to instantly terminating when launched, with
> > > > > > the brief output ‘abort’?   
> > > > > 
> > > > > Because of a bug? The abort message would probably be very
> > > > > informative. Or a gdb backtrace, etc.
> > > > > 
> > > > > > Is it possible to investigate why this happened?.  
> > > > > 
> > > > > Sure. For a simple start, you can check what package got
> > > > > upgraded in between, and then try to downgrade/upgrade to
> > > > > verify which exact package pose problems. But also simply
> > > > > actually pasting the abortion message would provide the list
> > > > > with *way* more information than just "terminates".
> > > > 
> > > > I don't know how to get a "register" of the upgrades realized at
> > > > that time (it was last week of 2025, i'm sure of it).
> > > 
> > > It's all in /var/log/dpkg.log
> > >  
> > 
> > Ok. Attached a short extract of /var/log/dpkg.log. I am 100% sure
> > that the changes that triggered the browsers falilure was made on
> > December 29th. Also including the day before and the day after.  
> 
> Then you can try to downgrade/reupgrade, to confirm which one poses
> problem.
>

I'd like downgrade the libjavascriptcoregtk and try to run the
browsers...
 
> > #6  0x000108e4bfcd in __GI_abort () at ./stdlib/abort.c:77
> > #7  0x000107ae1bec in ??? () at
> > /usr/lib/x86_64-gnu/libjavascriptcoregtk-4.1.so.0 #8
> > 0x000107f76912 in ??? () at
> > /usr/lib/x86_64-gnu/libjavascriptcoregtk-4.1.so.0  
> 
> So you want to install the -dbgsym package for that library, to get
> source code information
> 

First time using 'gdb' tool. Fast learning curve, i think. Attached
luakit_crash.log with the tool correctly applied.

> Samuel
> 

THX.
Jose Luis.

-- 
https://lordofunix.org/ 

Usuario GNU/Hurd no registrado.
Usuario BSD registrado 51101.
Usuario Linux registrado #213309.
Una vez mas cabalgaré con mis caballeros,
para defender lo que fue.
y el sueño de lo que pudo ser.

 
Not Registered GNU/Hurd User.
Registered BSD User 51101.
Registered Linux User #213309.
Memories. You are talking about memories.
Rick Deckard. Blade Runner.
Starting program: /usr/bin/luakit 
[bogus thread id 1 exited]
[bogus thread id 2 exited]
[bogus thread id 3 exited]
[New Thread 1688.5]
[New Thread 1688.6]
[New Thread 1688.7]
[New Thread 1688.8]
[New Thread 1688.9]

Thread 4 received signal SIGABRT, Aborted.
0x000108e4ccac in __GI___mach_msg_trap () at ./build-tree/hurd-amd64-libc/mach/mach_msg_trap.S:2
⚠️ warning: 2	./build-tree/hurd-amd64-libc/mach/mach_msg_trap.S: No existe el fichero o el directorio

Thread 9 (Thread 1688.9):
#0  0x000108e4ccac in __GI___mach_msg_trap () at ./build-tree/hurd-amd64-libc/mach/mach_msg_trap.S:2
#1  0x000108e4d38d in __GI___mach_msg (msg=msg@entry=0x11204bc28, option=option@entry=1, send_size=send_size@entry=80, rcv_size=rcv_size@entry=0, rcv_name=rcv_name@entry=0, timeout=timeout@entry=0, notify=0) at ./mach/msg.c:111
ret = 
#2  0x00010bfdf8a4 in __gsync_wake (task=, addr=, val=val@entry=0, flags=) at ./build-tree/hurd-amd64-libc/mach/RPC_gsync_wake.c:129
Mess = {In = {Head = {msgh_bits = 19, msgh_size = 80, {msgh_remote_port = 62, msgh_remote_port_do_not_use = 62}, {msgh_local_port = 0, msgh_protected_payload = 0}, msgh_seqno = 0, msgh_id = 4205}, addrType = {msgt_name = 11, msgt_size = 64, msgt_unused = 0, msgt_inline = 1, msgt_longform = 0, msgt_deallocate = 0, msgt_number = 1}, addr = 35184372791616, valType = {msgt_name = 2, msgt_size = 32, msgt_unused = 0, msgt_inline = 1, msgt_longform = 0, msgt_deallocate = 0, msgt_number = 1}, val = 0, valPad = "\000 \000", flagsType = {msgt_name = 2, msgt_size = 32, msgt_unused = 0, msgt_inline = 1, msgt_longform = 0, msgt_deallocate = 0, msgt_number = 1}, flags = 0, flagsPad = "\000\000\000"}}
InP = 0x11204bc28
#3  0x000108f0c6da in __GI___pthread_mutex_unlock (mtxp=) at ../sysdeps/mach/hurd/htl/pt-mutex-unlock.c:47
__iptr = 
self = 
flags = 
__iptr = 
#4  0x000101130d1a in g_mutex_unlock () at /usr/lib/x86_64-gnu/libglib-2.0.so.0
#5  0x000106b1bfa0 in ??? () at /usr/lib/x86_64-gnu/libgio-2.0.so.0
#6  0

Re: Have any difficulties arisen with the Lua scripting languaje into Hurd?

2026-02-11 Thread Samuel Thibault
Jose Luis Alarcon Sanchez, le mer. 11 févr. 2026 11:00:27 +0100, a ecrit:
> El Tue, 10 Feb 2026 22:53:39 +0100
> Samuel Thibault  escribió:
> > Jose Luis Alarcon Sanchez, le mar. 10 févr. 2026 20:41:38 +0100, a
> > ecrit:
> > > El Tue, 10 Feb 2026 19:23:17 +0100
> > > Samuel Thibault  escribió:  
> > > > Hello,
> > > > 
> > > > Jose Luis Alarcon Sanchez, le mar. 10 févr. 2026 15:37:08 +0100, a
> > > > ecrit:  
> > > > > How is it possible that at the end of last year, from one
> > > > > day to the next, the two best web browsers available on
> > > > > GNU/Hurd, Luakit and Surf, which are very similar and based on
> > > > > the same scripting language, Lua, went from working very
> > > > > satisfactorily to instantly terminating when launched, with the
> > > > > brief output ‘abort’? 
> > > > 
> > > > Because of a bug? The abort message would probably be very
> > > > informative. Or a gdb backtrace, etc.
> > > >   
> > > > > Is it possible to investigate why this happened?.
> > > > 
> > > > Sure. For a simple start, you can check what package got upgraded
> > > > in between, and then try to downgrade/upgrade to verify which
> > > > exact package pose problems. But also simply actually pasting the
> > > > abortion message would provide the list with *way* more
> > > > information than just "terminates".  
> > > 
> > > I don't know how to get a "register" of the upgrades realized at
> > > that time (it was last week of 2025, i'm sure of it).  
> > 
> > It's all in /var/log/dpkg.log
> >
> 
> Ok. Attached a short extract of /var/log/dpkg.log. I am 100% sure that
> the changes that triggered the browsers falilure was made on December
> 29th. Also including the day before and the day after.

Then you can try to downgrade/reupgrade, to confirm which one poses
problem.

> #6  0x000108e4bfcd in __GI_abort () at ./stdlib/abort.c:77
> #7  0x000107ae1bec in ??? () at 
> /usr/lib/x86_64-gnu/libjavascriptcoregtk-4.1.so.0
> #8  0x000107f76912 in ??? () at 
> /usr/lib/x86_64-gnu/libjavascriptcoregtk-4.1.so.0

So you want to install the -dbgsym package for that library, to get
source code information

Samuel



Re: Have any difficulties arisen with the Lua scripting languaje into Hurd?

2026-02-11 Thread Jose Luis Alarcon Sanchez
El Tue, 10 Feb 2026 22:53:39 +0100
Samuel Thibault  escribió:
> Jose Luis Alarcon Sanchez, le mar. 10 févr. 2026 20:41:38 +0100, a
> ecrit:
> > El Tue, 10 Feb 2026 19:23:17 +0100
> > Samuel Thibault  escribió:  
> > > Hello,
> > > 
> > > Jose Luis Alarcon Sanchez, le mar. 10 févr. 2026 15:37:08 +0100, a
> > > ecrit:  
> > > > How is it possible that at the end of last year, from one
> > > > day to the next, the two best web browsers available on
> > > > GNU/Hurd, Luakit and Surf, which are very similar and based on
> > > > the same scripting language, Lua, went from working very
> > > > satisfactorily to instantly terminating when launched, with the
> > > > brief output ‘abort’? 
> > > 
> > > Because of a bug? The abort message would probably be very
> > > informative. Or a gdb backtrace, etc.
> > >   
> > > > Is it possible to investigate why this happened?.
> > > 
> > > Sure. For a simple start, you can check what package got upgraded
> > > in between, and then try to downgrade/upgrade to verify which
> > > exact package pose problems. But also simply actually pasting the
> > > abortion message would provide the list with *way* more
> > > information than just "terminates".  
> > 
> > I don't know how to get a "register" of the upgrades realized at
> > that time (it was last week of 2025, i'm sure of it).  
> 
> It's all in /var/log/dpkg.log
>

Ok. Attached a short extract of /var/log/dpkg.log. I am 100% sure that
the changes that triggered the browsers falilure was made on December
29th. Also including the day before and the day after.
 
> > In the attached screenshot it's collected the proccess of typing
> > 'surf' and 'luakit' in the command line. If i can help in this
> > investigation in any way, let my know, cos i'm eager to help.  
> 
> Ok, it's indeed not informative. But then run it inside gdb and type
> "run" then "bt", having installed the -dbgsym packages as needed to
> get the source code lines, and the abortion backtrace will be
> informative.
> 

I Followed your instructions with gdb, the output is in the attached
file debug_luakit.txt. 

> Samuel

Thanks, Samuel. You're allways willing to help.

Regards.
Jose Luis.

-- 
https://lordofunix.org/ 

Not Registered GNU/Hurd User.
Registered BSD User 51101.
Registered Linux User #213309.
Memories. You are talking about memories.
Rick Deckard. Blade Runner.
2025-12-28 18:35:31 startup archives unpack
2025-12-28 18:35:33 upgrade task-ssh-server:all 3.85 3.86
2025-12-28 18:35:33 status half-configured task-ssh-server:all 3.85
2025-12-28 18:35:33 status unpacked task-ssh-server:all 3.85
2025-12-28 18:35:33 status half-installed task-ssh-server:all 3.85
2025-12-28 18:35:33 status unpacked task-ssh-server:all 3.86
2025-12-28 18:35:33 upgrade tasksel-data:all 3.85 3.86
2025-12-28 18:35:33 status half-configured tasksel-data:all 3.85
2025-12-28 18:35:33 status unpacked tasksel-data:all 3.85
2025-12-28 18:35:33 status half-installed tasksel-data:all 3.85
2025-12-28 18:35:33 status unpacked tasksel-data:all 3.86
2025-12-28 18:35:33 upgrade tasksel:all 3.85 3.86
2025-12-28 18:35:33 status half-configured tasksel:all 3.85
2025-12-28 18:35:33 status unpacked tasksel:all 3.85
2025-12-28 18:35:33 status half-installed tasksel:all 3.85
2025-12-28 18:35:34 status triggers-pending man-db:hurd-amd64 2.13.1-1
2025-12-28 18:35:34 status triggers-pending menu:hurd-amd64 2.1.51
2025-12-28 18:35:34 status triggers-awaited menu:hurd-amd64 2.1.51
2025-12-28 18:35:34 status unpacked tasksel:all 3.86
2025-12-28 18:35:34 upgrade libllvm21:hurd-amd64 1:21.1.8-1 1:21.1.8-1+b1
2025-12-28 18:35:34 status triggers-pending libc-bin:hurd-amd64 2.42-6
2025-12-28 18:35:34 status half-configured libllvm21:hurd-amd64 1:21.1.8-1
2025-12-28 18:35:34 status unpacked libllvm21:hurd-amd64 1:21.1.8-1
2025-12-28 18:35:34 status half-installed libllvm21:hurd-amd64 1:21.1.8-1
2025-12-28 18:35:40 status unpacked libllvm21:hurd-amd64 1:21.1.8-1+b1
2025-12-28 18:35:40 upgrade cargo:hurd-amd64 1.91.1+dfsg1-1~exp3 1.91.1+dfsg1-1
2025-12-28 18:35:40 status half-configured cargo:hurd-amd64 1.91.1+dfsg1-1~exp3
2025-12-28 18:35:40 status unpacked cargo:hurd-amd64 1.91.1+dfsg1-1~exp3
2025-12-28 18:35:40 status half-installed cargo:hurd-amd64 1.91.1+dfsg1-1~exp3
2025-12-28 18:35:41 status unpacked cargo:hurd-amd64 1.91.1+dfsg1-1
2025-12-28 18:35:41 upgrade rustc:hurd-amd64 1.91.1+dfsg1-1~exp3 1.91.1+dfsg1-1
2025-12-28 18:35:41 status half-configured rustc:hurd-amd64 1.91.1+dfsg1-1~exp3
2025-12-28 18:35:41 status unpacked rustc:hurd-amd64 1.91.1+dfsg1-1~exp3
2025-12-28 18:35:41 status half-installed rustc:hurd-amd64 1.91.1+dfsg1-1~exp3
2025-12-28 18:35:42 status unpacked rustc:hurd-amd64 1.91.1+dfsg1-1
2025-12-28 18:35:42 upgrade libstd-rust-dev:hurd-amd64 1.91.1+dfsg1-1~exp3 
1.91.1+dfsg1-1
2025-12-28 18:35:42 status half-configured libstd-rust-dev:hurd-amd64 
1.91.1+dfsg1-1~exp3
2025-12-28 18:35:42 status unpacked libstd-rust-dev:hurd-amd64 
1.91.1+dfsg1-1~exp3
2025-12-28 18:35:42 status half-i

Re: Have any difficulties arisen with the Lua scripting languaje into Hurd?

2026-02-10 Thread Samuel Thibault
Jose Luis Alarcon Sanchez, le mar. 10 févr. 2026 20:41:38 +0100, a ecrit:
> El Tue, 10 Feb 2026 19:23:17 +0100
> Samuel Thibault  escribió:
> > Hello,
> > 
> > Jose Luis Alarcon Sanchez, le mar. 10 févr. 2026 15:37:08 +0100, a
> > ecrit:
> > > How is it possible that at the end of last year, from one
> > > day to the next, the two best web browsers available on GNU/Hurd,
> > > Luakit and Surf, which are very similar and based on the same
> > > scripting language, Lua, went from working very satisfactorily to
> > > instantly terminating when launched, with the brief output ‘abort’?
> > >  
> > 
> > Because of a bug? The abort message would probably be very
> > informative. Or a gdb backtrace, etc.
> > 
> > > Is it possible to investigate why this happened?.  
> > 
> > Sure. For a simple start, you can check what package got upgraded
> > in between, and then try to downgrade/upgrade to verify which exact
> > package pose problems. But also simply actually pasting the abortion
> > message would provide the list with *way* more information than just
> > "terminates".
> 
> I don't know how to get a "register" of the upgrades realized at that
> time (it was last week of 2025, i'm sure of it).

It's all in /var/log/dpkg.log

> In the attached screenshot it's collected the proccess of typing 'surf'
> and 'luakit' in the command line. If i can help in this investigation
> in any way, let my know, cos i'm eager to help.

Ok, it's indeed not informative. But then run it inside gdb and type
"run" then "bt", having installed the -dbgsym packages as needed to get
the source code lines, and the abortion backtrace will be informative.

Samuel



Re: Have any difficulties arisen with the Lua scripting languaje into Hurd?

2026-02-10 Thread Samuel Thibault
Hello,

Jose Luis Alarcon Sanchez, le mar. 10 févr. 2026 15:37:08 +0100, a ecrit:
> How is it possible that at the end of last year, from one
> day to the next, the two best web browsers available on GNU/Hurd, Luakit
> and Surf, which are very similar and based on the same scripting
> language, Lua, went from working very satisfactorily to instantly
> terminating when launched, with the brief output ‘abort’?

Because of a bug? The abort message would probably be very informative.
Or a gdb backtrace, etc.

> Is it possible to investigate why this happened?.

Sure. For a simple start, you can check what package got upgraded
in between, and then try to downgrade/upgrade to verify which exact
package pose problems. But also simply actually pasting the abortion
message would provide the list with *way* more information than just
"terminates".

Samuel