GNU bug report logs - #75462
libtool wrongly reorder compiler deps on linux

Previous Next

To add a comment to this bug, you must first unarchive it, by sending
a message to control AT debbugs.gnu.org, with unarchive 75462 in the body.
You can then email your comments to 75462 AT debbugs.gnu.org in the normal way.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Frederic Berat <fberat <at> redhat.com>
To: bug-libtool <at> gnu.org
Subject: libtool wrongly reorder compiler deps on linux
Date: Thu, 9 Jan 2025 14:11:36 +0100

[Message part 1 (text/plain, inline)]

Hello,

I opened an issue a few years back that didn't get solved:
https://lists.gnu.org/archive/html/libtool/2022-02/msg00001.html

Out of curiosity, I decided to remove my local patch for it with the new
libtool release. It didn't take long for builds to fail again:
https://bugzilla.redhat.com/show_bug.cgi?id=2331361

Considering that more and more systems prevent deduplication over the
years, may it be considered to be the default behavior instead of the
exception ?

Frédéric Bérat

Principal Software Engineer, Platform Tools

Red Hat <https://www.redhat.com>

fberat <at> redhat.com
<https://www.redhat.com>

[Message part 2 (text/html, inline)]

Message #8 received at 75462 <at> debbugs.gnu.org (full text, mbox):

From: Ileana Dumitrescu <ileanadumitrescu95 <at> gmail.com>
To: 75462 <at> debbugs.gnu.org
Cc: Frederic Berat <fberat <at> redhat.com>
Subject: Re: bug#75462: libtool wrongly reorder compiler deps on linux
Date: Thu, 9 Jan 2025 20:56:50 +0200

[Message part 1 (text/plain, inline)]

On 09/01/2025 15:11, Frederic Berat wrote:
> Hello,
> 
> I opened an issue a few years back that didn't get solved: https:// 
> lists.gnu.org/archive/html/libtool/2022-02/msg00001.html <https:// 
> lists.gnu.org/archive/html/libtool/2022-02/msg00001.html>
> 
> Out of curiosity, I decided to remove my local patch for it with the new 
> libtool release. It didn't take long for builds to fail again:
> https://bugzilla.redhat.com/show_bug.cgi?id=2331361 <https:// 
> bugzilla.redhat.com/show_bug.cgi?id=2331361>
> 
> Considering that more and more systems prevent deduplication over the 
> years, may it be considered to be the default behavior instead of the 
> exception ?

To ensure GNU Libtool continues to work with older systems, this patch
should probably not be applied. Users can disable the deduplication
optimization with "--preserve-dup-deps" without applying the suggested
patch, but this affects more than just compiler generated dependencies.

I would be more willing to add an option to toggle
"$opt_duplicate_compiler_generated_deps" separate from
"$opt_preserve_dup_deps". Would this be a good alternative to the patch?

-- 
Ileana Dumitrescu

GPG Public Key: FA26 CA78 4BE1 8892 7F22 B99F 6570 EA01 146F 7354

[OpenPGP_0x6570EA01146F7354.asc (application/pgp-keys, attachment)]

[OpenPGP_signature.asc (application/pgp-signature, attachment)]

Message #11 received at 75462 <at> debbugs.gnu.org (full text, mbox):

From: Frederic Berat <fberat <at> redhat.com>
To: Ileana Dumitrescu <ileanadumitrescu95 <at> gmail.com>
Cc: 75462 <at> debbugs.gnu.org
Subject: Re: bug#75462: libtool wrongly reorder compiler deps on linux
Date: Thu, 9 Jan 2025 22:23:14 +0100

[Message part 1 (text/plain, inline)]

On Thu, Jan 9, 2025 at 7:57 PM Ileana Dumitrescu <
ileanadumitrescu95 <at> gmail.com> wrote:

> On 09/01/2025 15:11, Frederic Berat wrote:
> > Hello,
> >
> > I opened an issue a few years back that didn't get solved: https://
> > lists.gnu.org/archive/html/libtool/2022-02/msg00001.html <https://
> > lists.gnu.org/archive/html/libtool/2022-02/msg00001.html>
> >
> > Out of curiosity, I decided to remove my local patch for it with the new
> > libtool release. It didn't take long for builds to fail again:
> > https://bugzilla.redhat.com/show_bug.cgi?id=2331361 <https://
> > bugzilla.redhat.com/show_bug.cgi?id=2331361>
> >
> > Considering that more and more systems prevent deduplication over the
> > years, may it be considered to be the default behavior instead of the
> > exception ?
>
> To ensure GNU Libtool continues to work with older systems, this patch
> should probably not be applied. Users can disable the deduplication
> optimization with "--preserve-dup-deps" without applying the suggested
> patch, but this affects more than just compiler generated dependencies.
>
>
Well, I'm not sure I get why keeping duplicates would harm, but fair enough
my patch is clearly overkill.


> I would be more willing to add an option to toggle
> "$opt_duplicate_compiler_generated_deps" separate from
> "$opt_preserve_dup_deps". Would this be a good alternative to the patch?
>

Not sure that's a good alternative. you'd force users from the systems that
have been excluded over the years to toggle an option that they didn't need
so far.


>
> --
> Ileana Dumitrescu
>
> GPG Public Key: FA26 CA78 4BE1 8892 7F22 B99F 6570 EA01 146F 7354
>
>

[Message part 2 (text/html, inline)]

Message #14 received at 75462 <at> debbugs.gnu.org (full text, mbox):

From: Ileana Dumitrescu <ileanadumitrescu95 <at> gmail.com>
To: 75462 <at> debbugs.gnu.org
Cc: Frederic Berat <fberat <at> redhat.com>
Subject: Re: bug#75462: libtool wrongly reorder compiler deps on linux
Date: Fri, 10 Jan 2025 16:36:15 +0200

[Message part 1 (text/plain, inline)]

On 09/01/2025 23:23, Frederic Berat wrote:
...
>     To ensure GNU Libtool continues to work with older systems, this patch
>     should probably not be applied. Users can disable the deduplication
>     optimization with "--preserve-dup-deps" without applying the suggested
>     patch, but this affects more than just compiler generated dependencies.
> 
> 
> Well, I'm not sure I get why keeping duplicates would harm, but fair 
> enough my patch is clearly overkill.

I am willing to apply something similar if it is specific to systems
that have been verified to have a problem with reordered compiler
dependencies. I do not want to assume all of linux systems require this
patch, but a more specific subset would be good.

-- 
Ileana Dumitrescu

GPG Public Key: FA26 CA78 4BE1 8892 7F22 B99F 6570 EA01 146F 7354

[OpenPGP_0x6570EA01146F7354.asc (application/pgp-keys, attachment)]

[OpenPGP_signature.asc (application/pgp-signature, attachment)]

Message #17 received at 75462 <at> debbugs.gnu.org (full text, mbox):

From: Frederic Berat <fberat <at> redhat.com>
To: Ileana Dumitrescu <ileanadumitrescu95 <at> gmail.com>
Cc: 75462 <at> debbugs.gnu.org
Subject: Re: bug#75462: libtool wrongly reorder compiler deps on linux
Date: Mon, 13 Jan 2025 15:29:23 +0100

[Message part 1 (text/plain, inline)]

On Fri, Jan 10, 2025 at 3:36 PM Ileana Dumitrescu <
ileanadumitrescu95 <at> gmail.com> wrote:

> On 09/01/2025 23:23, Frederic Berat wrote:
> ...
> >     To ensure GNU Libtool continues to work with older systems, this
> patch
> >     should probably not be applied. Users can disable the deduplication
> >     optimization with "--preserve-dup-deps" without applying the
> suggested
> >     patch, but this affects more than just compiler generated
> dependencies.
> >
> >
> > Well, I'm not sure I get why keeping duplicates would harm, but fair
> > enough my patch is clearly overkill.
>
> I am willing to apply something similar if it is specific to systems
> that have been verified to have a problem with reordered compiler
> dependencies. I do not want to assume all of linux systems require this
> patch, but a more specific subset would be good.
>

More than a specific linux system, that are dependencies that include GCC
libraries that should not be re-ordered/de-duplicated in the case I've seen.


>
> --
> Ileana Dumitrescu
>
> GPG Public Key: FA26 CA78 4BE1 8892 7F22 B99F 6570 EA01 146F 7354
>
>

[Message part 2 (text/html, inline)]

Message #20 received at 75462 <at> debbugs.gnu.org (full text, mbox):

From: Ileana Dumitrescu <ileanadumitrescu95 <at> gmail.com>
To: 75462 <at> debbugs.gnu.org
Cc: Frederic Berat <fberat <at> redhat.com>
Subject: Re: bug#75462: libtool wrongly reorder compiler deps on linux
Date: Mon, 13 Jan 2025 19:44:48 +0200

[Message part 1 (text/plain, inline)]

On 13/01/2025 16:29, Frederic Berat wrote:
...
>     I am willing to apply something similar if it is specific to systems
>     that have been verified to have a problem with reordered compiler
>     dependencies. I do not want to assume all of linux systems require this
>     patch, but a more specific subset would be good.
> 
> 
> More than a specific linux system, that are dependencies that include 
> GCC libraries that should not be re-ordered/de-duplicated in the case 
> I've seen.

After some more thought on this, I will merge the suggested patch. If
there are any unlikely failures or performance issues, I will work to
find a better solution then, since the current ideas seem like a mess
to maintain.

-- 
Ileana Dumitrescu

GPG Public Key: FA26 CA78 4BE1 8892 7F22 B99F 6570 EA01 146F 7354

[OpenPGP_0x6570EA01146F7354.asc (application/pgp-keys, attachment)]

[OpenPGP_signature.asc (application/pgp-signature, attachment)]

Message #25 received at 75462-done <at> debbugs.gnu.org (full text, mbox):

From: Ileana Dumitrescu <ileanadumitrescu95 <at> gmail.com>
To: 75462-done <at> debbugs.gnu.org
Cc: Frederic Berat <fberat <at> redhat.com>
Subject: Re: bug#75462: libtool wrongly reorder compiler deps on linux
Date: Mon, 13 Jan 2025 19:56:35 +0200

[Message part 1 (text/plain, inline)]

Thank you again for the patch [1].

[1] 
https://git.savannah.gnu.org/cgit/libtool.git/commit/?h=development&id=18dc06ade3db7c4100e781d1c1ef8c5b018ca920

-- 
Ileana Dumitrescu

GPG Public Key: FA26 CA78 4BE1 8892 7F22 B99F 6570 EA01 146F 7354

[OpenPGP_0x6570EA01146F7354.asc (application/pgp-keys, attachment)]

[OpenPGP_signature.asc (application/pgp-signature, attachment)]

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.