Santiago Vila, le mer. 04 juin 2025 11:35:19 +0200, a ecrit:

# sorry for the previous broken utf8
retitle 1104835 edbrowse: FTBFS: jseng-quick.c:1493:34: error: implicit declaration of function '€˜JS_VALUE_GET_OBJ'€™
reassign 1104835 src:edbrowse
fixed 1104835 3.8.12-1
found 1104835 3.8.10-1
affects 1104835 =
thanks

El 4/6/25 a las 10:44, Samuel Thibault escribió:

reassign 1104835 edbrowse
close 1104835 3.8.12-1

Hi. Can you explain that a little bit, please?

Was this maybe some incompatibility between edbrowse 3.8.10-1
and the libquickjs in unstable which was fixed in edbrowse 3.8.12-1?

Yes, that is it, sorry for not making it explicit.

I had initially reassigned to quickjs, to make sure it doesn't migrate
to testing (thus breaking the edbrowse build) without the new version of edbrowse. But now that edbrowse is in unstable, it's indeed the new
version of edbrowse which actually fixes the build issue.

Samuel

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

# sorry for the previous broken utf8
retitle 1104835 edbrowse: FTBFS: jseng-quick.c:1493:34: error: implicit declaration of function '€˜JS_VALUE_GET_OBJ'€™
reassign 1104835 src:edbrowse
fixed 1104835 3.8.12-1
found 1104835 3.8.10-1
affects 1104835 =
thanks

El 4/6/25 a las 10:44, Samuel Thibault escribió:

reassign 1104835 edbrowse
close 1104835 3.8.12-1

Hi. Can you explain that a little bit, please?

Was this maybe some incompatibility between edbrowse 3.8.10-1
and the libquickjs in unstable which was fixed in edbrowse 3.8.12-1?

(My commands above try to be consistent with such theory. I hope
they are correct)

Thanks.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi,

On Wed, Jun 04, 2025 at 11:38:34AM +0200, Samuel Thibault wrote:

Santiago Vila, le mer. 04 juin 2025 11:35:19 +0200, a ecrit:

# sorry for the previous broken utf8
retitle 1104835 edbrowse: FTBFS: jseng-quick.c:1493:34: error: implicit declaration of function '€˜JS_VALUE_GET_OBJ'€™
reassign 1104835 src:edbrowse
fixed 1104835 3.8.12-1
found 1104835 3.8.10-1
affects 1104835 =
thanks

El 4/6/25 a las 10:44, Samuel Thibault escribió:

reassign 1104835 edbrowse
close 1104835 3.8.12-1

Hi. Can you explain that a little bit, please?

Was this maybe some incompatibility between edbrowse 3.8.10-1
and the libquickjs in unstable which was fixed in edbrowse 3.8.12-1?

Yes, that is it, sorry for not making it explicit.

I had initially reassigned to quickjs, to make sure it doesn't migrate
to testing (thus breaking the edbrowse build) without the new version of edbrowse. But now that edbrowse is in unstable, it's indeed the new
version of edbrowse which actually fixes the build issue.

Can you confirm that the build issue only happens with the version of quickjs currently in unstable?

That would mean that this bug doesn't affect edbrowse in testing.

The reproducible build history seems to suggest that this is the case, as 3.8.10-1 built fine in recent tests in trixie, but not in unstable.

https://tests.reproducible-builds.org/debian/history/edbrowse.html

Thanks,

Ivo

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Ivo De Decker, le mer. 04 juin 2025 10:27:08 +0000, a ecrit:

On Wed, Jun 04, 2025 at 11:38:34AM +0200, Samuel Thibault wrote:

Santiago Vila, le mer. 04 juin 2025 11:35:19 +0200, a ecrit:

# sorry for the previous broken utf8
retitle 1104835 edbrowse: FTBFS: jseng-quick.c:1493:34: error: implicit declaration of function '€˜JS_VALUE_GET_OBJ'€™
reassign 1104835 src:edbrowse
fixed 1104835 3.8.12-1
found 1104835 3.8.10-1
affects 1104835 =
thanks

El 4/6/25 a las 10:44, Samuel Thibault escribió:

reassign 1104835 edbrowse
close 1104835 3.8.12-1

Hi. Can you explain that a little bit, please?

Was this maybe some incompatibility between edbrowse 3.8.10-1
and the libquickjs in unstable which was fixed in edbrowse 3.8.12-1?

Yes, that is it, sorry for not making it explicit.

I had initially reassigned to quickjs, to make sure it doesn't migrate
to testing (thus breaking the edbrowse build) without the new version of edbrowse. But now that edbrowse is in unstable, it's indeed the new
version of edbrowse which actually fixes the build issue.

Can you confirm that the build issue only happens with the version of quickjs currently in unstable?

Yes, that is it.

That would mean that this bug doesn't affect edbrowse in testing.

When built against quickjs from testing, edbrowse from testing builds
fine, yes.

Samuel

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Hi,

On Wed, Jun 04, 2025 at 12:38:49PM +0200, Santiago Vila wrote:

I confirm that the package builds ok in both testing and unstable right now.

The problem would happen if the version of quickjs in unstable propagates
to testing before the updated edbrowse.

OK. That doesn't look likely, because the diff doesn't look like something
that would be unblocked.

It might also be useful to investigate adding an autopackagetest to edbrowse
to catch these type of changes (but that would probably be someting for after the freeze).

Not sure how to express that using BTS metadata, I hope somebody else
knows how to do that.

Good question :)

I don't know a good way to tell version tracking that the version in testing isn't affected, while the version number of that package is affected by the bug. The only workaround I can come up with is to remove the version tracking info entirely from this bug. As the bug is closed, this means the BTS now thinks it doesn't affect any suite, which is correct.

For people wondering: no, removing the 'trixie' tag is not the correct way to do this. Having a bug tagged 'sid', but not with the tag for the current testing will give unwanted and unexpected results.

Thanks,

Ivo

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Ivo De Decker, le mer. 04 juin 2025 11:47:06 +0000, a ecrit:

On Wed, Jun 04, 2025 at 12:38:49PM +0200, Santiago Vila wrote:

I confirm that the package builds ok in both testing and unstable right now.

The problem would happen if the version of quickjs in unstable propagates to testing before the updated edbrowse.

OK. That doesn't look likely, because the diff doesn't look like something that would be unblocked.

As discussed in #1104748, it's not only about the diff size, but about
being able to cope with security fixes on the long run.

It might also be useful to investigate adding an autopackagetest to edbrowse to catch these type of changes

It's not a runtime issue, but a build-time issue. quickjs does not
actually ship a shared library.

Samuel

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)