[Debian-med-packaging] Bug#1152371: oscar: unregistered vendor copy of botan
Sergio Durigan Junior
sergiodj at sergiodj.net
Sun Nov 8 03:40:12 GMT 2026
On Monday, 02 November 2026, Helmut Grohne wrote:
> Hi Sergio,
Hi Helmut,
> On Sun, Nov 01, 2026 at 09:31:27PM -0500, Sergio Durigan Junior wrote:
>> I don't vehemently disagree with you, but if upstream chose to keep
>> -Werror in the release then I tend to follow their lead here. For
>> example, OSCAR has this snippet in their .pro file:
>>
>> --8<---------------cut here---------------start------------->8---
>> # Always treat warnings as errors, even (especially!) in release
>> QMAKE_CFLAGS += -Werror
>> QMAKE_CXXFLAGS += -Werror
>> --8<---------------cut here---------------end--------------->8---
>>
>> So using -Werror in the release build was intentional. I understand we
>> may see FTBFSes due to that, but I'm personally willing to go through
>> the trouble of reporting problems upstream in this case.
>
> To me, that's precisely the problem. Keeping -Werror in your CI allows
> you to go through the trouble of handling those FTBFSes. Enabling it in
> the package makes *others* go through that trouble.
That's fair. I'll put in my TODO list to disable -Werror in the next
upload.
>> Ah, right. I thought this was a problem with OSCAR itself. In this
>> case, let's keep things in this bug.
>
>> It wasn't as hard as I was thinking, although TBH I haven't checked if
>> upstream is patching their amalgamation of botan (I don't think so).
>>
>> I've just uploaded the package to experimental; feel free to take a look
>> at the patch and let me know. I will probably give it a week before
>> uploading to unstable, unless upstream agrees with my changes.
>
> I spent a brief look and it mostly looks fine. In principle, this could
> be upstreamable if changing two more aspects. I guess upstream would
> want to keep the vendor copy, so adding a conditional for selecting the
> vendored vs system one would likely be required. The other aspect is
> using pkg-config to discover the flags for botan. Maybe you can try this
> snippet?
>
> CONFIG += link_pkgconfig
> PKGCONFIG += botan-3
Yes, that works, thanks. I wondered what it would take to make the
build use the pkgconfig file, but honestly I didn't have more time to
investigate.
>> Oh, I haven't unvendored the existing botan either. We can do that
>> later.
>
> I don't think there is much point in repacking the source to remove the
> vendored botan so long as it is distributable (and it is). The problem
> to solve here is not copying vulnerabilities and that is achieved by
> linking the system version. To me, your experimental upload fully
> addresses the reported problem.
Upstream released version 2.1.1 in the meantime, so I packaged that and
took the opportunity to repack the tarball and remove the vendored
botan. I've just uploaded the new package with the changes.
> How about also looking at bookworm and trixie? Unvendoring botan there
> probably is not appropriate. But maybe updating the
> debian-security-support package to call oscar unsupported in those
> releases would be worthwhile?
>
> See for example
> https://sources.debian.org/src/debian-security-support/1%3A14%2B2026.05.07/security-support.deb13
That's a good idea. As I said in my previous message my Debian time is
pretty limited nowadays, so I don't know when I'll be able to get to
this. Feel free to do it if you're in a hurry.
Thanks,
--
Sergio
GPG key ID: 237A 54B1 0287 28BF 00EF 31F4 D0EB 7628 65FC 5E36
Please send encrypted e-mail if possible
https://sergiodj.net/
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 865 bytes
Desc: not available
URL: <http://alioth-lists.debian.net/pipermail/debian-med-packaging/attachments/20261107/4cc8b150/attachment.sig>
More information about the Debian-med-packaging
mailing list