[Pkg-rust-maintainers] Bug#1055534: sq-wot - should the binary be dropped.

Peter Green plugwash at debian.org
Tue Nov 7 22:03:40 GMT 2023


Package: sq-wot
x-debbugs-cc: alexander.kjall at gmail.com, dkg at fifthhorseman.net

Since it seems we now (fingers crossed) finally have a version of
ring that should build on all release architectures I decided it was
time to fix sequoia-sq which has been languishing in a rc buggy
state since soon after the bookworm release. As part of doing this
I looked at sequoia-wot.

While checking what was going on, I discovered that the sq-wot binary
had been dropped in a git commit by Alexander Kjäll (capitol), even
more strangely the same commit that disabled building the binaries
added manpages for them.

I discussed this with capitol on irc

<plugwash> capitol, can you explain e2c9d5396d09ed7d3105dfd2cdf2ac4fbf594926
<plugwash> because it really doesn't make much sense to me
<plugwash> It seems to add a bunch of manpages for the binaries, yet disable actually building/packaging them!
<plugwash> and none of the (pretty substantial) changes other than the new upstream release are documented in the changelog.
<plugwash> (sorry if it wasn't clear in my first message, the hex string was a debcargo-conf commit ID https://salsa.debian.org/rust-team/debcargo-conf/-/commit/e2c9d5396d09ed7d3105dfd2cdf2ac4fbf594926 )
* hiddenman_ (~Andrew at broadband-46-242-8-154.ip.moscow.rt.ru) has joined
<capitol> plugwash: i followed the pattern of a couple of other binaries with man pages here, to add the man pages to the package, and add a <binary-name>.manpages file to get them into the correct place
<capitol> i have a memory of checking that they got into the package, but let me verify that again
<capitol> and sorry for not documenting this in the changelog, i'll make an effort on being more verbose in that
<plugwash> +bin = false
<plugwash>  overlay = "."
<plugwash> -uploaders = ["Daniel Kahn Gillmor <dkg at fifthhorseman.net>"]
<plugwash> -bin_name = "sq-wot"
<plugwash> +uploaders = ["Daniel Kahn Gillmor <dkg at fifthhorseman.net>", "Alexander Kjäll <alexander.kjall at gmail.com>"]
<plugwash> that's the bit I don't get, you seem to be disabling building of binaries at all.
<capitol> ohh...
<capitol> yes, now i realize, the sq-wot binary have been deprecated from sequoia, and replace with a 'wot' subcommand to the sq binary
<capitol> and since the original package had only gone to experimental i thought it best to not get it into debian
<capitol> i should have cleaned up the man pages, those where done before i talked to sequoia about it and they asked it to not be packaged
<capitol> and now the package seem to have bit-rotted and can't be built anymore, due to usage of a deprecated function, I can look at that tomorrow
<plugwash> umm
<plugwash> sq-wot is in stable, testing and unstable
<capitol> i must remember this totally wrong :/

After this conversation I decided to revert the change dropping the binary
package for now.

I don't object to removing the binary package in principle if it is clear
that people fully understand the situation and have decided that is the
best way forward but equally i'm not going to upload a version including
said removal until/unless it is clear to me that there is said consensus.
Especially given that while removing a binary package is easy, adding one
back requires dealing with the random delay and nitpicking of the NEW
queue.

I would particularly like dkg, as the person who initially packaged
sequoia-wot a year and a half ago to weigh in on this.



More information about the Pkg-rust-maintainers mailing list