Hi,
The libzstd package in testing is currently missing a couple of
build-time tests and a fix for a very rare data corruption bug.
Both of these have been fixed in libstd-1.5.4+dfsg2-5 in unstable;
however, I feel like I cannot ask for an unblock from the release
managers since the sparc64 buildd started failing on this package
at some point in February:
https://buildd.debian.org/status/logs.php?pkg=libzstd&arch=sparc64
Now, the -1, -2, and -4 failures I can explain: there were some
problems in the upstream test suite that were fixed in -3 and -5.
However, -3 and -5 should really not have failed: they built
successfully on all other architectures where the build dependencies
were installable, and they also passed autopkgtest runs:
https://ci.debian.net/packages/libz/libzstd/
I set up a qemu-based sbuild environment on my laptop, and
I built the libzstd package successfully. Is it possible that
something is wrong with the sparc64 buildd? Could somebody with
an actual sparc64 box try to build libstd-1.5.4+dfsg2-5 and
let me know if it works for them?
Thanks in advance!
Hi Peter,
On 06-04-2023 15:37, Peter Pentchev wrote:
I feel like I cannot ask for an unblock from the release
managers since the sparc64 buildd started failing on this package
at some point in February:
sparc64 is not a release architecture. sparc64 will not be better or worse
if something migrates. I suggest to consider those problem separately.
Sysop: | Keyop |
---|---|
Location: | Huddersfield, West Yorkshire, UK |
Users: | 429 |
Nodes: | 16 (2 / 14) |
Uptime: | 115:53:38 |
Calls: | 9,056 |
Calls today: | 3 |
Files: | 13,395 |
Messages: | 6,016,442 |