try to fix brotli test

[neheb]

No description provided.

[ghost]

👇 Click on the image for a new way to code review

Legend

[neheb]

ping @Biswa96

seems to be a regression in at least the UCRT package.

[Biswa96]

Which test are you asking about? Multiple tests failed here.

[neheb]

Look at other recent PRs

[Biswa96]

This PR fixes the test in ucrt64. But I can not figure out who is wrong here - exiv2 or brotli. Does any CI job use brotli 1.1.0 ? @kmilos may provide some info.

[neheb]

CL_SPACE should be the correct value. I don't know if this is an msys bug or upstream.

[kmilos]

Does any CI job use brotli 1.1.0 ?

It does for the UCRT64 job: https://github.com/Exiv2/exiv2/actions/runs/6140387211/job/16659286923#step:3:195

The Cygwin one is still on 1.0.9: https://github.com/Exiv2/exiv2/actions/runs/6140387211/job/16659286365#step:4:155

[kmilos]

Looks like this failed on Linux as well in #2762?

[neheb]

So upstream bug.

[kmilos]

So upstream bug.

Pinged upstream, looks like this is expected behaviour now from 1.1.0...

So we need to either detect the brotli version, and/or check both return values 😞