Make sure to actually wrap the action in 'wrapMonadClientSem'

[isovector]

Looking through #1781, I noticed we unintentionally changed the logic with respect to how exceptions are handled. We forgot to do any wrapping in wrapMonadClientSem!

Checklist

• The PR Title explains the impact of the change.
• The PR description provides context as to why the change should occur and what the code contributes to that effect. This could also be a link to a JIRA ticket or a Github issue, if there is one.
• changelog.d contains the following bits of information:
  • A file with the changelog entry in one or more suitable sub-sections. The sub-sections are marked by directories inside changelog.d.

[fisx]

I think this is just translating one internal server error into another one, but the idea was to keep the behavior completely indistinguishable, so, good catch!

[isovector]

@fisx good to merge, but apparently I don't have the right bits to do it myself.

[fisx]

all tests passed locally.

[jschaul]

@isovector @fisx please stop merging PRs until CI is green; #1781 introduced a new system-level dependency which resulted in breaking all the CI tests for everyone when it was merged:

docker run -it --entrypoint sh quay.io/wire/spar-schema:latest
 # ldd /usr/bin/spar-schema
	/lib/ld-musl-x86_64.so.1 (0x79110c8fc000)
Error loading shared library libncursesw.so.6: No such file or directory (needed by /usr/bin/spar-schema)
	libxml2.so.2 => /usr/lib/libxml2.so.2 (0x79110c7d3000)
	libstdc++.so.6 => /usr/lib/libstdc++.so.6 (0x79110c63a000)
	libz.so.1 => /lib/libz.so.1 (0x79110c620000)
	libssl.so.1.1 => /lib/libssl.so.1.1 (0x79110c59f000)
	libcrypto.so.1.1 => /lib/libcrypto.so.1.1 (0x79110c320000)
	libgmp.so.10 => /usr/lib/libgmp.so.10 (0x79110c2b9000)
	libc.musl-x86_64.so.1 => /lib/ld-musl-x86_64.so.1 (0x79110c8fc000)
	libgcc_s.so.1 => /usr/lib/libgcc_s.so.1 (0x79110c2a5000)
	liblzma.so.5 => /usr/lib/liblzma.so.5 (0x79110c28

all tests passed locally.

That's not entirely correct, ./deploy/dockerephemeral/run.sh fails with the same issue.

[pcapriotti]

That's not entirely correct, ./deploy/dockerephemeral/run.sh fails with the same issue.

To be fair, if you have the library installed locally and you were already running the ephemeral services in the background, you wouldn't notice the issue.

But I agree, we shouldn't merge when CI is red.

[fisx]

Sorry, will look into it now.

I my defense (@isovector had nothing to do with this): I saw and error that I didn't understand, but that did look like something related to concourse or docker trouble, which has happened in the past. When I did a rerun and tried to look at the logs again, concourse wouldn't show them (page wouldn't load). That's when I decided to take the risk.