Closed ni4 closed 3 years ago
Merging #69 (ba3fd68) into master (fdee8c9) will decrease coverage by
0.02%
. The diff coverage is66.66%
.
@@ Coverage Diff @@
## master #69 +/- ##
==========================================
- Coverage 98.94% 98.91% -0.03%
==========================================
Files 41 41
Lines 3325 3327 +2
==========================================
+ Hits 3290 3291 +1
- Misses 35 36 +1
Impacted Files | Coverage Δ | |
---|---|---|
lib/rnp/misc.rb | 100.00% <ø> (ø) |
|
spec/key/properties/4BE147BB22DF1E60_spec.rb | 98.61% <66.66%> (-1.39%) |
:arrow_down: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update fdee8c9...ba3fd68. Read the comment docs.
Ping @ronaldtse @dewyatt : could we merge this despite the CI failures? Or, what should be a workflow when some changes in rnp require tests here to be updated?
Ping @ronaldtse @dewyatt : could we merge this despite the CI failures? Or, what should be a workflow when some changes in rnp require tests here to be updated?
I'll take a closer look later but at a glance the test looks to be specific to the version of RNP and should be made compatible with all supported versions. As-is it looks like it would break support for all other versions of RNP.
I'll take a closer look later but at a glance the test looks to be specific to the version of RNP and should be made compatible with all supported versions. As-is it looks like it would break support for all other versions of RNP.
Thanks, @dewyatt . I updated the PR with version checks so it should work with every RNP version. However tests seems to behave weird, let's see whether some restarts would help.
Sorry I'm dragging my feet on this but I'll get back to it in the next few days here
@dewyatt Looks like we now have at least two issues:
I'll update the PR with latest timestamp check, but what would be a correct way to fix Botan's build? Use GCC 9.0? Report error to their tracker? Both :) ?
@dewyatt Looks like we now have at least two issues:
- latest Botan 3.0 alpha uses flag -Wextra-semi, which is available only since GCC 9.0, while 7.5 is used.
- there were some commits to the RNP master, so "LibRnp::rnp_version_commit_timestamp >= 1604156729" is not good check anymore.
I'll update the PR with latest timestamp check, but what would be a correct way to fix Botan's build? Use GCC 9.0? Report error to their tracker? Both :) ?
It's probably fine to just test against the latest release of botan and not master, which would take care of that issue.
Yeah typically what I would do here is merge the corresponding rnp PR first (with CI broken -- test manually) and then use that commit timestamp.
Also the version thing is something I would rather see in FEATURES
(which is maybe better named quirks) in lib/misc.rb
.
@dewyatt Thanks, updated PR. Does it look good now (except that we'll need to update the timestamp once https://github.com/rnpgp/rnp/pull/1337 is merged)?
@ronaldtse All tests except codecov are passing now, so let's merge this? P.S. Btw, where did we lose codecov for the rnp? CC @dewyatt
P.S. Btw, where did we lose codecov for the rnp?
@ni4 not sure why it's gone -- it's still enabled in Settings. Maybe we need this to make it work? https://github.com/codecov/codecov-action
This PR updates spec to be up-to-date with behavior changes from https://github.com/rnpgp/rnp/pull/1337 As it updates tests only, I don't think we should add some checks for rnp versions and so on. However, if there is a better approach all suggestions are appreciated.