Closed kyrofa closed 4 years ago
Merging #194 into master will increase coverage by
4.95%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## master #194 +/- ##
==========================================
+ Coverage 55.38% 60.34% +4.95%
==========================================
Files 17 17
Lines 585 585
Branches 52 52
==========================================
+ Hits 324 353 +29
+ Misses 247 218 -29
Partials 14 14
Flag | Coverage Δ | |
---|---|---|
#unittests | 60.34% <100.00%> (+4.95%) |
:arrow_up: |
Impacted Files | Coverage Δ | |
---|---|---|
sros2/sros2/api/_permission.py | 92.30% <100.00%> (+30.76%) |
:arrow_up: |
sros2/sros2/verb/create_permission.py | 68.00% <0.00%> (+68.00%) |
:arrow_up: |
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 1b41a2a...e7156ce. Read the comment docs.
It changes it to a... segfault? What the heck? Any chance that's Focal/SSL-related? I just ran ros2 security create_permission
for the basic talker/listener example and got AttributeError
s without this patch, and success with it (on Bionic).
I just ran ros2 security create_permission for the basic talker/listener example and got AttributeErrors without this patch, and success with it.
I confirmed I have a segfault on Focal with the repo state before all the refactoring so not related to this PR. Not great but at least clears this ...
As mentionned offline, committing a simple test along this PR would be great, full coverage can be done at another time.
committing a simple test along this PR would be great, full coverage can be done at another time.
Done. Couldn't let codecov own me like that. I'm afraid that I rebased instead of merged though, I'm sorry. At least the diff is tiny.
Later on, it might be easier to just compare the permission.xml tree to a reference xml file, and just pre-processes steps to remove/sanitize known diffs, like <validity>
, or <domains>
matches the ros domain env, or exceptional rules for logging and discovery. Once you filter out those tags, the rest of the tree for the _test_identity
grant name your checking should be the same when pretty printed.
Proof that we're still completely missing test coverage of some verbs (logged as #193) :cry: . Added some here.