Closed
Bug 1496286
Opened 6 years ago
Closed 6 years ago
XML coverage is missing in grcov data
Categories
(Testing :: Code Coverage, defect)
Testing
Code Coverage
Tracking
(Not tracked)
RESOLVED
INVALID
People
(Reporter: sparky, Unassigned)
References
(Blocks 1 open bug)
Details
Looking at the data from the description attachment in bug 1496285, specifically in this push and task: https://treeherder.mozilla.org/#/jobs?repo=try&revision=f90b8cdb2cefb9e39f37db54415645df0055ae69&selectedJob=201089393 The XML data exists in the raw JSVM data, but I can't find it in the GRCOV data. So it seems like it is being lost when it is processed by grcov: https://dxr.mozilla.org/mozilla-central/source/testing/mozharness/mozharness/mozilla/testing/codecoverage.py#314 This data also can't be found in activedata.
Comment 1•6 years ago
|
||
There seems to be no JS data at all in the grcov output file, not only the XML data is missing.
Comment 2•6 years ago
|
||
Ah, this is actually expected. We are not parsing the JSVM data with grcov. In normal coverage builds, we are uploading both the jsvm and the grcov artifacts. In per-test coverage builds we are merging them.
Status: NEW → RESOLVED
Closed: 6 years ago
Resolution: --- → INVALID
Reporter | ||
Comment 3•6 years ago
|
||
Is there a reason for not merging? I think we should be merging them since we are missing the coverage in AD.
Comment 4•6 years ago
|
||
AD should be parsing both artifacts, isn't it? The reason for not merging is that it's more expensive for no benefit (grcov is going to parse the jsvm file simply to output it again without any modification).
Flags: needinfo?(klahnakoski)
Reporter | ||
Comment 5•6 years ago
|
||
Oh ok. :marco, I looked at the code for the etl and it looks like it should be ingested: https://github.com/klahnakoski/ActiveData-ETL/blob/e154cb609eeb99c4490e6fbcb63c9206e43bdc15/activedata_etl/transforms/cov_to_es.py#L106 And here's the function that handles JSVM specifically: https://github.com/klahnakoski/ActiveData-ETL/blob/dev/activedata_etl/transforms/jsvm_to_es.py#L30 :ekyle, any ideas why we would be missing the data in AD? It sounds like it's failing to ingest them properly.
Comment 6•6 years ago
|
||
Looking right now...
Comment 7•6 years ago
|
||
I confirmed both files will get ingested. If they are missing, then it is the same bug that appears to be dropping coverage. http://54.149.21.8/tools/query.html#query_id=P28lFRbX
Depends on: 1496521
Flags: needinfo?(klahnakoski)
Reporter | ||
Comment 8•6 years ago
|
||
Thanks ekyle, they have been ingested now and I can see them in the query. So it is caused by bug 1496521, and there is a partial solution to this in bug 1384643.
Updated•6 years ago
|
Version: Version 3 → Trunk
You need to log in
before you can comment on or make changes to this bug.
Description
•