coverage: count number of executions per line #1265
Conversation
|
I think either the workers will be frequently sending events (quite bad) or some counter will not be properly updated (not so bad, still not great). |
|
This is awesome, I want this feature so bad 😆 |
elopez
force-pushed
the
dev-cov-exec-counts
branch
from
6c0a717
to
97007bc
Compare
|
I did a second attempt at implementing this a safer way. The code is quite ugly right now but if anyone wants to test/review it before I spend some more time on it I'd appreciate it 👍 |
|
I don't see this branch under Actions to download the binary 🤔 |
|
I think it is because it is a draft, let me convert it a proper PR |
elopez
force-pushed
the
dev-cov-exec-counts
branch
from
97007bc
to
559bbc1
Compare
|
@rappie There was a small git conflict, hopefully it builds now that I resolved it 👍 |
elopez
force-pushed
the
dev-cov-exec-counts
branch
from
0b5d6a0
to
8d2cb5e
Compare
|
@samalws-tob can you take a look to this PR? |
| @@ -304,13 +314,13 @@ execTxWithCov tx = do | |||
| -- bug in another place, investigate. | |||
| -- ... this should be fixed now, since we use `codeContract` instead | |||
| -- of `contract` for everything; it may be safe to remove this check. | |||
| when (pc < VMut.length vec) $ | |||
| when (pc < VMut.length vec) $ do | |||
| VMut.modify (fromJust maybeStatsVec) (\(execQty, revertQty) -> (execQty + 1, revertQty)) opIx | |||
There was a problem hiding this comment.
I don't see anywhere where we increment revertQty
There was a problem hiding this comment.
yeah, I have not implemented counting of reverts. I meant to do it at first but then it wasn't clear to me where I should be counting them. If you have any pointers there I'd appreciate them, otherwise I may remove the extra var at this time.
elopez
force-pushed
the
dev-cov-exec-counts
branch
from
839d2d3
to
924dfb2
Compare
Revert counting is still not implemented.
No description provided.