xapi-project
/
xen-api
/
23233711099
80%
master: 80%
master: 80%
|
Ran
|
Jobs
1
|
Files
34
|
Run time
1min
|
Badge
Embed ▾
README BADGES
|
coverage: 80.459%. Remained the same
push
github
web-flowCP-311892: optimize Pervasiveext.finally (#6954) The important part of this optimization was merged in f642b347a, which optimized `Mutex.execute`. I also had an optimization for `Pervasiveext.finally`, but initially it didn't seem like the improvement is significant. I wrote a better benchmark now, and it is: performance can be improved 55x on the error path. Previously I didn't call `Debug.with_thread_associated` on the benchmark program's startup, which disabled most of the `xapi-backtrace` code. When it is active, then `Backtrace.is_important` becomes a lot slower. Optimizing that would be the topic of another PR, meanwhile with a very small change (2 lines) we can improve the performance of `finally`. (Usually code shouldn't be raising exceptions all the time, but we've got quite a few places in the code that raise `Not_found` and equivalents). The optimization passed a BVT+BST, and I tested that backtraces still work.
3504 of 4355 relevant lines covered (80.46%)
0.8 hits per line
Jobs
| ID | Job ID | Ran | Files | Coverage | |
|---|---|---|---|---|---|
| 1 | python3.11 - 23233711099.1 | 34 |
80.46 |
GitHub Action Run |
