Closed JKRT closed 2 years ago
Merging #54 (0f88f75) into master (e45d452) will increase coverage by
12.55%
. The diff coverage is50.67%
.
@@ Coverage Diff @@
## master #54 +/- ##
===========================================
+ Coverage 2.77% 15.32% +12.55%
===========================================
Files 164 166 +2
Lines 61104 59128 -1976
===========================================
+ Hits 1693 9061 +7368
+ Misses 59411 50067 -9344
Impacted Files | Coverage Δ | |
---|---|---|
src/NewFrontend/BindingExpression.jl | 19.27% <0.00%> (+19.27%) |
:arrow_up: |
src/NewFrontend/NFBuiltin.jl | 0.00% <ø> (ø) |
|
src/NewFrontend/NFCeval.jl | 6.84% <0.00%> (+6.84%) |
:arrow_up: |
src/NewFrontend/NFClass.jl | 35.68% <0.00%> (+35.68%) |
:arrow_up: |
src/NewFrontend/NFComponent.jl | 35.19% <0.00%> (+35.19%) |
:arrow_up: |
src/NewFrontend/NFEvalFunction.jl | 0.00% <0.00%> (ø) |
|
src/NewFrontend/NFExpandableConnectors.jl | 8.53% <0.00%> (+8.13%) |
:arrow_up: |
src/NewFrontend/NFFlatten.jl | 35.78% <0.00%> (+35.67%) |
:arrow_up: |
src/NewFrontend/NFFunction.jl | 27.69% <0.00%> (+27.69%) |
:arrow_up: |
src/NewFrontend/NFOperatorOverloading.jl | 0.00% <0.00%> (ø) |
|
... and 81 more |
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 e45d452...0f88f75. Read the comment docs.
@adrpo I fixed this now should be possible to add more MSL tests. One change here is that I added more precompilation during compiletime so that a user need to spend less time waiting when he or she uses the package.
The initial precompilation will take more time than before, this is also true for importing the package. However, using the package to compile models after it has been used once will be much faster. Overall, this PR improves the performance of the frontend by about 2X:)
This PR introduces more efficient precompilation. It also fixes the performance regression resulting from Julia 1.7 where the compiler went into an infinite loop when inferring types. It also improves the performance of doing lookups