Closed Razican closed 6 years ago
So, apart from the windows build issue that it's really easy to fix, builds are taking too long when we generate packages. This can be an issue, since I don't know if the deployment scripts count towards the time limit. Example.
An option I'm thinking about would be to divide the package generation with a Travis build matrix. Setting up a variable for the package generation. We can use job exclusion so that we don't end up with 50 jobs.
Deployments are already only done in stable Rust channels and only on Linux builds, so if we generate packages in multiple jobs, all of them will be uploaded. This also means that we can check our 4 targets in each build, and that builds will be shorter, which is nice.
Merging #164 into develop will increase coverage by
1.24%
. The diff coverage is100%
.
@@ Coverage Diff @@
## develop #164 +/- ##
===========================================
+ Coverage 39.7% 40.95% +1.24%
===========================================
Files 16 15 -1
Lines 4480 4344 -136
===========================================
Hits 1779 1779
+ Misses 2701 2565 -136
Impacted Files | Coverage Δ | |
---|---|---|
src/results/report/json.rs | 0% <ø> (ø) |
:arrow_up: |
src/main.rs | 100% <ø> (ø) |
:arrow_up: |
src/results/report/handlebars.rs | 25% <ø> (ø) |
:arrow_up: |
src/static_analysis/mod.rs | 0% <ø> (ø) |
:arrow_up: |
src/config.rs | 81.96% <ø> (ø) |
:arrow_up: |
src/lib.rs | 30.32% <ø> (ø) |
:arrow_up: |
src/static_analysis/code.rs | 66.53% <100%> (ø) |
:arrow_up: |
src/results/mod.rs | 0% <0%> (ø) |
:arrow_up: |
src/static_analysis/certificate.rs |
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 126540d...08b130f. Read the comment docs.
Everything seems to work as expected, packages will be build for each distribution in every build, so I'm guessing that it's enough for now. I finally used Matrix include statement instead of exclude statements, since it's shorter and cleaner.
The only thing that was mentioned in #40 but that is not implemented yet, is the package installation and running tests, which I think it's an overkill if we don't find any further issue. I'm thus closing all the 3 (#40, #149 and #159, this last one already closed).
About that terminal coloring, I will create a new issue to add the explanation in the README that on Windows it should be built with the no-color
feature.
This is meant to fix #40 and #149.
Still WIP: packages are not being installed and tested.