Closed odow closed 4 years ago
Merging #100 into master will not change coverage by
%
. The diff coverage isn/a
.
@@ Coverage Diff @@
## master #100 +/- ##
=======================================
Coverage 83.83% 83.83%
=======================================
Files 4 4
Lines 464 464
=======================================
Hits 389 389
Misses 75 75
Impacted Files | Coverage Δ | |
---|---|---|
src/ECOS.jl | 92.85% <ø> (ø) |
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 90565f2...3235c57. Read the comment docs.
Any ideas? BinaryProvider installs 2.0.5, whereas ECOS_jll installs 2.0.7.
For some reason, 2.0.7 no longer detects infeasibility in one of the MPB tests, and it hits an iteration limit: https://ci.appveyor.com/project/JuliaOpt/ecos-jl/builds/31805100/job/gfs908v4n55d4y78#L384
Nothing in the diff suggests a reason for the test failing: https://github.com/embotech/ecos/compare/v2.0.5...2.0.7
I can build ECOS 2.0.5
with Yggdrasil if you want to keep the same version with Julia < 1.3.
@amontoison yes please. That's a good short-term fix.
In the long-term, it would be nice to understand why things changed.
It's still picking 2.0.7.
https://github.com/JuliaRegistries/General/pull/11905 was not merged.
I merged it. Automerge is not working it seems. @giordano I hope its ok to merge manually.
Close/open this PR again?
Huh - not sure why it is not picking up 2.0.5 on travis.
I kicked it off again. Travis must have got confused during the restart somehow. The other builds passed.
Am I missing something in https://travis-ci.org/github/JuliaOpt/ECOS.jl/builds/669469878? I read
[c2c64177] + ECOS_jll v2.0.5+0
in the log and all tests are passing
x-ref: https://github.com/JuliaPackaging/Yggdrasil/issues/509#issuecomment-605516668