These are chat archives for symengine/symengine

16th
Mar 2016
Isuru Fernando
@isuruf
Mar 16 2016 01:44
See coeff method
Abhinav Agarwal
@abhinavagarwal07
Mar 16 2016 10:50
What does outArg do?
Isuru Fernando
@isuruf
Mar 16 2016 11:02
Creates a pointer to a RCP
Rajith Vidanaarachchi
@rajithv
Mar 16 2016 16:05

I have drafted a proposal for GSoC 2016, for the Ruby Wrappers for the SymEngine project, under SciRuby. Can you guys please check and give some feedback?

https://github.com/rajithv/symengine.rb/wiki/GSoC-2016-Rajith-Vidanaarachchi:-Ruby-Wrappers-for-SymEngine

Thanks!

Isuru Fernando
@isuruf
Mar 16 2016 18:19
@certik, can you enable codecov? It's much better than coveralls. It also has a nice browser extension.
Ondřej Čertík
@certik
Mar 16 2016 18:22
@isuruf sure.
I just logged in and gave codecov.io permissions to access the symengine org.
Then I think we just need to put this into our .travis.yml and that's it:
before_install:
  - pip install codecov
after_success:
  - codecov
Isuru Fernando
@isuruf
Mar 16 2016 18:25
That should be it. There's also some settings to change. I'll let you know the details
symengine/symengine#874
#874 and #873 both have this status. AppVeyor was unable to build non-mergeable pull request
Ondřej Čertík
@certik
Mar 16 2016 18:27
I've seen that. Don't know why.
Ondřej Čertík
@certik
Mar 16 2016 18:34
@isuruf it seems to be working.
There are two Travis builds that upload the code coverage.
What happens --- does it get overwritten?
Isuru Fernando
@isuruf
Mar 16 2016 18:36
Not sure. I'll remove the first one
I pushed in a new commit to the PR and appveyor seems to be working
Looking at a file like, https://github.com/isuruf/symengine/blob/master/symengine/add.cpp with the browser extension is really cool
Ondřej Čertík
@certik
Mar 16 2016 19:14
@isuruf Can you send a screenshot?
Isuru Fernando
@isuruf
Mar 16 2016 19:15
blob
Ondřej Čertík
@certik
Mar 16 2016 19:15
nice!
it's a completely dead code.
Isuru Fernando
@isuruf
Mar 16 2016 19:17
Ah yes, there are tests for individual series implementations, but the generic series function is not tested
Ondřej Čertík
@certik
Mar 16 2016 19:17
Is there some easy way to see just the changes in the PR?
In a PR.
In other words, to make sure that the PR itself is 100% tested, even if otherwise symengine is not?
Isuru Fernando
@isuruf
Mar 16 2016 19:19
yes
isuruf/symengine#13
Click the link at the Coverage Diff
Abinash Meher
@abinashmeher999
Mar 16 2016 19:20
@rajithv Few suggestions.
  • You should elaborate more on how you are going to do what you have mentioned in the timeline. There should be a little more of how you will be doing things.
  • You might want to incorporate exception/error handling from the beginning of the project rather than keeping it for the end.
  • I would suggest you to keep well defined deliverables that you wish to achieve with code samples. Like some ruby code that should work correctly both for the end of midterm evaluations and the endterm evaluations.
  • You should mention how you will be integrating the other gems in greater detail.
  • You should add atleast two paragraphs for the work that you will be doing each week (that is what SciRuby demands).
Ondřej Čertík
@certik
Mar 16 2016 19:24
@isuruf nice, very useful! I think the code coverage is cumulative -- and so perhaps we can submit it multiple times, so that we can test symengine with, say, different integer implementations.
Isuru Fernando
@isuruf
Mar 16 2016 19:25
Yes, that's good
Ondřej Čertík
@certik
Mar 16 2016 19:26
The only thing is that I think the gcc instruments the source code when we run it, so we probably don't want to run coverage on all our tests, to ensure that things are actually working without the coverage.
@isuruf I think we don't need do make codecov comment in our PRs, because you can get to all the information by just clicking on the commit status symbol, which leads to the codecov page.
What do you think?
Isuru Fernando
@isuruf
Mar 16 2016 19:30
Yes. The comment is a giant distraction
Ondřej Čertík
@certik
Mar 16 2016 19:33
Exactly.