Opened 7 years ago
Closed 17 months ago
#16605 closed defect (fixed)
Nonsymmetric Macdonald polynomials tests too slow
Reported by:  vbraun  Owned by:  

Priority:  major  Milestone:  sage8.9 
Component:  combinatorics  Keywords:  
Cc:  schilling, nthiery, tscrim  Merged in:  
Authors:  Frédéric Chapoton  Reviewers:  Markus Wageringel 
Report Upstream:  N/A  Work issues:  
Branch:  b1cedc0 (Commits)  Commit:  b1cedc08dcc696bd3c0ca7ff9b9d67c7b0764af4 
Dependencies:  Stopgaps: 
Description
There is a small number of extremely longrunning doctests, eating quite a chunk of memory. Is it really true that it can only be tested with 100+ seconds long tests?
#long doctests should still be < 10s. It is about test coverage, not showing off what you can compute given enough time. For every ticket somebody is (hopefully) running all the tests...
Change History (9)
comment:1 Changed 7 years ago by
comment:2 Changed 7 years ago by
Try sage t warnlong 10
. There are also fairly accurate comments about the runtime. You could just loop over fewer Cartan types in some tests, e.g. pick randomly 5.
comment:3 Changed 6 years ago by
 Milestone changed from sage6.3 to sage6.4
comment:4 Changed 17 months ago by
 Branch set to u/chapoton/16605
 Cc tscrim added
 Commit set to b1cedc08dcc696bd3c0ca7ff9b9d67c7b0764af4
 Status changed from new to needs_review
New commits:
b1cedc0  trac 16605 mark some long tests as not tested in nonsym Macdonald polynomials

comment:5 Changed 17 months ago by
bot is morally green, please review
comment:6 Changed 17 months ago by
 Milestone changed from sage6.4 to sage8.9
comment:7 Changed 17 months ago by
review, someone ? please !
comment:8 Changed 17 months ago by
 Reviewers set to Markus Wageringel
 Status changed from needs_review to positive_review
Looks good to me.
comment:9 Changed 17 months ago by
 Branch changed from u/chapoton/16605 to b1cedc08dcc696bd3c0ca7ff9b9d67c7b0764af4
 Resolution set to fixed
 Status changed from positive_review to closed
Hi Volker!
This piece of code was quite tricky to get right for all Cartan types, and it did take us some far testing to get convinced that it was indeed correct. Now, those far tests are indeed most likely only really relevant if someone someday changes the code of non symmetric macdonald polynomials itself. Any change elsewhere in Sage should either break lighter tests or be orthogonal.
So I would be fine with putting #not tested on the very expensive tests and only keep the lighter ones (or scaled down versions thereof). Do you have a specific list of tests that are very expensive?
Cheers,