#25806 closed defect (fixed)
Pass noreadline to R
Reported by:  Julian Rüth  Owned by:  

Priority:  major  Milestone:  sage8.4 
Component:  interfaces  Keywords:  
Cc:  François Bissey, Antonio Rojas, Ximin Luo, Jeroen Demeyer, Erik Bray  Merged in:  
Authors:  Ximin Luo, Julian Rüth  Reviewers:  Erik Bray 
Report Upstream:  N/A  Work issues:  
Branch:  9e6ea7e (Commits, GitHub, GitLab)  Commit:  
Dependencies:  Stopgaps: 
Description (last modified by )
Conda and Gentoo (and apparently also ArchLinux and Debian) suffer from hangs when talking to R through the pexpect interface on the first invocation in a Sage session. It is unclear what is the underlying issue. Shortening Sage's R prompt seems to fix the issue as does passing noreadline
to R.
Is there any reason why we should keep readline support when talking to R with the pexpect interface? Otherwise, we should just pass on that flag even though we do not fully understand why it fixes that particular issue.
See https://github.com/condaforge/sagelibfeedstock/issues/19#issuecomment402557595, and https://github.com/cschwan/sageongentoo/blob/master/scimathematics/sage/files/dtrnoreadline.patch.
Change History (10)
comment:1 Changed 4 years ago by
Branch:  → u/saraedum/25806 

comment:2 Changed 4 years ago by
Commit:  → 2d8f4bc9e6acd82f386a92ad087580c9b9d1d1fe 

Status:  new → needs_review 
Work issues:  → is the patchbot happy? 
comment:3 Changed 4 years ago by
Description:  modified (diff) 

comment:4 Changed 4 years ago by
Commit:  2d8f4bc9e6acd82f386a92ad087580c9b9d1d1fe → 9e6ea7e3a2eb5c9d28159942b6d794a6e830a148 

Branch pushed to git repo; I updated commit sha1. New commits:
9e6ea7e  Link to #25806 to explain the noreadline switch

comment:5 Changed 4 years ago by
Reviewers:  → Erik Bray 

Status:  needs_review → positive_review 
This makes sense to me. There is a similar issue #25363 where it was needed to disable prompt_toolkit for Cantor's pexect interface to Sage to work (I believe Cantor is wrong to use pexpect to wrap Sage in the first place, but that's a different story).
I don't know why this would be a problem on some platforms but not others.
New commits:
9e6ea7e  Link to #25806 to explain the noreadline switch

comment:6 Changed 4 years ago by
Status:  positive_review → needs_review 

Work issues:  is the patchbot happy? → is the patchbot happy ⇒ positive review 
Great. Let's wait for the patchbot to actually pass as I have not run doctests for this myself.
comment:7 Changed 4 years ago by
#25503 does this already, and should ease downstream packaging (still needs review).
Replying to embray: [...]
I don't know why this would be a problem on some platforms but not others.
To what i understand, this is not a platform issue, but a version issue: our interface does not work with readline 7.
comment:8 Changed 4 years ago by
Status:  needs_review → positive_review 

Work issues:  is the patchbot happy ⇒ positive review 
comment:9 Changed 4 years ago by
Branch:  u/saraedum/25806 → 9e6ea7e3a2eb5c9d28159942b6d794a6e830a148 

Resolution:  → fixed 
Status:  positive_review → closed 
comment:10 Changed 4 years ago by
Commit:  9e6ea7e3a2eb5c9d28159942b6d794a6e830a148 

Milestone:  sage8.3 → sage8.4 
I might have misunderstood what the milestone stands for but I thought it would be the release of Sage that includes this. So, this is not in 8.3 yet but will only be in 8.4.
New commits:
Pass noreadline to R otherwise Pexpect hangs