Opened 3 years ago
Closed 3 years ago
#28596 closed defect (fixed)
Fix jmol permissions
Reported by: | jhpalmieri | Owned by: | |
---|---|---|---|
Priority: | major | Milestone: | sage-9.0 |
Component: | packages: standard | Keywords: | |
Cc: | jdemeyer, embray, novoselt | Merged in: | |
Authors: | Erik Bray | Reviewers: | John Palmieri |
Report Upstream: | N/A | Work issues: | |
Branch: | 87c694f (Commits, GitHub, GitLab) | Commit: | 87c694f059e839a4bb12db80037f3c4398b56c2b |
Dependencies: | Stopgaps: |
Description (last modified by )
This is a followup to #25026: the permissions of SAGE_ROOT/local/share/jmol/
are wrong:
drwxr-xr-x 24 palmieri staff 768 Sep 12 00:11 info drwx------ 21 palmieri staff 672 Oct 12 10:28 jmol drwxr-xr-x 80 palmieri staff 2560 Oct 12 10:28 jsmol drwxr-xr-x 4 palmieri staff 128 Sep 12 00:11 jupyter
Change History (16)
comment:1 Changed 3 years ago by
- Description modified (diff)
comment:2 Changed 3 years ago by
- Cc jdemeyer embray novoselt added
comment:3 Changed 3 years ago by
- Branch set to u/jhpalmieri/jmol-permissions
comment:4 Changed 3 years ago by
- Commit set to 82e736b7c1f5f93042d61d34149150ab7d9cec2b
- Status changed from new to needs_review
New commits:
82e736b | trac 28596: change permissions on src after unpacking each package's tarball.
|
comment:5 Changed 3 years ago by
I want to do this for all packages to avoid similar situations in the future.
comment:6 follow-ups: ↓ 7 ↓ 10 Changed 3 years ago by
This does not seem to fix the doc building issue.
comment:7 in reply to: ↑ 6 Changed 3 years ago by
Replying to chapoton:
This does not seem to fix the doc building issue.
I don't know what doc building issue you're referencing.
comment:8 Changed 3 years ago by
Here is the docbuild log where the problem appears:
This prevents me from using this machine as a patchbot.
comment:9 Changed 3 years ago by
- Status changed from needs_review to needs_work
This behavior originates from #24567, and would be different if we'd gone with Jeroen's suggestion of a less restricted mask on the files than I used.
I wrote at the time:
this should have no effect on the permissions applied when the files are installed, but I guess it depends in part on the installation mechanism
Which, incidentally was true at the time. But the part of "depends on the installation mechanism" was also true, and the installation mechanism did change after that, in such a way that this did have an effect on the installed files.
So I think the better option here might be to try Jeroen's original suggestion of a less restrictive permission mask on the source tarball's extraction location.
comment:10 in reply to: ↑ 6 Changed 3 years ago by
- Branch changed from u/jhpalmieri/jmol-permissions to u/embray/ticket-28596
- Commit changed from 82e736b7c1f5f93042d61d34149150ab7d9cec2b to 87c694f059e839a4bb12db80037f3c4398b56c2b
This approach fixes the issue closer to the source. However, it would be good to test a full rebuild. I don't know for sure if there aren't some pathological packages that can break because of this (and if so those packages should probably be fixed...).
New commits:
df58d4b | Trac #28596: Apply less restricted permission mask to extracted source
|
87c694f | Add an spkg-legacy-uninstall for jmol.
|
comment:11 Changed 3 years ago by
- Status changed from needs_work to needs_review
comment:12 Changed 3 years ago by
All tests pass for me on OS X, with a build from scratch, and the permissions look okay.
comment:13 Changed 3 years ago by
then it should be ok, no ?
comment:14 Changed 3 years ago by
- Reviewers set to John Palmieri
- Status changed from needs_review to positive_review
Okay with me.
comment:15 Changed 3 years ago by
I'm sure, if he were paying attention, Jeroen would be pleased to be vindicated :)
comment:16 Changed 3 years ago by
- Branch changed from u/embray/ticket-28596 to 87c694f059e839a4bb12db80037f3c4398b56c2b
- Resolution set to fixed
- Status changed from positive_review to closed
I can see two solutions: change jmol's spkg-install script so it sets the right permissions, or set the permissions differently for all packages. Here is a branch which does the second of these: after unpacking the tarball into
src
, runchmod +rx src
. Is this a bad idea for some reason?