New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
SCIF v0.8.2 does not substitute [e]ENV_VAR correctly #70
Comments
To repicate using the tutorial containers... Expected substitution with older SCIF version in the image referenced in tutorials:
Update to v0.8.2 does and substitution does not work correctly
|
It looks to me as if the root cause is the change here: 67ea105#diff-9a3e86c7ce958be3e6d12390d98a2b51380fbd09434ecb030d839edc92574c87R175 From the PR: #68
Now, It's not clear to me exactly how to get the shell substitution back, while preserving the behaviour in #68 Not sure whether the shell substitution is more important, so #68 should be rolled back, or the exit code behaviour is more important - and we should update our tests for the non-substituted env var? |
lol! That was fast. I didn't realize the python scif was being used here. @dtrudg how do you want to fix it? Should I just revert the PR for now? The issue is just having the return code for run. |
I'll roll back for now and we can figure out another fix for the return code. |
Having looked around, I notice the Hopefully @dmachi might have an idea for an alternative approach to their code that doesn't impact the substitution? |
I don't think that interactive=False change is the problem here. It was originally interactive=interactive, and in one of my intermediate patches before the PR I changed it to interactive=True, exit=True and then in my final patch I changed it back to interactive=interactive. I must have missed something else. I will test this and get it figured out. |
As an aside... @vsoch ... is there any chance you'd be able to add tag this repo with the version used at each PyPi release? That would make it easier to do a |
yep can do. |
Over on sylabs/singularity, we noticed that some of our end-to-end tests that run Docker SCIF containers (not Singularity containers) started failing overnight, coincident with the release of v0.8.2.
We do an
exec hello-world-echo echo
of the SCIF apphello-world-echo
, with a command line that includes the[e]SCIF_APPNAME
construct that previously substituted the value of$SCIF_APPNAME
:i.e.
resulted in the following output on SCIF v0.8.1:
Now, with SCIF v0.8.2, the
[e]SCIF_APPNAME
is substituted with a literal$SCIF_APPNAME
, not its value...The text was updated successfully, but these errors were encountered: