Closed MichaelClerx closed 1 year ago
The error code 123 seems to be a bit of a red herring: https://stackoverflow.com/questions/54016035/does-tar-exit-code-123-always-mean-empty-archive
Tar only returns 0, 1, or 2; and xars turns 1 and 2 into 123
Not sure it's anything to do with the additional repo either? Seems more likely to be something to do with the number of args being sent to tar --- but my understanding of the script is very limited :D
I think it isn't parsing the package name correctly and then fails to archive the files (which causes the tar error). That is my cursory guess though. I can take a look in a few days. What would help in the meantime is to run this with debug
set to true
. Can you do that and post the results?
I added a test for this but am unable to replicate the issue.
https://github.com/awalsh128/cache-apt-pkgs-action-ci/actions/runs/3796943120/jobs/6457479412
Could you follow the instructions from my last comment? This would help since I can't seem to replicate the issue.
NOTE: I inlined your Ubuntu dependencies script as part of the test.
Thanks for looking into this! Sorry for the late reply, was away for christmas.
Have rewritten and simplified the test, but the issue still occurs: https://github.com/myokit/myokit/actions/runs/3803268251
Adding debug: true
results in a warning Unexpected input(s) 'debug', valid inputs are ['packages', 'version', 'refresh']
This goes away when I switch from @latest to @dev: https://github.com/myokit/myokit/actions/runs/3803276847/jobs/6469547524
- echo 'opt/intel/oneapi/lib/licensing/compiler/Intel Developer Tools EULA'
- IFS=
- read -r f tar: opt/intel/oneapi/lib/licensing/compiler/Intel: Cannot stat: No such file or directory tar: Developer: Cannot stat: No such file or directory tar: Tools: Cannot stat: No such file or directory tar: EULA: Cannot stat: No such file or directory
So looks as simple as spaces in the directory name? (I'm guessing these will only occur with non-standard repos, but not sure if that's true, or guaranteed to be true in the future?)
No problem, sorry the flag isn't working for you. :/
Thanks for digging. I'll treat it as a space issue and see if it works. Will update in the next few days (holidays and all).
This is due to the latest
tag not being on the actual latest rev (see #72). I was letting v1.2.0+ soak but then didn't rev soon enough. Sorry for this silly problem. This explains the replication problem and why the debug
input wasn't working.
The tag has been updated. Try to the run again and confirm it is working.
This is what I get now: https://github.com/myokit/myokit/actions/runs/3830142331/jobs/6517643436
Thanks, I'll look into it. FYI, I need to revert the latest
and v1
tag due to #83.
Hi, Thanks for creating this, it looks like it could really speed up our unit tests.
The slowest install we have is from a non-standard repo, and (maybe because of this?) the caching script runs into an error. The final lines shown are:
A full log is here: https://github.com/myokit/myokit/actions/runs/3707556270/jobs/6284113435
Any help with this would be much appreciated