Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
Brett Smith
@brettcs:matrix.org
[m]
It's something like "bound method manifest_text"
Peter Amstutz
@tetron:matrix.org
[m]
ah
Brett Smith
@brettcs:matrix.org
[m]
Which, again, makes sense, the instance of the method manifest_text that's bound to c1 obviously can't be the same as the one bound to c2
Peter Amstutz
@tetron:matrix.org
[m]
I have to admit I'm getting a bit of a "how did this ever work" vibe
Brett Smith
@brettcs:matrix.org
[m]
Yeah me too, which is one of my least favorite vibes
Peter Amstutz
@tetron:matrix.org
[m]
what version of Python are you using to run tests?
Brett Smith
@brettcs:matrix.org
[m]
3.9, the stock one in bullseye
Peter Amstutz
@tetron:matrix.org
[m]
huh
print(c1.manifest_text, c2.manifest_text, c1.manifest_text == c2.manifest_text) -> <bound method RichCollectionBase.manifest_text of <arvados.collection.Collection object at 0x7f1fc77b6ac8>> <bound method RichCollectionBase.manifest_text of <arvados.collection.Collection object at 0x7f1fc6468940>> True
Brett Smith
@brettcs:matrix.org
[m]
Hmm, what I said earlier about method instances "obviously" being different is only true when the names c1 and c2 refer to different objects. If they were the same object, the test would pass...
Well okay now I have two questions
Based on that output, I don't think what the test is doing is what the test intended
But also how the heck is that comparison true
Peter Amstutz
@tetron:matrix.org
[m]
the method definition manifest_text is the same which means possibly the comparison is of function pointers and not taking the bound object id into account which seems odd but this might be a box of horrors best left unopened
I agree 100% that the intended behavior was to test the value returned by the manifest_text() m ethod
Brett Smith
@brettcs:matrix.org
[m]
Your "box of horrors" is my "unexplored lower depths of the iceberg"
But yeah for now I can just adjust the text. Should I prep a small branch for that?
Peter Amstutz
@tetron:matrix.org
[m]
yes please
FWIW arvbox uses Python 3.7, should double check what version jenkins uses
that python 3.7 and 3.9 might have different opinions about this corner case seems most likely
Brett Smith
@brettcs:matrix.org
[m]
Peter Amstutz: I have confirmed that the obvious fix of just calling those methods is the only fix necessary to get the tests passing. OK to just push directly?
Peter Amstutz
@tetron:matrix.org
[m]
yes
Brett Smith
@brettcs:matrix.org
[m]
Thanks.
Peter Amstutz
@tetron:matrix.org
[m]
Lucas Di Pentima
@ldipenti
@tetron:matrix.org goo morning! https://dev.arvados.org/issues/19699#note-19 back to you
1 reply
Peter Amstutz
@tetron:matrix.org
[m]
thanks
Lucas Di Pentima
@ldipenti
@tetron:matrix.org 19699-cwl-dl-docs LGTM, thanks
Peter Amstutz
@tetron:matrix.org
[m]
Brett Smith
@brettcs:matrix.org
[m]
Thanks
Brett Smith
@brettcs:matrix.org
[m]
Peter Amstutz: For the structure of the create body, the API docs make it look like update should follow the same pattern. Is that right?
Peter Amstutz
@tetron:matrix.org
[m]
correct
Stephen Smith
@stephen304
@tetron:matrix.org https://dev.arvados.org/issues/19700#note-8 is ready for review
Peter Amstutz
@tetron:matrix.org
[m]
great!
Peter Amstutz
@tetron:matrix.org
[m]
@stephen304: HTTP URL rendering works great, scrolling is fixed, but I'm still seeing empty JSON tab for inputs (but not outputs), ummm, what would you suggest doing to debug it?
I'm seeing this when I visit the process page while it is in the "Running" state
Peter Amstutz
@tetron:matrix.org
[m]
Stephen Smith
@stephen304
Hmm I'll have to dig around, it shouldn't be able to show preview values if the json doesn't get loaded
Peter Amstutz
@tetron:matrix.org
[m]
right, I think it's loaded, but something about the Running state is different
try launching a job and then visiting it
see if you can reproduce it that way
Stephen Smith
@stephen304
I tried a couple but they're stuck at queued on tordo
Peter Amstutz
@tetron:matrix.org
[m]
Ah, they should start eventually
It has been getting compute node startup failures I haven't chased down
Peter Amstutz
@tetron:matrix.org
[m]
@stephen304: 19700-http-input LGTM with that last fix
@tomclegg: 18842-arv-mount-disk-config was back to you a few days ago, do you have time to look at it today?
Tom Clegg
@tomclegg
Peter Amstutz
@tetron:matrix.org
[m]
Daniel KutyƂa
@DannyNs
ticket ready to be reviewed https://dev.arvados.org/issues/19786 by whoever takes the review task :)
thank you @stephen304 for the test hint :)