Out of interest, is there a default timeout for `p...
# general
a
Out of interest, is there a default timeout for
pantsd
, or will it keep running indefinitely until some change requires a restart?
w
Infinity and beyond. Anytime you change a global option (as you noted - pants.toml), it'll restart. There are also global options you can pass in at the command line which would cause the same. Hopefully some upcoming changes can reduce that startup time slightly. But yeah, it'll live forever functionally - unless I think the .pantsd folder is deleted
a
Great to know, thanks for the explanation!
👍 1
Would a change to any BUILD files also trigger a
pantsd
restart?
w
No
👍 1
At least, it doesn't for me - more accurately
a
Ah! I think I've found the issue. I've been alternating calls to
pants package
and
pants dependencies
. I ran the former with
-ldebug
but the latter at regular info level. I think this difference in logging level was what triggered the daemon restart.
@wide-midnight-78598 Thanks for your help (esp. your mentioning of global options in your initial reply)!
w
I've done exactly the same thing more times than I care to admit 😊
😆 1
And I've done it in the Pants mainline repo, and changes there take a lonnnng time to schedule
😴 1
a
Probably a silly question, but is there a specific reason why a log level change requires a daemon restart?
w
I can't answer that past it being a global option
a
Yeah, fair enough
h
More specifically, a "bootstrap option"
I don't think all global options restart pantsd, just "bootstrap options", which admittedly is many of them
w
do we maintain a list of bootstrap items? I can't recall a time I've edited pants.toml and not had it restart the scheduler
Ah hang on, I've also messed up terminology - not everything in pants.toml is a "global" option, as we have subsystem-specific ones
h
Yes, and also not all global-scope options are bootstrap options
c
w
🎉 TIL
👍 1