pantsbuild #development

Channels

welcome

github-notifications

announce

hundreds-breakfast-49010

08/17/2020, 10:31 PM

I'm not sure why that's being printed

👍 1

hundreds-breakfast-49010

08/17/2020, 10:33 PM

the draft PR I had for filtering log messages by area of code had an option that would include the rust crate where the log message originated, so if that flag was turned on it would print

workunit_store

(or similar) there

hundreds-breakfast-49010

08/17/2020, 10:33 PM

I don't think tom's merged that yet though?

hundreds-breakfast-49010

08/17/2020, 10:34 PM

yeah I don't see that in the code I just rebased from master with

fast-nail-55400

08/17/2020, 11:10 PM

I put that PR aside until I fix these logging deadlocks

fast-nail-55400

08/17/2020, 11:10 PM

I can put up what I have

fast-nail-55400

08/17/2020, 11:10 PM

I managed to wire up the --log-domain-levels option

fast-nail-55400

08/17/2020, 11:11 PM

let me know if that would be useful

hundreds-father-404

08/17/2020, 11:15 PM

That’s okay, I’m not at all blocked by things. Only trying to do some low-hanging fruit cleanup to logs

hundreds-breakfast-49010

08/17/2020, 11:24 PM

@fast-nail-55400 yeah IIRC the --log-domain-levels option in that code would put something like "workunit_store" in that position in the log message. but that should be optional, we don't want it in logs all the time

hundreds-breakfast-49010

08/17/2020, 11:25 PM

and if that code isn't merged yet anyway I'm not sure where the string in the message eric posted is coming from

👍 1

fast-nail-55400

08/17/2020, 11:25 PM

there was --show-log-domain for that

fast-nail-55400

08/17/2020, 11:25 PM

--log-domain-levels was the dict for filtering

hundreds-breakfast-49010

08/18/2020, 5:14 AM

re: https://github.com/pantsbuild/pants/issues/10129, I am very sure that the multiple item-uncacheable Nodes are being invoked from the public

get()

function exposed by the

graph

crate

hundreds-breakfast-49010

08/18/2020, 5:16 AM

also partially-relatedly, it would be helpful for debugging this (and other problems) if there was a good way to use gdb or a similar debugger in pants rust code

hundreds-breakfast-49010

08/18/2020, 5:17 AM

I feel like this came up at some point a while ago and there was a reason why it wouldn't work, but maybe it can be made to work now that we've switched the cffi framework?

wonderful-iron-54019

08/18/2020, 7:00 PM

hey folks was talking to @hundreds-father-404 about the env var proposall here: https://github.com/pantsbuild/pants/issues/10644, and while i have no issues with it, was wondering if there had been any consideration to preserve the entire original environment. This technically does not. violate hermeticity and can be quite useful for CI/CD environments that load multiple credentials or other things via env

hundreds-breakfast-49010

08/18/2020, 10:41 PM

logging-related thing I noticed - it looks like the

Copy code

15:17:20 [INFO] Initializing pantsd...
15:17:26 [INFO] pantsd initialized`

log messages from pantsd are happening before the code that initializes the rust logger for pantsd is called

witty-crayon-22786

08/18/2020, 11:51 PM

@enough-analyst-54434: while (prematurely: fingers crossed) looking for tickets that will be closed by https://github.com/pantsbuild/pants/pull/10645, i came across https://github.com/pantsbuild/pants/issues/6845: the talking i did to myself there might answer your question about why we memoize based on Params rather than the literal arguments to a rule

hundreds-breakfast-49010

08/19/2020, 2:34 AM

https://github.com/pantsbuild/pants/blob/master/src/python/pants/init/logging.py#L29

hundreds-breakfast-49010

08/19/2020, 2:34 AM

is it just me, or is the

NativeHandler

logging class never actually doing anything if configured to write to a file?

hundreds-breakfast-49010

08/19/2020, 2:34 AM

which might have something to do with https://github.com/pantsbuild/pants/issues/9988

hundreds-father-404

08/19/2020, 2:51 AM

Yeah I noticed something like that yesterday! I’m afk and didn’t look super closely yesterday, but I was wondering if some of that code could be removed

hundreds-breakfast-49010

08/19/2020, 7:52 PM

re: https://github.com/pantsbuild/pants/issues/10518 , the root cause of this seems to be that certain python types we use as parameters a lot, particularly with the new target API, are gigantic python objects

hundreds-breakfast-49010

08/19/2020, 7:52 PM

so their string representation gets to be sizeable

hundreds-breakfast-49010

08/19/2020, 7:53 PM

it's easy enough to truncate this in rust, but I'm wondering if maybe what we should be doing is overriding

__str__

for certain types instead

aloof-angle-91616

08/19/2020, 7:55 PM

i had a patch that would output process_executor command lines for the Process str method

aloof-angle-91616

08/19/2020, 7:55 PM

made it super easy to debug things

aloof-angle-91616

08/19/2020, 7:56 PM

it would be nice if we could have an option to toggle str representations of params between like full/abbreviated

hundreds-breakfast-49010

08/19/2020, 7:56 PM

@hundreds-father-404 do you have any thoughts on how to control the size of the string representation of target-API-related types?