[01:43:11] <ryankemper>	 @dcausse looks like it failed on `self._prepare_wdqs_target_path()`'s rm command https://www.irccloud.com/pastebin/1iugeaCg/
[07:48:52] <dcausse>	 hm... I don't quite get how this could fail... these are only very simple rm/mkdir/test commands :/
[07:49:22] <dcausse>	 oh
[07:49:45] <dcausse>	 missing comma :/
[07:52:12] <dcausse>	 def func(*par: str), even mypy won't help you if you call func("one" "two") instead of func("one", "two")
[08:04:16] <volans>	 that's string concatenation, it's valid
[08:04:26] <volans>	 >>> "foo" "bar"
[08:04:26] <volans>	 'foobar'
[08:05:21] <pfischer>	 Hm, might make sense with vars in between but with a chain of literals the IDE could at least give a hint
[08:06:14] <volans>	 it's used a lot to split long strings into multiple lines, so maybe the IDE could hint only if there are spaces, not newlines
[08:06:45] <volans>	 the other similar syntax confusion is making a tuple of 1 item, without a comma is just a string
[08:06:48] <volans>	 >>> ("foo")
[08:06:50] <volans>	 'foo'
[08:07:26] <pfischer>	 Oh, so that get’s implicitly converted?
[08:07:54] <pfischer>	 s/get’s/gets/
[08:08:43] <volans>	 no, that's just a string, with parentheses that allow to use the string concatenation
[08:16:33] <claime>	 ebernhardson: https://logstash.wikimedia.org/goto/0266f02c804026e80e666a3d5b7805b2
[08:17:14] <claime>	 I can see one with the new UA I bet https://logstash.wikimedia.org/app/discover#/doc/0fade920-6712-11eb-8327-370b46f9e7a5/ecs-default-1-1.11.0-6-2024.24?id=e-aEC5AB0bw1V4rvhOFM
[08:34:51] <dcausse>	 volans: yes... thing is that I make use of this string concatenation quite a log esp. to split long strings over multilines, which is very similar to splitting long args over multiline too :)
[08:34:53] <dcausse>	 https://gerrit.wikimedia.org/r/c/operations/cookbooks/+/1038904/24..25/cookbooks/sre/wdqs/data-reload.py
[08:35:52] <volans>	 I'm totally with you, it's one of the python syntax that I find fairly annoying and quite error prone
[09:32:55] <dcausse>	 wdqs2023 does not have the latest code... trying scap deploy -l 'wdqs2023.codfw.wmnet' but getting "No targets selected, check limits and dsh_targets"
[10:05:08] <dcausse>	 apparently test hosts are pulled individually as scap targets
[10:08:17] <dcausse>	 lunch
[12:56:49] <inflatador>	 <o/
[13:00:18] <pfischer>	 o/
[13:10:42] <inflatador>	 ryankemper I assigned T364861 to you, feel free to close if it's no longer relevant
[13:10:42] <stashbot>	 T364861: Check SLO impact of Elastic cluster rolling restarts/mitigate if necessary - https://phabricator.wikimedia.org/T364861
[13:15:04] <dcausse>	 o/
[14:04:16] <inflatador>	 workout/taking my son to camp...back in ~1h
[14:58:40] <gehel>	 P&T staff meeting is conflicting with Wednesday meeting. Do what make sense to you!
[15:54:46] <ryankemper>	 Eating some food real quick btw meetings
[16:21:57] <dcausse>	 ryankemper: we made few iterations with Balthazar on the cookbook, I think we're close, uploaded a last set of fixes so if you have time to make another try that'd be great! :) 
[16:21:58] <dcausse>	 dinner
[16:24:58] <ryankemper>	 dcausse: kicking off another run now
[16:41:57] <ryankemper>	 Failed fast this time:
[16:41:59] <ryankemper>	 https://www.irccloud.com/pastebin/HRms4C3Q/
[16:42:11] <ryankemper>	 TLDR `current_age = (right_now_date - dump_date).days` leads to `TypeError: can't subtract offset-naive and offset-aware datetimes`
[16:44:01] <ryankemper>	 https://stackoverflow.com/questions/4530069/how-do-i-get-a-value-of-datetime-today-in-python-that-is-timezone-aware/49059780#49059780 looks like we can just use `datetime.now(timezone.utc).astimezone()`. Getting patch up now
[18:11:28] <inflatador>	 lunch, back in ~1h
[18:46:31] * ebernhardson is not sure List<Function<Predicate<String>, Boolean>> is any simpler :P
[19:01:44] <pfischer>	 ebernhardson: where does that list live and does the outer Function just delegate to its Predicate parameter?
[19:05:18] <ebernhardson>	 pfischer:  this is me trying to make anther patch more java-ish, sec.
[19:06:05] <ebernhardson>	 pfischer:  from https://gerrit.wikimedia.org/r/c/wikidata/query/rdf/+/1041789/1/blazegraph/src/main/java/org/wikidata/query/rdf/blazegraph/throttling/SimpleBooleanExpression.java to https://phabricator.wikimedia.org/P64743
[19:06:12] <ebernhardson>	 but it feels like i'm still not on the right path
[19:07:54] <ebernhardson>	 the validation expression isn't tested yet, suspect it needs ^ and $ but i was about to add some tests to be sure
[21:05:46] <pfischer>	 The regex looks good to me, ran a few tests on regex101.com
[21:06:27] <pfischer>	 Oh, you mean the one in the Paste…
[21:20:34] <ebernhardson>	 pfischer: yea, it did need the ^$ because Pattern.asPredicate uses Pattern.find, although i guess there is no great reason to make it a predicate, other than to simplify the available options
[21:20:56] <ebernhardson>	 i pondered switching over to some dedicated classes for pos/neg expressions, but really this is probably fine
[21:21:00] <ebernhardson>	 it just looks a little awkward
[21:23:00] <pfischer>	 Dedicated classes sound too much.
[21:26:18] <pfischer>	 It’s not intended to be a pluggable API, so the list of expressions seems suitable. Debugging might get a little harder, since expression only gives you a list of cryptic method references, but I’d be fine with that.
[21:30:29] <pfischer>	 ebernhardson: BTW: I would like to test the SUP rate-limiting tomorrow and I thought of either stopping staging for some time and let it catch up or run a batch on staging. Or would you suggest yet another deployment?
[21:32:52] <ebernhardson>	 pfischer: you could probably delete staging and start it with an older timestamp, should be fine
[21:33:02] <ebernhardson>	 but pausing it ahead of time works too
[21:47:59] <ryankemper>	 dcausse: I got a successful run on the `data-reload` cookbook. I did set the `--no-position-kafka-offsets` flag since there's no updater on this host so there is some logic that went untested but I think we're ready to merge https://gerrit.wikimedia.org/r/c/operations/cookbooks/+/1038904
[21:52:32] <ebernhardson>	 \o/