Log file Viewer

[01:37:25] <cooperlees> di_codes / dstufft / EWDurbin: Does the xmlrpc have rate limits on pypi.org? Think I've hit them if it does debugging a bandersnatch bug with me moving it all to asyncio

[01:37:50] <cooperlees> I'll mock it what a fake return if so - Next time I can dump the returned xml. haha

[01:38:40] <EWDurbin> cooperlees: it actually doesn’t. what response are you getting?

[01:39:25] <EWDurbin> reminds me to dust off that PR some day :|

[01:39:26] <cooperlees> O ok, I'll debug more. Seemed to be timing out. Leave it with me then.

[01:39:51] <cooperlees> I gave up waiting to get rid of it and async'd the xmlrpc client :(

[01:39:51] <EWDurbin> How big of a request rate were you pulling?

[01:39:57] <cooperlees> Luckily someone had written one.

[01:40:23] <cooperlees> Empty mirror (so every package) with list_packages_with_serial I need to repro

[01:40:31] <cooperlees> So MBs ?

[01:40:59] <cooperlees> I also think my comcast is doing weird dropouts it seems.

[01:41:11] <cooperlees> But, also happy to try land that API with you too

[01:41:37] <cooperlees> asmacdo and I didn't want to spend the hours on the unit tests unless it was going to get accepted :(

[01:41:40] <cooperlees> Was the main killer

[01:42:13] <cooperlees> HA - I think I just set my timeout to low. And it takes to long to respond

[01:42:26] <EWDurbin> seems plausible sometimes :)

[01:43:37] <cooperlees> Wow. My async refactor of bandersnatch works. Now the hours of refactoring unit tests begins.

[01:43:41] <EWDurbin> cooperlees: what was your timeout?

[01:43:55] <cooperlees> 10 seconds

[01:44:07] <EWDurbin> Oh, wow

[01:44:24] <cooperlees> Lets just kill xmlrpc :D

[01:44:39] <cooperlees> Or at least start the path

[01:44:58] <EWDurbin> I don’t see anything to indicate 10s requests in the metrics..

[01:45:22] <cooperlees> Let me get some time around my calls and have data

[01:45:28] <cooperlees> Could be something clown in my refactor

[01:45:40] <cooperlees> To many changes going on over here

[12:01:47] <dustfinger> I asked this same question on pypa, but I will ask it here as well becaue I might get a different perspective from devs. If I am going to use the build-backend setuptools.build_meta what advantage is there in using pep517.build over pip wheel --use-pep517?

[12:07:52] <pradyunsg> dustfinger: this group is gonna be a subset of pypa mostly. 🙃

[12:16:29] <dustfinger> pradyunsg: There is definitely cross-over. Point understood.

[12:29:59] <travis-ci> pypa/twine#1308 (master - fb03835 : Brian Rutledge): The build passed.

[12:29:59] <travis-ci> Change view : https://github.com/pypa/twine/compare/e9d628c5764e...fb03835ebf12

[12:29:59] <travis-ci> Build details : https://travis-ci.org/pypa/twine/builds/654053951

[22:08:54] <travis-ci> pypa/twine#1309 (master - fb03835 : Brian Rutledge): The build passed.

[22:08:54] <travis-ci> Change view : https://github.com/pypa/twine/compare/613a6f641fdefb815f6ad9cd20ee98220c5529c5...fb03835ebf124ed2f9a447a5f0bd607c283172fb

[22:08:54] <travis-ci> Build details : https://travis-ci.org/pypa/twine/builds/654183313