IRC logs for #buildstream for Wednesday, 2019-10-02

« Tuesday, 2019-10-01 Index Thursday, 2019-10-03 »
*** narispo has quit IRC03:34
*** narispo has joined #buildstream03:35
*** slaf_ has joined #buildstream07:01
*** slaf_ has joined #buildstream07:01
*** slaf has quit IRC07:03
*** slaf_ is now known as slaf07:03
*** tpollard has joined #buildstream08:13
*** tiagogomes has joined #buildstream08:34
*** phoenix has joined #buildstream09:03
*** slaf has quit IRC09:04
*** slaf has joined #buildstream09:05
*** slaf has joined #buildstream09:05
*** slaf has joined #buildstream09:06
*** bochecha has joined #buildstream09:07
*** bochecha has joined #buildstream09:09
*** phoenix has quit IRC09:16
*** jonathanmaw has joined #buildstream09:18
*** phoenix has joined #buildstream09:19
*** Kinnison has joined #buildstream09:22
Kinnisonjuergbi: !1611 👍09:22
gitlab-br-botMR !1611: Use buildbox-casd CaptureTree() to import directories https://gitlab.com/BuildStream/buildstream/merge_requests/161109:22
juergbita09:22
*** phoenix has quit IRC09:25
gitlab-br-bottlater opened issue #1149 (Autocompletion without buildbox-casd will stacktrace) on buildstream https://gitlab.com/BuildStream/buildstream/issues/114910:07
gitlab-br-bottlater opened MR !1615 (tlater/fix-casdless-completion->master: Do not check for casd in bash completion) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/161510:07
gitlab-br-botmarge-bot123 closed issue #1132 (Batch CAS writes to reduce round trips between bst and casd) on buildstream https://gitlab.com/BuildStream/buildstream/issues/113210:18
gitlab-br-botmarge-bot123 merged MR !1611 (juerg/capture-tree->master: Use buildbox-casd CaptureTree() to import directories) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/161110:18
*** toscalix has joined #buildstream10:25
Kinnisontlater[m]: !1615 👍10:35
gitlab-br-botjuergbi approved MR !1615 (tlater/fix-casdless-completion->master: Do not check for casd in bash completion) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/161510:36
gitlab-br-botmarge-bot123 closed issue #1149 (Autocompletion without buildbox-casd will stacktrace) on buildstream https://gitlab.com/BuildStream/buildstream/issues/114911:17
gitlab-br-botmarge-bot123 merged MR !1615 (tlater/fix-casdless-completion->master: Do not check for casd in bash completion) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/161511:17
*** phoenix has joined #buildstream11:58
*** phoenix has quit IRC12:02
*** narispo has quit IRC12:04
*** benschubert has joined #buildstream12:33
gitlab-br-botjuergbi approved MR !1605 (bschubert/ensure-buildbox-alive->master: Report when Buildbox-casd is not alive at the end of a run) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/160512:45
gitlab-br-botmarge-bot123 merged MR !1605 (bschubert/ensure-buildbox-alive->master: Report when Buildbox-casd is not alive at the end of a run) on buildstream https://gitlab.com/BuildStream/buildstream/merge_requests/160512:55
*** lachlan has joined #buildstream13:11
*** lachlan has quit IRC13:54
*** lachlan has joined #buildstream13:59
Kinnisonjuergbi: Without wanting to release too many felines among the rats-with-wings -- I just realised REAPI file paths are represented as *valid utf8* strings which means they cannot represent UNIX paths safely.  Have I missed something which mitigates that?14:03
juergbiKinnison: you mean they cannot represent *all* UNIX paths safely. do you anticipate a real issue with this in build actions? I suspect it's mostly a theoretical issue in this context14:11
juergbiREAPI would have to use 'bytes' instead of 'string' everywhere, which would probably be a pain to use with some bindings14:12
* Kinnison is just terrified that now, having noticed, all REAPI calls will fail :D14:12
juergbiin weird environments ;)14:12
KinnisonBut yes, there's not a lot of cases where it'll fail14:12
KinnisonIn morph (baserock) we had to carefully not treat file paths as utf8 strings because in those days glibc's test suite had some tests which had non-utf8 filenames14:13
juergbiah, interesting14:13
juergbiI'm not sure whether protobuf actually enforces this constraint14:13
juergbiPython might14:14
KinnisonSo https://developers.google.com/protocol-buffers/docs/proto3#scalar says it must be utf8 (or 7 bit ascii)14:14
juergbiyes, that's how it's specified. doesn't say whether the implementation actually validates it, though14:15
juergbiit might not, for performance reasons (leaving it up to bindings)14:15
juergbiin any case, we certainly shouldn't rely on missing validation for functionality14:15
*** ikerperez has joined #buildstream14:17
*** ikerperez has quit IRC14:18
*** lachlan has quit IRC14:18
*** ikerperez has joined #buildstream14:20
*** ikerperez has quit IRC14:20
*** ikerperez has joined #buildstream14:21
*** ikerperez has quit IRC14:21
*** lachlan has joined #buildstream14:21
KinnisonSo a brief skim tells me that Prost (a Rust protobuf impl) will reject non-valid-utf8 for string fields14:22
tpollardprost!14:23
*** ikerperez has joined #buildstream14:24
KinnisonThe Java impl seems to have variable strictness from (silently ignore errors) to throw IOException, to lazy throw later14:25
*** ikerperez has quit IRC14:29
KinnisonThe core of the Python impl seems to use six.text_type(bytestr, 'utf-8')14:30
KinnisonAnd then has a strictness check for surrogate pairs14:30
*** Iker is now known as ikerperez14:35
*** lachlan has quit IRC14:49
gitlab-br-bottraveltissues opened issue #1150 (Remove concept of cache key instability) on buildstream https://gitlab.com/BuildStream/buildstream/issues/115014:51
*** lachlan has joined #buildstream15:02
Kinnisontpollard: I've put a few skim review comments onto !1613 for you.15:02
gitlab-br-botMR !1613: WIP: Run bst build under a multiprocessing subprocess https://gitlab.com/BuildStream/buildstream/merge_requests/161315:02
*** bochecha_ has joined #buildstream15:04
tpollardthanks :)15:05
*** bochecha has quit IRC15:06
*** bochecha_ is now known as bochecha15:06
*** lachlan has quit IRC15:18
*** bochecha_ has joined #buildstream15:34
*** lachlan has joined #buildstream15:34
Kinnisonjuergbi: I've pinged you on one of the discussions on !161315:35
*** bochecha has quit IRC15:36
*** bochecha_ is now known as bochecha15:36
Kinnisoncoldtom: Any idea if you'll be able to progress on !1607 ?15:37
gitlab-br-botMR !1607: WIP: Run standardised source tests from external plugins https://gitlab.com/BuildStream/buildstream/merge_requests/160715:37
*** lachlan has quit IRC15:44
*** lachlan has joined #buildstream15:46
coldtomKinnison: im just waiting on review really15:55
gitlab-br-bottlater opened issue #1151 (`bst shell --use-buildtree ask` uses `--pull` inconsistently) on buildstream https://gitlab.com/BuildStream/buildstream/issues/115116:01
*** tpollard has quit IRC16:03
*** bochecha has quit IRC16:03
gitlab-br-bottlater opened issue #1152 (`bst shell` could use splitting into artifact/source commands) on buildstream https://gitlab.com/BuildStream/buildstream/issues/115216:11
*** lachlan has quit IRC16:11
*** lachlan has joined #buildstream16:19
KinnisonWorth pinging people for that?16:25
coldtomyes, I mainly thought it would be best to wait until after the big build meetup16:29
*** lachlan has quit IRC16:31
*** toscalix has quit IRC17:12
*** tiagogomes has quit IRC17:13
*** bethw has quit IRC17:31
*** valentind has quit IRC17:31
*** laurence has quit IRC17:31
*** jward has quit IRC17:31
*** WSalmon has quit IRC17:31
*** valentind has joined #buildstream17:33
*** bethw has joined #buildstream17:34
*** WSalmon has joined #buildstream17:35
*** laurence has joined #buildstream17:35
*** jward has joined #buildstream17:36
*** jonathanmaw has quit IRC17:41
*** adds68 has quit IRC17:49
*** ikerperez has quit IRC17:49
*** qinusty has quit IRC17:49
*** jward has quit IRC18:33
*** laurence has quit IRC18:34
*** jward has joined #buildstream18:34
*** jward has joined #buildstream18:34
*** laurence has joined #buildstream18:34
*** paulsherwood has quit IRC18:37
*** paulsherwood has joined #buildstream18:38
*** paulsherwood has quit IRC18:41
benschubertcoldtom: will look at it on friday, thanks!19:48
benschubertthough you do have tests fialing, mind taking care of those before? :)19:49
coldtombenschubert: I was hoping a review might shed light on why those are failing :P I'll take another look when I get a chance19:50
benschubertah ok no worries then, I will have a look and see if I can understand the failure19:56
gitlab-br-bottraveltissues opened issue #1153 (Provision freedesktop-sdk performance benchmarking) on buildstream https://gitlab.com/BuildStream/buildstream/issues/115320:15
*** adds68 has joined #buildstream20:37
*** phoenix has joined #buildstream21:38
*** phoenix has quit IRC21:38
« Tuesday, 2019-10-01 Index Thursday, 2019-10-03 »

Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!