00:17
[Coke]_ joined
02:13
[Coke]_ left
|
|||
Geth | rakudo/windows_latest_azure_pipeline: ffb4705887 | (Timo Paulssen)++ | azure-pipelines.yml make telemetry, spesh log, and jit dumps downloadable |
02:57 | |
rakudo/windows_latest_azure_pipeline: 4d62c647cf | (Timo Paulssen)++ | azure-pipelines.yml also publish crash dumps |
02:59 | ||
timo | if you look at my azure pipeline "improvements" you will see that my approach is mostly "what if ... but longer!" | 03:00 | |
03:12
[Coke]_ joined
|
|||
Geth | rakudo/windows_latest_azure_pipeline: 5b011eb1d3 | (Timo Paulssen)++ | azure-pipelines.yml create artifacts even when previous build fails |
03:13 | |
rakudo/windows_latest_azure_pipeline: 83aaf9175f | (Timo Paulssen)++ | azure-pipelines.yml even more embarassing mistakes … |
03:15 | ||
03:18
[Coke]_ left
|
|||
Geth | rakudo/windows_latest_azure_pipeline: d537b74a69 | (Timo Paulssen)++ | azure-pipelines.yml try to use jit-bisect.pl on this problem |
03:47 | |
rakudo/windows_latest_azure_pipeline: 10cf6bf2a3 | (Timo Paulssen)++ | azure-pipelines.yml blabla |
03:55 | ||
rakudo/windows_latest_azure_pipeline: 79cdd4a170 | (Timo Paulssen)++ | azure-pipelines.yml is this how to correctly invoke jit-bisect? |
04:03 | ||
rakudo/windows_latest_azure_pipeline: 9f70624086 | (Timo Paulssen)++ | azure-pipelines.yml yeah "dump bytecode" is not going to make the jit magically work you dork |
04:07 | ||
rakudo/windows_latest_azure_pipeline: c72f34b96d | (Timo Paulssen)++ | azure-pipelines.yml fix a later less-important step in the pipeline |
04:08 | ||
04:11
[Coke]_ joined
04:18
[Coke]_ left
|
|||
Geth | rakudo/windows_latest_azure_pipeline: 59bf5a9564 | (Timo Paulssen)++ | azure-pipelines.yml try jit-bisect's spesh mode |
04:19 | |
05:12
[Coke]_ joined
05:17
[Coke]_ left
05:45
[Coke]_ joined
05:49
[Coke]_ left
06:08
[Coke]_ joined
06:17
[Coke]_ left
06:44
[Coke]_ joined
06:51
[Coke]_ left
07:20
djinni` left
07:24
djinni` joined
08:18
[Coke]_ joined
08:22
[Coke]_ left
08:41
sena_kun joined
08:49
[Coke]_ joined
08:54
[Coke]_ left
09:16
[Coke]_ joined
09:23
[Coke]_ left
09:37
sena_kun left
|
|||
Geth | rakudo/windows_latest_azure_pipeline: 42ddb82548 | (Timo Paulssen)++ | azure-pipelines.yml trying our best to get a crash dump outta WERFault.exe |
09:56 | |
10:03
sena_kun joined
|
|||
Geth | rakudo/windows_latest_azure_pipeline: 075e20c7aa | (Timo Paulssen)++ | azure-pipelines.yml try getting the crashdump in the rakudo/ folder |
10:11 | |
rakudo/windows_latest_azure_pipeline: 939a6772f6 | (Timo Paulssen)++ | azure-pipelines.yml try to make log retrieval not crash |
10:14 | ||
rakudo/windows_latest_azure_pipeline: 067cd6f3e7 | (Timo Paulssen)++ | azure-pipelines.yml try using ProcDump.exe |
10:44 | ||
10:54
sena_kun left
|
|||
Geth | rakudo/windows_latest_azure_pipeline: b7b7f64d4a | (Timo Paulssen)++ | azure-pipelines.yml fix Receive-Job, also wait for it to finish |
11:00 | |
timo | i wonder how we could actually make rakudo/moar internals development interesting enough to gain a developer who's familiar enough with windows to make stuff not suck as much | 11:01 | |
i wonder what it would take to make raku code usable as powershell cmdlets for example | 11:02 | ||
i assume cmdlets can only be built by creating some classes for the .NET runtime to consume | 11:06 | ||
slightly annoyed that microsoft calls starting a debugger when a program encounters an unhandled exception "Just In Time (JIT) Debugging" | 11:11 | ||
Geth | rakudo/windows_latest_azure_pipeline: aad280187a | (Timo Paulssen)++ | azure-pipelines.yml cache chocolatey packages |
11:15 | |
11:18
[Coke]_ joined
11:23
[Coke]_ left
|
|||
Geth | rakudo/windows_latest_azure_pipeline: 54dae3190d | (Timo Paulssen)++ | azure-pipelines.yml how are we missing the crash? try dumping multiple times?? |
11:56 | |
12:00
sena_kun joined
|
|||
Geth | rakudo/windows_latest_azure_pipeline: 9a6bfa70c2 | (Timo Paulssen)++ | azure-pipelines.yml maybe -wer was yoinking the file away from here! |
12:07 | |
12:08
sena_kun left
12:09
sena_kun joined
12:19
[Coke]_ joined
12:27
[Coke]_ left
|
|||
timo | leont: can i bother you about the TAP::Harness raku module? | 12:44 | |
Geth | rakudo/windows_latest_azure_pipeline: 4ddda58dcc | (Timo Paulssen)++ | azure-pipelines.yml fix procdump invocation? |
12:47 | |
13:20
[Coke]_ joined
13:27
[Coke]_ left
|
|||
Geth | rakudo/windows_latest_azure_pipeline: c53187e1f4 | (Timo Paulssen)++ | azure-pipelines.yml don't fully compile rakudo for now |
13:43 | |
rakudo/windows_latest_azure_pipeline: 012ffb2ce1 | (Timo Paulssen)++ | azure-pipelines.yml dump up to 100 dumps, no clue how many we need |
13:47 | ||
rakudo/windows_latest_azure_pipeline: c8f7c053db | (Timo Paulssen)++ | azure-pipelines.yml include moar.exe and libmoar.dll in debug output artifact |
13:55 | ||
rakudo/windows_latest_azure_pipeline: 92d2101071 | (Timo Paulssen)++ | azure-pipelines.yml maybe fix copying bin and lib? |
14:11 | ||
rakudo/windows_latest_azure_pipeline: 178f3befad | (Timo Paulssen)++ | azure-pipelines.yml looks like lib doesn't exist after all, only bin |
14:20 | ||
rakudo/windows_latest_azure_pipeline: 7e228a9754 | (Timo Paulssen)++ | azure-pipelines.yml steal moar+nqp cache from azure_improvements |
14:29 | ||
rakudo/windows_latest_azure_pipeline: e27f4a70a3 | (Timo Paulssen)++ | azure-pipelines.yml steal moar+nqp cache from azure_improvements |
14:32 | ||
rakudo/windows_latest_azure_pipeline: 426afe42e6 | (Timo Paulssen)++ | azure-pipelines.yml throw out some stuff that doesn't concern windows+mvm |
14:33 | ||
timo | späm späm späm späm späm späm späm | 14:34 | |
azure windows minidump and späm | |||
Geth | rakudo/azure_improvements: f218012a59 | (Timo Paulssen)++ | azure-pipelines.yml Make extra sure libs and bins are chmodded |
14:39 | |
rakudo/windows_latest_azure_pipeline: c5e955dcfd | (Timo Paulssen)++ | azure-pipelines.yml wow procdump will literally just write how many dumps you ask for??? |
14:45 | ||
leont | Timo: sure | 14:46 | |
timo | i'm looking into making azure more useful or at least pretty to look at; there's two things of interest for the harness (if i switch to the harness6 for azure spec test runs) which are: | 14:47 | |
- you can output a percentage to stdout/stderr that azure will pick up and assign to the current step somehow | |||
14:48
[Coke]_ joined
|
|||
timo | - the PublishTestResults task can ingest a few different test result formats, like xunit, junit, ctest, vstest. i would like to write out which-ever is the simplest one of these so azure can show us our tests | 14:48 | |
harness6 dynamically loads TAP::Harness after making sure it's cloned and up to date, so getting at the symbols at run-time is not possible, so i now have something terrible that 1) creates a TAP::Reporter implementing class that creates a TAP::Reporter::Console to forward all calls to and 2) adds a mixin to the TAP::Reporter::Console::Session that open-test creates to intercept/forward method | 14:55 | ||
calls on that to the original | |||
do you have a better idea how to get both the regular console output visible in the log, plus a (for example) junit output file, and/or maybe writing the verbose tap output to a file (or multiple) | |||
14:58
[Coke]_ left
|
|||
Geth | rakudo/azure_improvements: 8a3c4cfde2 | (Timo Paulssen)++ | azure-pipelines.yml For now, make fullspectest imply relocatable, and use that |
15:08 | |
rakudo/azure_improvements: dd5df2c02a | (Timo Paulssen)++ | azure-pipelines.yml For now, make fullspectest imply relocatable, and use that |
15:10 | ||
15:13
[Coke]_ joined
15:18
[Coke]_ left
|
|||
Geth | rakudo/windows_latest_azure_pipeline: f5804954ff | (Timo Paulssen)++ | azure-pipelines.yml fix caching of nqp&moar |
15:27 | |
rakudo/azure_improvements: f2a3c9dc28 | (Timo Paulssen)++ | azure-pipelines.yml try only "relocatable" variant for now for spectest |
15:34 | ||
leont | timo: sounds like what you want is an alternative Formatter that outputs xunit (or whatever) | 15:41 | |
The system is pluggible, this is possible and depending on the format might not even be hard. | |||
Or actually a Reporter not a Formatter, I forgot I split things up a bit in later versions | 15:42 | ||
timo | right, at least the junit format wants the total amount of tests in the root node of the resulting xml document, so a formatter itself doesn't seem like the right way | 15:44 | |
leont | Much of the architecture was inspired by perl's TAP::Harness, one of its extensions may be a good starting point. TAP::Formatter::JUnit looks sensible enough. | ||
(to steal from) | 15:45 | ||
15:48
[Coke]_ joined
15:53
[Coke]_ left
|
|||
timo | hrm, it pulls in a whole XML package, it would be cool if i could stay with very minimal dependencies, but i'll look at how the code flows anyway | 15:55 | |
leont | Yeah, but I think xunit is also XML based? I don't really know the other formats | ||
timo | you don't happen to know how to work with a minidump and a little heap of .pdb files? | 15:57 | |
leont | No, Windows is not my platform | ||
timo | Windows 11 Error Code: 0xcfffffff - Microsoft Community | 15:58 | |
Harassment is any behavior intended to disturb or upset a person or group of people. Threats include any threat of violence, or harm to another. | |||
leont | (well, just enough to know those are Windows tools) | ||
timo | ^- amazing, the quality of search results i find | ||
leont | I can't find proper documentation for ctest. Not sure if it doesn't exist or is just hidden well. I would expect that not to be XML | 16:00 | |
timo | maybe it's the exact same syntax as you write in CMake files haha | 16:01 | |
16:08
[Coke]_ joined
|
|||
leont | Honestly, TAP may have its issues but people underappreciate "It's simple to implement" | 16:09 | |
16:25
[Coke]_ left
16:31
[Coke]_ joined
16:36
[Coke]_ left
16:38
[Coke]_ joined
16:43
[Coke]_ left
16:52
[Coke]_ joined
|
|||
ab5tract | notable6__: weekly - 5ab5traction5.bearblog.dev/introdu...-comma-20/ | 16:52 | |
notable6__ | ab5tract, Noted! (weekly) | ||
16:58
[Coke]_ left
17:19
bloatable6 left,
bloatable6 joined
17:20
[Coke]_ joined
17:26
[Coke]_ left
17:28
[Coke]_ joined
17:45
|Tux| left
17:57
|Tux| joined
17:58
Voldenet left,
Voldenet joined
18:47
MasterDuke joined
18:50
librasteve_ joined
|
|||
Geth | HTML-Strip/main: 66cb197edd | (Elizabeth Mattijsen)++ | 14 files CI test for 0.1.2 release in zef ecosystem |
19:18 | |
HTML-Strip/main: 2126b4b3d7 | (Elizabeth Mattijsen)++ | META6.json A space in a depends confuses zef |
19:28 | ||
HTML-Strip/main: 5c37ac48be | (Elizabeth Mattijsen)++ | 2 files 0.1.2 |
19:29 | ||
HTML-Strip/main: 15e0f90294 | (Elizabeth Mattijsen)++ | lib/HTML/Strip.rakumod Fix docu typos |
19:34 | ||
19:50
[Coke]_ left
20:12
[Coke]_ joined
|
|||
Geth | rakudo/windows_latest_azure_pipeline: 5e289783e6 | (Timo Paulssen)++ | azure-pipelines.yml fix caching of nqp&moar |
20:22 | |
rakudo/windows_latest_azure_pipeline: 9cb3ff8e63 | (Timo Paulssen)++ | azure-pipelines.yml try getting symbols from the public MS symbol server |
|||
rakudo/windows_latest_azure_pipeline: fd69730e5d | (Timo Paulssen)++ | azure-pipelines.yml fix script download, move it earlier in the jobs list |
20:37 | ||
rakudo/windows_latest_azure_pipeline: 6ef3b22a62 | (Timo Paulssen)++ | azure-pipelines.yml pwsh syntax |
20:43 | ||
20:49
[Coke]_ left
20:56
[Coke]_ joined
21:37
[Coke]_ left
21:42
[Coke]_ joined
22:03
[Coke]_ left
22:11
[Coke]_ joined
|
|||
Geth | rakudo/windows_latest_azure_pipeline: 3cf1f5281c | (Timo Paulssen)++ | azure-pipelines.yml back to WER minidump, moar with some debug output |
22:34 | |
rakudo/windows_latest_azure_pipeline: db104c581c | (Timo Paulssen)++ | azure-pipelines.yml offer alternatively the entire workspace as download |
22:37 | ||
22:37
sena_kun left
22:39
[Coke]_ left
22:47
[Coke]_ joined
|
|||
Geth | rakudo/windows_latest_azure_pipeline: 0064596f84 | (Timo Paulssen)++ | azure-pipelines.yml try procdump -i only, retry windows event log thing |
23:02 | |
rakudo/windows_latest_azure_pipeline: f56c586a8e | (Timo Paulssen)++ | azure-pipelines.yml procdump changes again |
|||
rakudo/windows_latest_azure_pipeline: cffedb34c1 | (Timo Paulssen)++ | azure-pipelines.yml make a relocatable build |
23:17 |