🦋 Welcome to the IRC channel of the core developers of the Raku Programming Language (raku.org #rakulang). This channel is logged for the purpose of history keeping about its development | evalbot usage: 'm: say 3;' or /msg camelia m: ... | log inspection situation still under development | For MoarVM see #moarvm
Set by lizmat on 22 May 2021.
00:01 [Coke] joined 00:06 reportable6 left 00:07 reportable6 joined 02:05 evalable6 left, linkable6 left 03:06 evalable6 joined 04:06 greppable6 left, reportable6 left, sourceable6 left, coverable6 left, evalable6 left, bisectable6 left, unicodable6 left, bloatable6 left, shareable6 left, tellable6 left, nativecallable6 left, committable6 left, notable6 left, quotable6 left, releasable6 left, statisfiable6 left, benchable6 left, squashable6 left 04:07 greppable6 joined, statisfiable6 joined, coverable6 joined, shareable6 joined 04:08 committable6 joined, bisectable6 joined 04:09 benchable6 joined, squashable6 joined, reportable6 joined, unicodable6 joined 05:07 notable6 joined, linkable6 joined, tellable6 joined 05:08 quotable6 joined, bloatable6 joined 05:09 sourceable6 joined 05:51 Xliff left 06:04 frost joined 06:07 reportable6 left 07:06 nativecallable6 joined 07:07 releasable6 joined 07:10 reportable6 joined 08:09 evalable6 joined 08:46 TheAthlete joined 11:01 evalable6 left, linkable6 left 12:01 evalable6 joined 12:08 reportable6 left 12:09 reportable6 joined 12:42 Altai-man joined 13:03 linkable6 joined 14:10 TempIRCLogger left, TempIRCLogger joined 15:10 shareable6 left, squashable6 left, nativecallable6 left, greppable6 left, bisectable6 left, bloatable6 left, benchable6 left, linkable6 left, evalable6 left, statisfiable6 left, coverable6 left, unicodable6 left, reportable6 left, notable6 left, sourceable6 left, releasable6 left, quotable6 left, tellable6 left, committable6 left 15:11 frost left, bisectable6 joined, evalable6 joined, unicodable6 joined 15:12 benchable6 joined, linkable6 joined, committable6 joined 15:13 tellable6 joined 16:11 quotable6 joined, coverable6 joined 16:12 statisfiable6 joined, squashable6 joined, reportable6 joined 16:13 notable6 joined, releasable6 joined, sourceable6 joined 16:28 Xliff joined 17:08 TheAthlete left 17:12 shareable6 joined 17:14 bloatable6 joined
Geth nqp: bb14317951 | (Elizabeth Mattijsen)++ | tools/templates/MOAR_REVISION
Bump MoarVM to get nine++'s latest fixes
18:09 reportable6 left 18:10 reportable6 joined
jdv isn't there a nqp bump needed? 18:12
lizmat yes, that's coming after I ran all the tests :-)
jdv ah, cool
Geth rakudo: 96919e52fd | (Elizabeth Mattijsen)++ | tools/templates/NQP_REVISION
Bump NQP to get nine++'s MoarVM fixes
lizmat jdv ^^
jdv ok, thanks 18:19
Geth rakudo: 835d0f12a9 | (Stefan Seifert)++ | 2 files
Fix segfaults due to concurrent hash access in parallel module loading

Read access to the %!seen and %!loaded hashes in CompUnit::Repository:: Installation and FileSystem was not protected by a lock. This could lead to us fetching an only half-setup entry from the hash with a key that was still a NULL pointer which we then dereferenced.
Fix by extending (or introducing) locks to all access to those hashes.
jdv i imagine between today and tomorrow i will make the first attempt, assuming a new blin run works out
Geth rakudo: f4d7c4f366 | niner++ (committed using GitHub Web editor) | 2 files
Merge pull request #4679 from rakudo/fix_concurrent_module_loading

Fix segfaults due to concurrent hash access in parallel module loading
jdv i guess i'll wait a bit longer... 18:22
Geth rakudo: 0b286ed21c | Xliff++ (committed using GitHub Web editor) | src/core.c/Main.pm6
MAIN default values (#4646)

  * - Adds default values to .WHY.trailing arguments when used in sub MAIN
  * - Adds default valeus to USAGE in MAIN
lizmat hmmm... that's odd, but good to see merged (it was I who merged that one) 18:26
18:26 linkable6 left 18:45 Altai-man left
Geth rakudo: a6fad8798b | (Stefan Seifert)++ | 6 files
Fix removing proxies in native calls

While removing proxies by passing args through a slurpy was clever and very little code, it had the unfortunate side effect of containerizing those arguments. This would then lead to issues when the same callsite could pass defined and undefined objects for the same parameter.
Fix by replacing the clever but slow and buggy implementation with usage of the ProxyReaderFactory that Rakudo also uses for multi dispatch. This will make calls correcter, faster and get rid of the raku-native-dispatch-deproxy frame in the callstack (and backtraces).
rakudo: f0713c42b4 | (Stefan Seifert)++ | lib/NativeCall/Dispatcher.rakumod
Make native calls with Proxy args more efficient

Split off raku-nativecall-core from raku-nativecall and use this as resumption target for raku-nativecall-deproxy-resume. This way we get rid of redundant checking for proxies in arguments and the redundant check for whether the callee is fully set up yet. Both were done already at the start of the dispatch.
rakudo: 26dd2f4824 | niner++ (committed using GitHub Web editor) | 6 files
Merge pull request #4676 from rakudo/fix_removing_proxies_in_native_calls

Fix removing proxies in native calls
19:26 linkable6 joined 19:30 japhb joined 20:11 nativecallable6 joined
jdv releasable6: status 21:45
releasable6 jdv, Next release will happen when it's ready. 1 blocker. 198 out of 198 commits logged
jdv, Details: gist.github.com/97c4d07d4b3ccb6a5a...fe96379321
jdv a blin on head just finished and its clean enough. the only potential "blockers" i'm aware of are how to handle the regressions related to ro vars. 21:46
the rakudo and moarvm changelogs are on the respective wiki pages 21:47
lizmat jdv++
jdv anyone have any strong opinions about the ro vars regressions? 21:48
i could go either way. if the fallout is unacceptable we could revert in next release or we could revert for this release...
i haven't seen any chatter from any of the parties involved but it is holiday season... 21:49
lizmat jdv: please remind us what that is about again?
jdv Its the regression in the blin ticket for Data::Record, Auth::SAML2, and XML::Signature 21:50
github.com/rakudo/rakudo/pull/4536 seems to be the offender
lizmat yeah, ok, well looks like in Data::Record, the problem is actually in the *test*, not in the module itself 21:52
lemme see if I can make a PR for that one 21:53
jdv but in terms of release, regardless of PR, its still a regression, right? 21:56
lizmat jdv: problem is already fixed, Data::Record just needs a new release
jdv nice 22:00
and of course its not XML::Signature but XML::Canonical. i didn't look closely enough. 22:06
lizmat jdv: but XML::Canonical is not in github.com/rakudo/rakudo/issues/4667 ? 22:08
jdv i know - blin reported XML::Signature but if you read the failure better or try to actually debug it locally its XML::Canonical (at least) 22:09
its a little perplexing 22:10
22:13 greppable6 joined
lizmat m: sub a(Any $a) { $a := 42 } 22:14
camelia ===SORRY!=== Error while compiling <tmp>
Cannot bind to '$a' because it was bound in a signature and variables
bound in signatures cannot be rebound unless they were declared with the
'is rw' or 'is copy' traits
at <tmp>:1
------> su…
lizmat jdv ^^ that's basically the issue
m: sub a(Any $a is copy) { $a := 42 } 22:15
camelia ( no output )
jdv i see that. just not sure yet what the fix in situ is.
lizmat add an "is copy" to the sig
jdv oh, i think i see what's up 22:16
i did that and assumed the test breakage indicated it was wrong but github.com/retupmoca/P6-XML-Canonical/issues/6 22:17
lizmat yeah, I see that same error
jdv huh. so, does that mean we just do a PR and forget about it? looks like that dist is not active.
wonder what if anything depends on those 2 dists 22:18
lizmat looks like Andrew Egeler is more into Rust these days
perhaps their modules should be moved to raku-community-modules 22:19
for now, I'd say, the issue can be fixed easily 22:20
the issue of the failing test has been open for a year now
for now I'd not make the XML::Signature failure a blocker, as it can be fixed by a module update 22:21
jdv and of course Auth::SAML2 has the same root cause
lizmat I've left a ping in the issue, hope that will get someone's attention :-) 22:23
jdv cool. i'll make a pr for fun and update the tickets and stuff... 22:24
i'm about done for today. not feeling well. i'll start the actual release tomorrow morning. likely ~9am EST. 22:26
thanks lizmat!
lizmat ++jdv 22:27
sleep well! Get well!
MasterDuke when i've emailed Andrew in the past about issues he's responded pretty quickly 22:28
22:44 japhb left