|greppable6||vrurg, Found nothing!|
|Geth||rakudo: vrurg++ created pull request #3891:
[WIP] The first draft of new coercion semantics
03:55 sourceable6 left, committable6 left, benchable6 left, notable6 left, linkable6 left, coverable6 left, quotable6 left, greppable6 left, releasable6 left, shareable6 left, statisfiable6 left, tellable6 left, unicodable6 left, reportable6 left, evalable6 left, bisectable6 left, squashable6 left, nativecallable6 left, bloatable6 left 03:56 statisfiable6 joined, bisectable6 joined 03:57 quotable6 joined, squashable6 joined, tellable6 joined, committable6 joined, shareable6 joined, unicodable6 joined, linkable6 joined, bloatable6 joined 03:58 nativecallable6 joined, evalable6 joined, reportable6 joined, notable6 joined, coverable6 joined 03:59 sourceable6 joined, releasable6 joined, greppable6 joined, benchable6 joined 04:57 reportable6 left 04:58 reportable6 joined, benchable6 left, benchable6 joined 05:01 nebuchadnezzar left 05:02 yuplushi left, nebuchadnezzar joined 05:03 AlexDani` left, AlexDani` joined 05:08 greppable6_ joined 05:09 yuplushi joined, releasable6_ joined 05:11 greppable6 left, releasable6 left 07:31 maggotbrain left 07:32 maggotbrain joined 08:02 maggotbrain left 10:52 leont joined
|Geth||rakudo/precomp_store_parallelism: 1b8849e312 | (Stefan Seifert)++ | src/core.c/CompUnit/PrecompilationStore/File.pm6
Allow for more parallelism when precompiling modules
Instead of locking the whole precomp store whenever we precompile a module or re-check dependencies, we now use one lock file per target file. With this multiple processes may precompile different files and store them in the same precomp store.
Note: it would be even better to not need those .lock files at all and instead lock the target files. But that caused weird issues with multiple processes being able to lock the same file simultaneously. So I'm leaving this next step for a future change.
|rakudo/precomp_store_parallelism: 19961794c3 | (Stefan Seifert)++ | src/core.c/CompUnit/PrecompilationRepository.pm6
Check for circular module loading earlier to avoid deadlock
Since precomp processes now take care of their own locking we have to take that into account and check for circular loading before trying to lock a destination that's potentially already locked by a parent process.
12:02 AlexDani` is now known as AlexDaniel, AlexDaniel left, AlexDaniel joined
|tbrowder||hi, jnthn, i'm making progress with cro except for ssl dependencies||14:45|
|ssl modules can't find libssl.so in a non-standard place||14:47|
|that's at least two modules so i'm not sure where to file the issue, maybe nativecall||14:48|
|erg, never mind, false alarm. as a workaround i've installed the package libssl-dev which is not ideal (since i prefer to use latest openssl) but it works.||15:06|
15:19 maggotbrain joined
|tbrowder||jnthn: \o/ cro non-tls reverse proxy success! check florida-candidates.us||15:25|
|tellable6||tbrowder, I'll pass your message to jnthn||15:26|
|tbrowder||going tls next..l|
|Geth||¦ problem-solving: fecundf assigned to jnthn Issue Extended identifiers- reserving :sym<...> scares me github.com/Raku/problem-solving/issues/225||16:36|
19:45 committable6 left, greppable6_ left, releasable6_ left, reportable6 left, benchable6 left, sourceable6 left, unicodable6 left, squashable6 left, statisfiable6 left, nativecallable6 left, quotable6 left, bisectable6 left, tellable6 left, shareable6 left, linkable6 left, bloatable6 left, evalable6 left, notable6 left, coverable6 left, committable6 joined, statisfiable6 joined 19:46 nativecallable6 joined, quotable6 joined, sourceable6 joined, bisectable6 joined, linkable6 joined, evalable6 joined, unicodable6 joined 19:47 releasable6 joined, benchable6 joined, greppable6 joined, squashable6 joined, shareable6 joined 19:48 reportable6 joined, tellable6 joined, bloatable6 joined, notable6 joined, coverable6 joined 21:15 oddp joined 22:29 oddp left 22:51 Xliff joined