Welcome to the main channel on the development of MoarVM, a virtual machine for NQP and Rakudo (moarvm.org). This channel is being logged for historical purposes.
Set by lizmat on 24 May 2021.
nine Yesterday snow, today sun. Just beautiful :) 10:43
lizmat we appeared to have some snow last night, but this morning it was all gone already 10:45
but it *is* sunny atm :-)
nine Well it is an ephemeral beauty at 2.7 degrees 10:46
Geth MoarVM/fix_unwind_missing_frame: 66b9f87dce | (Stefan Seifert)++ | 2 files
Fix wrong result from bind-will-resume-on-failure due to region record

The bind-will-resume-on-failure syscall could give a false negative when between the frame record and the bind control record the syscall was looking for a start region record was encountered. This led to multi dispatch failures. Fix by using the new MVM_callstack_prev_significant_record function which ignores region records.		 11:59
MoarVM/fix_unwind_missing_frame: 7f07c19120 | (Stefan Seifert)++ | src/core/callstack.c
Fix "No frame found in callstack region" error

When a frame did not fit into the same callstack region as the continuation tag continuationcontrol would end up in a "No frame found in callstack region" panic.
Fix by continuing the search in the following regions. This is ok because MVM_callstack_continuation_slice does not just splice off the one region it returns, but the full chain following it.
MoarVM/fix_unwind_missing_frame: a3a04f47c2 | (Stefan Seifert)++ | 3 files
Fix unwinding a nested runloop leaving a region start record on the call stack

In case we had to allocate a new callstack region for a nested runloop record, cleaning up this record left a region start record on top of the callstack. This would later lead to assertion failures or even worse, confused runloops.
nine With those fixes I can't provoke any callstack related crashes anymore. Not even by allocating a new region for every single record 12:46
Geth MoarVM/fix_uninitialized_reg_after_pea_and_inlining: dee757d359 | (Stefan Seifert)++ | src/spesh/inline.c
Fix uninitialized register after PEA and inlining

The register in question should have been written to by a PHI node and carried the appropriate deopt usage markers. The other side of the PHI (just one version) didn't. So PEA optimized away the writer.
Fix by copying the dopt_users from the PHIs target to the PHI's sources when rewriting returns. This way PEA knows that it has to keep the writer alive.		 13:03
MoarVM: niner++ created pull request #1612:
Fix uninitialized register after PEA and inlining		 13:05
MoarVM: 2a124542e9 | (Ben Davies)++ | src/io/asyncsocket.c
Close socket on async error during an async connection

Related to github.com/rakudo/rakudo/issues/4370.		 13:13
MoarVM: 54aa83655e | niner++ (committed using GitHub Web editor) | src/io/asyncsocket.c
Merge pull request #1504 from Kaiepi/asyncconnect-error-close

Close socket on async error during an async connection
nine The snow has melted. 14:06
The sun is gone.
Bugs were fixed.
Lockdown in Austria
jnthnwrthngtn No snow here yet, alas. 14:28
Nicholas nine: our snow has not yet melted. It's working on it. And the sun is gone. 14:41

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!