github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
nwc10 jnthn: "I think the second bit is an accident" - "Drat," said Damain, "they spotted my cunning plan" 05:14
dah, I typoed my punchline.
Geth MoarVM: 5ab2a1d292 | (Daniel Green)++ | azure-pipelines.yml
Remove unnecessary Windows-specific jobs from CI

We can use Powershell on all OSes, so do that, but then only set the few Windows-specific options at runtime.
Also stick the install destination in a variable, so for relocatable builds we can change that after moving the installation and subsequent jobs can be simplified.
06:24
MoarVM: e31b57be4d | MasterDuke17++ (committed using GitHub Web editor) | azure-pipelines.yml
Merge pull request #1483 from MasterDuke17/remove_unnecessary_windows_specific_tasks_from_ci

Remove unnecessary Windows specific tasks from CI
Geth MoarVM/probe-toolchain-cleanup: 10 commits pushed by (Nicholas Clark)++
review: github.com/MoarVM/MoarVM/compare/3...58a77c176c
07:57
jnthn nwc10: I realized on my walk that a word boundary *does* indeed mean `:tempest;` would no longer be parsed there, but my example with `-` will, 'cus that is a word boundary. So I guess the cunning plan largely survives... :) 09:17
nwc10 this was an all-night walk? 09:21
nine those are the best 09:22
jnthn It wasn't, although I'm feeling sleepy enough this morning that it might as well have been... 09:39
nwc10 I was like that yesterday. For some reason last night I slept much better 09:41
but I had this strange dream about getting a replacement passport. From some guy with a big beard. Who did it very quicky. And then I realised that the outer cover was had the wrong language on it. And it was a whole sequence of covers going inwards, each with a different language on it, like it was some do-it-yourself forgery kit. 09:42
nine So....your unconcious expresses doubt towards your passport forger? :D 09:44
nwc10 seemingly yes. But this was supposed to be a genine passport. So clearly I was conned
MasterDuke is a jvm-backeded nqp/rakudo relocatable? 09:54
patrickb MasterDuke: JVM has no fancy wrapper programs as MoarVM does, only plain bat/sh scripts with their inherent drawbacks. I think I tried to make them relocatable as well, but I possibly never actually tried. 10:16
MasterDuke cool. we'll see what happens with github.com/Raku/nqp/pull/719
[Coke] Yo. 13:55
brrt \o 14:18
nwc10 o/ 14:22
[Coke] tries to imagine "Yo" as a variant of "o/" 14:50
maybe an Ent. 14:51
nwc10 Yo => ¥o => \o 14:53
[Coke] chuckles at nwc10's transformation path 15:48
MasterDuke arg. of course it's the job that takes an hour that fails...and at the end of the hour... 18:24
and i'm not sure why... 18:32
geth_ MoarVM/inline-inf-nan: 2442077453 | (Nicholas Clark)++ | 9 files
Convert MVM_num_{isnanorinf,posinf,neginf,nan} to inline functions.

They are all one-liners - likely the size of the inlined code is smaller than the calling convention setup.
  (Unless one is building with Link Time Optimisation, in which case *that*
will do the transformation, but even then, why make more work for ourselves?)
This has the side effect of moving `#include <math.h>` to a header file, meaning that various C files no longer need to include it.
20:17
MoarVM: nwc10++ created pull request #1486:
Convert MVM_num_{isnanorinf,posinf,neginf,nan} to inline functions.
20:19