github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
samcv jnthn: timotimo: these entries still need to be put into a proper changelog github.com/MoarVM/MoarVM/wiki/ChangeLog-TODO 00:05
i've dealt with most of the changelog entries but these still have to be dealt with 00:06
github.com/MoarVM/MoarVM/wiki/ChangeLog-TODO
timotimo oh, OK
samcv there were like 370 things. i have it down to like 115 00:07
timotimo nice work
jnthn samcv++
I can't promise much; I've over-committed for the next couple of days
Will see what I can do with the couple of minutes I have before going to bed 00:08
samcv ok :) 00:09
jnthn I'm also not sure what level of granularity we're aiming for 00:10
A bunch of them are related to context traversal refactoring, though I'm not quite sure what {foo} I should use for something that cross-cuts multiple subsystems... 00:11
Traditionally I'd have just written "* Refactored context traversal to better cope with inlining" 00:12
And left it at that 00:13
samcv i can put it in a Misc section or just without a section
just put {General} actually
jnthn Yeah
tbh, I mostly feel confused :) 00:14
My feeling was "ChangeLog is hand-wavey and folks who want details can `git log`"
samcv jnthn: yeah just write something general then 00:15
jnthn OK, but I guess we want to keep track of the SHA-1s that are summarized, or? 00:16
samcv yeah, i'm trying to do that now 00:16
jnthn On the Rakudo one I tried to help summarize my changes by `git log --author=jnthn` and writing them up, then went and looked at the change log, saw it was all sections and SHA-1s, sighed and that was about the end of my motivation to help with it. :S 00:17
samcv hah 00:18
jnthn Anyway, I gathered the context traversal ones 00:23
It feels like it did pure date ordering; grabbing the git log with --topo-order so that commits on a branch are together would perhaps be a bit easier 00:24
sleep time; 'night 00:26
timotimo gnite jnthn
Geth MoarVM: 45ff2b130c | (Samantha McVey)++ | tools/update-changelog.p6
Add ability to merge commits in update-changelog.p6

Remove for now the code to extract exprjit template additions for now.
03:33
MoarVM: b0b90623ee | (Samantha McVey)++ | tools/update-changelog.p6
Fix ability to add custom categories to update-changelog.p6

Add ability to dump things to a file, though I haven't yet added the option to choose which file to output to.
MoarVM: 633604d301 | (Samantha McVey)++ | tools/update-changelog.p6
Use --topo-order in update-changelog.p6

  jnthn++. So things part of the same merge show up next to each other.
travis-ci MoarVM build passed. Samantha McVey 'Use --topo-order in update-changelog.p6 03:52
travis-ci.org/MoarVM/MoarVM/builds/421877292 github.com/MoarVM/MoarVM/compare/0...3604d301e6
brrt we can make MoarVM bytecode be ELFs.... 12:03
and make them executables..... 12:04
jnthn We should deliver it at Christmas... :P 12:07
brrt :-D 12:08
hehe
I was actually just looking into it for the purpose of making the JIT output debuggable 12:09
it's not so hard, at least if you have the linux elf headers
nwc10 "other platforms are avaailable" - ie can it be done nicely without breaking stuff on platforms which are ELF but we don't have headers for? 12:10
brrt yes, but it is slightly harder since you have to basically do the definitions by yourself 12:11
With elf documentation in hand, writing an ELF formatter in C (or perl for that matter) is a job worth a day or two of work 12:14
robertle I think guile does that as well 12:17
brrt right 12:18
MasterDuke that would be cool 12:25
brrt puts it on the ever lenghtening list 12:31
masak .oO( "TODO: make this list shorter" ) 12:44
masak .oO( ...dammit! )
brrt on the other hand, having to redo all of MoarVM's serialization logic... is less than awesome 12:51
dogbert2 jnthn: should postrelease-opts be merged directly after the 2018.08 release or do you want to wait? 13:38
jnthn dogbert2: Wait. It breaks stuff. 13:58
dogbert2 jnthn: not many bugs left though :) 14:01