github.com/moarvm/moarvm | IRC logs at colabti.org/irclogger/irclogger_logs/moarvm
Set by AlexDaniel on 12 June 2018.
timotimo so the getcode + push stuff comes from the `list_b` desugar, and is used in QASTCompilerMAST in `method deserialization_code` 10:25
brrt \o 17:43
tellable6 2020-02-03T19:47:39Z #moarvm <patrickb> brrt: Spontaneous GSoC project idea: A mod_moar Apache or Nginx module. You did mod_parrot back in the day. What's your opinion about such a thing? Worth it? Could you imagine mentoring?
2020-02-03T19:48:23Z #moarvm <patrickb> brrt: By the way, did you hear back from pamplemousse?
brrt ohai patrickb 17:43
patrickb o/
brrt I haven't heard from pamplemousse recently (though I also forgot whether I reached out)
I don't see much in in-server modules at this point, to be honest 17:44
but maybe there is a FCGI-style protocol that someone could implement
jnthn One could in principle host a Cro app via the FCGI protocol, though that ain't implemented yet :) 17:45
patrickb Would a Cro FCGI connector be worthwhile to have from a technical point of view?' 17:50
jnthn I guess it's useful to have a range of deployment options 17:51
Though it's the kind of thing I suspect I could hack out in < 1 day, and so not on its own a GSoC project.
Though I suspect there may be various Cro things that together could be of use, if we can find an overarching theme for them. 17:52
brrt hmm, there's that, too :-) 18:00
yeah, infrastructure libraries probably aren't the best fit for a GSoC project
re: backward branches, my two best solutions are: 18:04
- keep a bitmap of all emitted labels, so we know which ones are backwards
- ensure, somehow, that the labels emitted always increase, and maintain the highest one that's recently been emitted
I think we're almost at the second of these options already, except that we have different classes of labels, 18:10
timotimo jnthn: would it be a very bad idea to have frames share bytecode if they have the same? like the extremely common "internal error: execution of block eliminated by optimizer" one that shows up more than a thousand times in CORE.c.setting 23:54
or maybe we can do something smarter with the blocks that were eliminated, that'd be better of course 23:55