logs archiveBotHelp.net / Freenode / #ada / 2015 / July / 21 / 3
charlie5
windows
marble_visions
has anyone tried to do functional programming in some sense with ada?
darkestkhan
marble_visions: there was one guy, he even succeeded with it
although it was much more of subprogram access abuse than natural way it is done FP langs
marble_visions
ok, i might look into that
jk4
morning (+ planet correction)
i wish people would stop trying to force paradigms onto languages for which they weren't created
you can macro the bejeebes out of C to look like Ada but it still ain't Ada
darkestkhan
wow, seems like a ton of interesting stuff was added to Debian in last 24h
libregf - a library to access the Windows NT Registry File (REGF) format
no idea of what use this would be but whateve
r
Lucretia
charlie5: re swig, it would be nice if it were more visible, and iirc a bi**h to build. it would be cool to get the patches required into the mainline of the project, but make it easy to build.
charlie5
hi luke
Lucretia
hi
charlie5
just sat back down
yes, re vis
it isn;t a bi**h to build, i think ... i just built it on freebsd yesterday
Lucretia
and didn't you write the Ada extension to swig in Ada, twas amistake imo
charlie5
$ ./build_swig
Lucretia
and change your licence to match that of the project.
but yeah, I think it is worth carrying on with.
the more Ada projects, the better imo
charlie5
mm, not really imposed any licence on it yet :) ... but it is free as far as i'm concerned
"didn't you write the Ada extension to swig in Ada" ... huh, but but but, it is ada sir !
you thinki'd do it in c/c++ ?!?
:)
Lucretia
just to match the project :D
hey ho
yo ushould contact them about including it into their source tree
charlie5
i did
Lucretia
did they say "fu*k off"?
charlie5
they said i had to pass all 30 million of their tests 1st
Lucretia
ah
does it?
charlie5
i'd said i'd wait for volunteers to help and get back to them
that was about 10 years ago
Lucretia
aren't the tests automated?
charlie5
not especially, unless you do it yourself
Lucretia
hmm
charlie5
heh, and you should see their tests ... i had no idea just how gory c/c++ can get ... take a look out of interest ... better than any horror movie
but, yes ... i'm just bi**hing really ... it probly is worht continuing with given the alternatives
Lucretia
having many generators is better than none
and having a generator that can produce useable output is better than the gcc/g++ alternative
(Action) wonders if the objc/objc++ compilers also hav that switch
charlie5
gcc produced stuff struck me as ghastly ... they don't bother to use i.c.pointers ... and i doubt if they have support for layered c+ libs (ie one on top of another)
system.Address for pointers
ick
Lucretia
no, it's basically a sh*tty dump of the ir
I just went through sdlada remvoing address in favour of access to null record with conventionc
charlie5
one of my biggest probs was sorting out how to avoid circular deps
Lucretia
is hard enough by hand
but you'd need to implement a dep graph really for that and traverse it looking for cycles
charlie5
(Action) didn't g quite that far
go
Lucretia
well, today i need to work out how to get my gl bindings output
Natacha
btw in this topic, do you know whether you can have convention-C accesses to convention-Ada stuff?
charlie5
but just organising ada pkgs to cater for C core types + C types that use them + C types that use them/etc can be a challenge
Natacha
(typically to put a "this" pointer in a void* C callback)
Lucretia
Natacha: you can pass an Ada access type to a C function for a void*
I did it in my original wxAda bindings
whether it's right to do, I dunno
Visaoni
charlie5: just about to head to sleep, but re my project I think you might be right about CLA. probably need to spec out what I need well and just ask... pretty sure there is no good answer though
some of my latest ideas are probably as good as it's going to get, but have their flaws (that hopefully won't be run into by most people)
charlie5
Visaoni: yeah, cla is great place to get a 'definitive' answer (or consensus (or lack thereof :) ) from the Ada community
here is good also of course, for quick q's and social stuff
Visaoni
maybe I'll see if I can get it going with tagged types first. not convinced even that'll go nicely, but I can't say it won't either :)
charlie5
we even got a plug on cla the other day :)
tagged type classes fit my head better than generic OO ... i would prototype with tagged types and only switch to generic if there were some real performance gain, probly
Visaoni
well, part of the draw of this kind of implementation is that it would actually fit real nice on micros :)
Lucretia
Visaoni: must've missed it, what's the project?
Visaoni
PRNG, if you want to look at the thing more it goes by the name PCG
figured an Ada implementation would be nice
(and honestly thought this would be simple and done with long, long ago)
Lucretia
hmm, pcg is not an easy google
ahh
Natacha
WE ARE PCG \o/
Visaoni
try pcg rng
Lucretia
gotit
does Ada's rng suffer from the same problems?
Visaoni
it uses the Mersenne Twister
pretty much what you see as the common C++11 thing
Lucretia
(Action) shrugs
Visaoni
I think the pcg page has a quick summary of some of the pros/cons
just checked. Twister is the first thing after PCG itself :)
Lucretia
aye
are you wanting to know how to structre the code?
I take it, as it says "pcg family," that there are multiple algorithms with the same interface?
Natacha
I wasn't aware of problems in Mersenne Twister, I guess it's not as ultimate as it was during my PhD
Visaoni
it's not that Mersenne Twister is /bad/, it's just has some tradeoffs
charlie5
(Action) jots down Mersenne Twister as potential spell name for rpg
Visaoni
Lucretia: you could say that... it's sort of a framework almost
Lucretia
along with titty
so
type Root_RNG is abstract null record;
or
type Root_RNG is interface;
Visaoni
I've just run into problem after problem getting it all to work with a half-way reasonable interface and without copy/pasting sh*t
I've always been avoiding tags
*also
« prev 1 2 3 4 5 6 next »