nullcone has quit [Quit: Connection closed for inactivity]
rgherdt has quit [Quit: Leaving]
nullcone has joined #racket
catonano_ has joined #racket
badkins has joined #racket
catonano has quit [Ping timeout: 256 seconds]
catonano_ is now known as catonano
badkins has quit [Ping timeout: 265 seconds]
TCZ has quit [Quit: Leaving]
wreed has quit [Quit: Goodbye]
bitmapper has quit [Ping timeout: 265 seconds]
<winny>
erkin: it's all gold :)
_whitelogger has joined #racket
orivej has quit [Ping timeout: 256 seconds]
orivej_ has joined #racket
<erkin>
winny: Worth sacrificing a tree? ;-P
Fare has joined #racket
<winny>
my ereader kills virtual trees daily :)
<erkin>
Haha
<winny>
screen is large enough to permit PDF reading
<winny>
maybe racket needs a packaging guide
<winny>
and recommendations for QA testing upcoming releases, since above discussion suggests it's not entirely clear how to do it
<winny>
(QA testing the packaging processes, that is)
<winny>
[for what it's worth, it's not possible to use a snapshot tarball for a "live" or "current" package on some distros, such as gentoo, where "live" packages are supposed to be from SCM only]
<samth>
winny: the packaging process is pretty heavily QA'ed already, but we're adding something to avoid this problem too :)
<samth>
that's a silly rule
<samth>
bremner: i think git being different from tarballs is pretty common in complex-to-build software
<samth>
certainly it's the case for web browsers, for example
dddddd has quit [Remote host closed the connection]
Fare has quit [Ping timeout: 246 seconds]
<winny>
samth: i think the issue comes down with as it is, there is no easy way to build git reproducibly with the same checkouts
Fare has joined #racket
<winny>
since there isn't a way to pin git revisions to the external artifacts at this time
<samth>
winny: but that's not a tarball vs git issue, the problems are the same in both cases
<winny>
right, so the issue is more universal than building from git
<samth>
and there is, of course, a way to do that, since that's how the release work
<winny>
but it's not pinned per se, but pinned based on time of creation alone?
<samth>
but anyway, I understand why having that be easier would be nice, I just don't understand bremner's issue with builds
<samth>
winny: I'm not sure what you mean by that
<samth>
git checkouts serve a quite different audience than tarballs
<winny>
i think bremner was expressing how the steps to build from git and tarball differ quite a bit, making it challenging for packagers to correctly package git (so issues like the above don't happen)
<samth>
winny: I think the right way to fix that, if you don't want to use snapshots, is to just create the tarballs, which is easy
<winny>
it's worth noting *nix packaging was impacted on 7.4, 7.6, 7.7 releases and this comes down to packagers not putting in the effort to build and package from git
<winny>
(if packagers did, they would have reported the build issues before release day)
ArthurStrong has quit [Ping timeout: 260 seconds]
ArthurStrong has joined #racket
badkins has joined #racket
badkins has quit [Ping timeout: 256 seconds]
corpix has quit [Ping timeout: 240 seconds]
efm has quit [Ping timeout: 264 seconds]
badkins has joined #racket
<winny>
for what it's worth, snapshot releases are still releases, with all the upkeep they entail. hence the motivation to use a SCM managed "live" repository for source. but as it is, it appears in order to build a full racket release from git, you need network access when building (instead of fetching all artifacts before build phases). it also means distfiles/git repos needed aren't 100% trackable by the distro
<winny>
package manager (if it supports such a thing, such as gentoo ebuild). Please correct me if i'm wrong. the build process is confusing at best, and i've basically given up on doing git builds
corpix has joined #racket
efm has joined #racket
<winny>
i could start issueing packages from snapshots, but i see the amount of effort to correctly maintain them getting out of hand pretty quickly (deleting old ones, adding new ones as they appear on the snapshot site, and having to do a full packaging cycle for each snapshot)
<winny>
this also implies each user of the repository syncs (whereas a true "live" package is buildable without versioning -- it's just built from SCM repositories [or in our case, perhaps some additional artifacts fetched as a tarball somehow(?)])
badkins has quit [Remote host closed the connection]
<bremner>
constptr: it can. I don't know it it's required. I'd guess not.
<bremner>
constptr: ooc, why?
<constptr>
I just wanted to know which backend does racket uses to compile
<bremner>
it uses liblightning as a jit
<constptr>
Ok
<bremner>
although with Chez Scheme that's presumably different
<bremner>
Racket CS probably uses whatever chez scheme uses
<constptr>
Isn't there AOT compiler for racket ?
<bremner>
not really
<bremner>
although, again, that might change with racket CS.
<constptr>
Wouldn't AOT improve speed of execution ?
<constptr>
For racket? I mean
orivej has joined #racket
<samth>
constptr: it all depends on what you mean by AOT, but Racket CS (mostly) compiles to machine code when you run `raco make`. Racket CS compiles to machine code always, but it will do it on demand if the file isn't already compiled.
<tlcu>
emacs has a really good racket mode, might wanna invest in evil for that alone.
laxask has joined #racket
<narimiran>
'slimv' was quite nice when i tried it with common lisp, but with racket i only got some strange hang-ups. i'm guessing i need to do some configuring, and i don't know exactly how....
<bremner>
at least on the emacs side, slime is common lisp only.
<narimiran>
in slimv's readme it mentions clojure too, so i've guessed that any lisp might work :/
<bremner>
ah. that might be optimistic
<narimiran>
probably :)
<narimiran>
"Lisp (any Common Lisp with SLIME support) or Clojure or MIT Scheme (Linux only)"
<narimiran>
so my workflow with an open file and racket's REPL is the best i can do in vim?
<winny>
samth: thanks, in the case of building an installer from git, i'm not sure how to express that dependency, since it needs the same software it's to be building, at build phase (bootstrapping problem). usually that's handled in tree, but in this case it isn't.
<winny>
anyway it seems from what you said indeed it's not really packagable from git as it is
<winny>
since there is no way to grab all artifacts before build phase, without having the same software you're trying to install, installed
badkins has quit [Remote host closed the connection]
badkins has joined #racket
<winny>
hopefully I manage to find some time to use the snapshots =) I'll share any challenges that might ensue (first thought: how do i automate this?)
badkins has quit [Ping timeout: 260 seconds]
davidl_ has joined #racket
notzmv has joined #racket
ArthurStrong has joined #racket
catonano has quit [Ping timeout: 258 seconds]
jmiven has quit [Quit: bye]
jmiven has joined #racket
tlcu has joined #racket
true-grue has quit [Read error: Connection reset by peer]
TCZ has quit [Quit: Leaving]
badkins has joined #racket
selimcan has quit [Ping timeout: 264 seconds]
dvdmuckle has quit [Remote host closed the connection]
dvdmuckle has joined #racket
Fare has joined #racket
Ekho has joined #racket
SGASAU has quit [Remote host closed the connection]