A suggestion to add a racket-5.2 ebuild Reproducible: Always It looks like racket-5.2 is ok for inclusion in portage. Current latest in portage is 5.1.2. No changes needed to be made to the ebuild. I just copied it into my own overlay as racket-5.2.ebuild, and installed. Just tested it out with some (cool!) new features added in 5.2, and looks good so far. FYI, my USE flags (global & package specific) were: + X cairo doc futures jit places plot threads - backtrace
Sorry, forgot to add, I installed & tested on amd64
Unfortunately I've encountered multiple yet-unsolved big bugs, both with the new plot and with the syntax checker. I'm helping upstream debug these issues, but until they're solved I won't be adding any new releases.
Marijn, I just checked your mail on racket-dev "check-syntax (require plot) hangs drracket". I just tried the following snippet in drracket: #lang racket (require plot) (plot-new-window? #t) (plot3d (list (surface3d (λ (x y) (+ (sqr x) (sqr y))) -1 1 -1 1 #:label "z = x^2 + y^2") (surface3d (λ (x y) (- (+ (sqr x) (sqr y)))) -1 1 -1 1 #:color 4 #:line-color 4 #:label "z = -x^2 - y^2"))) And it works fine (check-syntax doesn't freeze, and the plot works as advertised). You mention in that thread that you're working from a local checkout from git master. I don't know if any post-5.2 commits might be causing the issue. All I can say is that a fresh install from a racket-5.2.ebuild isn't showing the same problems on my end. Perhaps if you just added the 5.2 ebuild to the tree, and others also tried it - we might see if the problems you're hitting are replicated by others. PS: are you compiling with any different configure directives than my USE flags (see above) would enable/disable?
dev-scheme/racket-999 was built with the following: USE="X cairo futures jit (multilib) places plot threads -backtrace -doc" so the only difference to you seems to be the doc use flag and the higher version. I guess it's possible that 5.2.0 works fine while all my git checkouts don't, but I find this hard to believe and also given that upstream has such a hard time reproducing it seems more likely to have been overlooked. Anyway, I might give 5.2.0 a spin.
(In reply to comment #4) > dev-scheme/racket-999 was built with the following: > USE="X cairo futures jit (multilib) places plot threads -backtrace -doc" > > so the only difference to you seems to be the doc use flag and the higher > version. I guess it's possible that 5.2.0 works fine while all my git checkouts > don't, but I find this hard to believe and also given that upstream has such a > hard time reproducing it seems more likely to have been overlooked. Anyway, I > might give 5.2.0 a spin. Marijn, With all due respect (no offence intended), I think you're going about this the wrong way. racket-5.2 does build and install correctly on gentoo, just by renaming the most recent ebuild. The issues you're running into are coming from your local builds straight from racket's git/master (as per your mails on racket-dev). Thats got nothing to do with racket-5.2. If it builds, installs and runs (as I have verified with some of the new racket 5.2 features), then it should be included in the tree. If only to allow other users to try it out, and then file bug reports of their own upstream, if racket-5.2 breaks in some other ways. I'm sure racket-5.2 has other bugs. I mostly use chicken-scheme and librep myself, so I haven't banged too hard on it (I really only use racket for its integration with vim). But thats not the point. *ALL* software has bugs. racket-5.2 *does* build, install and run on gentoo. Surely thats all thats needed to include it in the tree. If its a better or worse release than 5.1 is besides the point, and not a gentoo issue. Upstream saw fit to cut a 5.2 release, and its buildable, installable and runnable on gentoo. Right now. Please reconsider.
