Updates aren't automatic yet. If you change your interpreter, please let me know so I can pull the updates.
Helen
@Helen0903
Alright
stasoid
@stasoid
@DennisMitchell Thank you. Did you have problems running phase0 or did you try to build latest Pyret version? Looks like machine with a lot of RAM is needed to build it (-max-old-space-size in Makefile is 8GB). Compiler flags are in src/arr/compiler/pyret.arr, -compiled-dir is useful for the wrapper not to write to /opt/pyret/compiled/ (I didn't try it myself).
petelomax
@petelomax
re phix: Thanks for trying - I spent 10 mins proving it works fine here before getting a similar segfault after a tiny change. Will get back to you when/if I manage to fix it.
DennisMitchell
@DennisMitchell
@petelomax I found the following errors in my logs, for the 64-bit and 32-bit versions.
p64: Uhuuh, elf segment at 0000000000400000 requested but the memory is mapped already
p32: Uhuuh, elf segment at 0000000008048000 requested but the memory is mapped already
DennisMitchell
@DennisMitchell
@stasoid I was able to build phaseA and phaseB just fine; TIO's 2 GiB of RAM and 4 GiB of zram seem to be sufficient. phaseC didn't build, but not because of lack of RAM. While it worked most of time, it sometimes failed to run for no apparent reason. I got stuck on that and didn't even try to make it work in the sandbox...
Helen
@Helen0903
@DennisMitchell I've pushed some updates to bitch
DennisMitchell
@DennisMitchell
@Helen0903 Pulled.
Jens Reidel
@Gelbpunkt
@DennisMitchell you may pull Mamba :) I'll fix one last bug when I got the time to, but as of now there's only improvements and v 0.2.0 is released
Jens Reidel
@Gelbpunkt
It now also features a CLI tool shipped with the language, which didn't work before. mamba --version can be used to check if it installed correctly.
Helen
@Helen0903
@DennisMitchell Found some bugs in bitch so I've pushed again
@DennisMitchell Pushed some changes to fix input problems in tio if you don't mind pulling
stasoid
@stasoid
@DennisMitchell Can we change DOBELA implementation to this one? Old asm interpreter doesn't follow the spec in some cases, which has bitten me once already. I believe new implementation is better in many respects, including maintainability, following the spec and handling many corner cases which dobelx64 doesn't. Currently DOBELA answers exist only in Hello World and Add a language to a polyglot threads. Changing the interpreter will break several old polyglot answers (207-210, all of them are mine) and the program provided in the comments of Hello World answer. All these breakages are due to dobelx64 not following the spec. If this is a problem I can fix all those answers for new interpreter, but I see no point in doing this. If that is fine, I'll write a pull request.
Tesla Ice Zhang
@ice1000
Requesting Agda 2.6.0.1 release, and a stdlib update
This example does not typecheck due to old compiler impl
DennisMitchell
@DennisMitchell
@stasoid I could host both implementations, but if dobelx64 just deviates from the spec in undocumented ways, it's probably not worth it and we should just switch to your implementation.
@ice1000 I'll have to update to Fedora 30 to update Agda (so it will take a while), but even Fedora 30 is still on Agda 2.5.4. When was Adga 2.6 released?
stasoid
@stasoid
@DennisMitchell No, dobelx64 behaves strangely due to bug(s?), not because it deliberately deviates from spec. The author of dobelx64 is actually the one who wrote the spec. Now that I think of it, dobelx64 has some value of being a "classical" implementation, so it's better to keep it. I will write a PR today or tomorrow.
It would be great to add syntax colouring at least for the most used languages.
Taylor Scott
@TaylorAScott
Hey there, I was hoping that we could get the version of Yabasic updated from 2.81 (Oct 2019) to the current verion 2.85 (Nov 2019). There are quite a few feature differences that are really important for golfing
(namely the additions of bitwise shifting functions, removal of parens from while statements and the addition of the chomp$ function)