Anyone know how I make with -ggdb?

Probably you just need to run the binary under gdb, but it may help to use the racket3m binary in the src/build directory

I will open an issue with all this information and revisit it in the future.

Getting SIGSEGV maperr in scheme_gmp_tls_unload, s is not null, line 5822, running https://github.com/racket/gui/issues/118

It doesn’t make any sense to me, why would we sigsegv with a valid ptr here, and why only on the assignment, not the read?

@macocio That’s not an actual crash. It’s an intentional fault set up by the GC to implement write barriers. In gdb, use handle SIGSEGV nostop noprint and then continue.

@macocio I have this alias: alias gdbracket='gdb -eval-command "handle SIGSEGV nostop noprint" --args $PLTHOME/racket/src/build/racket/racket3m'

I see. Maybe it’s just a GTK and GLFW conflict then

Here’s the backtrace if anyone’s interested https://pastebin.com/gMDGU9fU

I answered a question about Racket on reddit - please let me know if I got anything wrong.

https://www.reddit.com/r/Racket/comments/9td40m/why_not_use_racket/?st=JO05VWS8&sh=6958b525

Sounds right: https://github.com/racket/racket/issues/1570

Oh - missed you had multiple comments :slightly_smiling_face:

@spdegabrielle I would suggest the community is now big enough that it would be easy to find a Racket consultant in your region. That’s quite optimistic but how are you getting this information?

I like the linkt to: https://www.techrepublic.com/google-amp/article/famous-developers-reveal-the-programming-languages-theyre-glad-they-learned/

I don’t think it is optimistic at all. Racket is loved by many developers who would jump at paying gigs. There is quite an active [Racket-Money] list and a significant portion of the community that doesn’t participate often on the list/Twitter/reddit/etc.

I’m not a recruiter but I have no doubt even I could find good people to do consulting work.

S.

Something about this change broke the package dependency checker for me: https://github.com/racket/racket/commit/8bed64f44f26d63bfdde816fb4b3d55a60b6d67f I now consistently get this output when I run raco setup: raco setup: --- checking package dependencies --- cdddr: contract violation expected: (cons/c any/c (cons/c any/c pair?)) given: '("6.90.0.28" ("6637c2dff775d4d7f586b77d93572c9d070cd00b" . "f9065fa7389750e16fe00d7ba36748f61d3e0df6")) context...: /Users/alexis/gits/racket/racket/racket/collects/setup/private/pkg-deps.rkt:478:8: for-loop /Users/alexis/gits/racket/racket/racket/collects/setup/private/pkg-deps.rkt:474:4: for-loop /Users/alexis/gits/racket/racket/racket/collects/setup/private/pkg-deps.rkt:465:2: for-loop /Users/alexis/gits/racket/racket/racket/collects/setup/private/pkg-deps.rkt:26:0: check-package-dependencies /Users/alexis/gits/racket/racket/racket/collects/setup/setup-core.rkt:71:0: setup-core "/Users/alexis/gits/racket/racket/racket/collects/setup/main.rkt": [running body] temp37_0 for-loop run-module-instance!125 "/Users/alexis/gits/racket/racket/racket/collects/raco/main.rkt": [running body] temp37_0 for-loop run-module-instance!125 perform-require!78 Do other people on HEAD get this, too? Or is it just me?

How about remove all compiled directories in the packages you installed and re-run raco setup?

Unfortunately, deleting compiled/ and recompiling does not seem to have fixed the problem.

those are very different programs

the first one has a submodule, the second one has a top-level module

so the “answer” is that the set of scopes expander changed something about how modules at the top-level inside begin work

which is not super-helpful

@mflatt will probably have some more detailed thoughts on exactly what changed

I changed the format of the “.dep” file rceently, and a file that claims “6.90.0.28" is clearly from before that change. So, raco setup should not try to inspect old “.dep” files. I suspect that @mbutterick saw something similarly related to the “.dep” change, and probably it was difficult to replicate because the problem “.dep” file got replaced/updated.

Oh, this might be because I have PLTCOMPILEDROOTS set to compiled/@(version):, and IIRC the dependency checker is not very intelligent when that environment variable is set.

@mbutterick can you just remove the begin?