Clojurians Log v2
inf-clojure 2020-08-29
@dpsutton I added more notes to the ticket. I'm going to be AFK for a bit while we set up dinner, but can probably try more stuff later if you need me to...
i saw. i'm working through it. i can repro now. its weird and i don't know why Β―\(γ)/Β―
but i'm investigating. disappointed in how brittle it is π
Those bugs are almost the most fun π
yeah. its a thin wrapper around comint-mode so i'm hoping i'm not getting in there to figure it out
I confirmed that a docstring of "some \"thing\" like this" doesn't cause it BTW.
(so my hypothesis in the issue comment was wrong)
although i just killed the docstring from the ns and it eval'd just fine. i think it may be some escaping in that
get-datasource also blows up
yeah. just evaling the docstring blows up
Yeah, after trying to eval a few forms in next.jdbc, I just assumed there was something widely broken but now I've pushed at it a bit more, def seems to be something in the string π
Back in a bit
cool i'll bisect the string down to what's causing it
i think it might be related to the size of the form actually
so some headway. i tried running a scheme mode based on comint and got the same results. view the bytes and they are just repeating. was wondering if there was an encoding issue. but it seems if the form is too large it gets batched and that totally breaks
Oh, that's interesting... Interesting that it's always 07 (ctl-G)?
Yeah. Iβve opened a bug on the emacs mailing list. Can reproduce in scheme mode. Itβs in coming (and send-process-string) at the base
Iβll put a link to the bug after dinner. Sorry this is broken.
an older bug talking about the same issue although missing that its in comint i think
No worries. I'm not trying to use it in anger -- it is purely curiosity that is leading me to try this.
Unfortunately, this is just another reason for me to continue to avoid Emacs and its ecosystem π For all the years I used it (from 17.x through 19.x and then 24.x onward, on and off), I've found it to be a frustrating and brittle ecosystem π
Part of why I stopped using Emacs was the fragility of CIDER etc which kept breaking stuff on each upgrade early on -- and Bug was very dismissive about it and just told folks to uninstall and reinstall and then deal with the changed π
I gather it's all a lot more stable now...
From the bug report I found it might just be a buffer bug on OS X. Do you have a wsl handy? I can check on Linux in a bit
I was unable to update Emacs on WSL beyond 25.2 so I can't install packages (SSL / TLS restrictions).
I haven't bothered installing Emacs on my Windows machine for all the reasons discussed above π
Ha totally fair. Iβll check in about an hour
i've confirmed its a bug in emacs on mac. it doesn't seem to use process-send-string correctly
I wonder if it is also present if you run it in -nw mode? (i.e., inside a terminal)
I thought http://emacsforosx.com was supposed to be a fairly bare bones build of core Emacs without much/any Mac-specific stuff?
I build on Linux and itβs fine. On OS X I have a built version and a brew version with patches for OS X. Havenβt tried the built one yet
Gonna have to try my hand at some C soon :)
You're very dedicated π
i feel bad. the hope was for a simple interaction with a repl or a socket repl with essentially no tooling in the way π
wow. this is ... not trivial
our bug: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=43099