IRC channel logs

2021-10-22.log

back to list of logs

<jab29>RhodiumToad That's true...I suppose that mutating state is one way to achieve high performance
<robin>jab29, mostly-functional style is pretty common for scheme
<drakonis>a #lang construct you say?
<jab29>drakonis That's what racket does.
<jab29>It's a way to change the "language". They have a logic programming language, that gets converted to racket via macros...
<jab29> https://beautifulracket.com/explainer/lang-line.html
<oriansj>any word on when guile is going to officially have a bootstrapped psyntax as part of the default release?
<jab29>oriansj What is psyntax?
<oriansj>jab29: it is the scheme code which enables guile to support syntax-case
<oriansj>that is written uses syntax-case and a generated expansion is distributed
<oriansj>that is why https://github.com/schierlm/guile-psyntax-bootstrapping was created for guile
<oriansj>as the goal should be to enable one to build guile from source and not require pregenerated binary blobs.
<daviid>lloda: i think we have a problem with the exception truncate patch and guile-3, though i can't yet tell exactly what/where, and so far can only reproduce it while using guile-cv, i might later today or tomorrow produce a test that does not depend on it (guile-cv) - but here is a very simple example, using guile-cv in 2.2 then 3.0, where i deliberately call a procedure using an unbound variable, in 2.2 we get the exécted error
<daviid>msg, in 3.0, it crashes guile - let me prepare and paste ... in a min
<daviid>lloda: here https://paste.debian.net/1216359/
<daviid>lloda: not that expect you to really use/nor understand guile-cv', but fwiw the im-rgba->rgb proc def is here - https://git.savannah.gnu.org/cgit/guile-cv.git/tree/cv/adds.scm?h=devel#n205 - basically, a series of nested calls to match, then parallel calls to im-rgba-channel->rgb-channel
<daviid>lloda: have to go, i'll try to write a similar code that doesn't depend on guile-cv tomorow ... sorry to bother you with incomplete bug/info, did it in case it rings a bell wrt your own large array processing experience... bbtomorow
<daviid>before i leave, here is a sinilar crash, but using a 'fake' 4x3-4(channels) image - https://paste.debian.net/1216360/
***taw10_ is now known as taw10
<lloda>hmm
<dsmith-work>Happy Friday, Guilers!!
<civodul>hey, happy Friday!
<civodul>hmm, am i right that we can't use (ice-9 vlist) in psyntax?
<civodul>or anything sophisticated
<cwebber>back from my trip!
<cwebber>hi hi
<cwebber>hi civodul !
<civodul>howdy cwebber!
<cwebber>:)
<cwebber>I'm SO RELIEVED to be home again
<cwebber>also
<cwebber>I used a lot of this trip to start bringing the tooling I've been working on to Guile
<cwebber>and planning how Guile fits into the future of it
<cwebber>and! I've been enjoying playing with dthompson's Chickadee
<cwebber>I nearly have Spritely Goblins ported to Guile
<cwebber>and
<cwebber>I've been figuring out how to get Goblins, including its time-traveling features, on top of Chickadee
<cwebber>civodul: it may be interesting at some point to look at putting our macro system on top of Matthew Flatt's "bindings as sets of scopes" stuff, maybe maybe maybe
<civodul>cwebber: woow, lots of nice things that you did! :-)
<civodul>re psyntax, i was looking at whether/how we could use a data structure more efficient than alists for the expansion-time environment
<civodul>not ambitious, yet pretty challenging
<civodul>it's all car cdr set-cdr! in there
<cwebber>civodul: oh interesting
<cwebber>civodul: I've not spent too much time looking at the vlist implementation but I do appreciate that we have them
<cwebber> https://www.cs.utah.edu/plt/scope-sets/
<civodul>cool, thanks for the link!
*civodul -> home
<civodul>later!
<cwebber>later civodul
<dsmith-work>Is a release immanent? (noticed a NEWS update...)
***jackhill is now known as KM4MBG
***KM4MBG is now known as jackhill