Search
Monday, 26th of June 2017, 13:41:49 UTC
13:47:27
Bike
is the dev-array branch on github in use? it looks like it was merged
13:47:58
drmeister
No - there are no dev-xxx branches in use at the moment. They can all be cleared
13:48:28
drmeister
Not dev-cffi - leave that one - and any you are using
13:49:08
drmeister
I should check that one first
13:51:29
Bike
dev-mbx is one commit ahead, devtype is eight ahead. though that's of master. i'll just leave them.
13:54:16
Bike
well, whatever, i just wanted to know if i could use the name dev-array
0:19:40
Bike
ok i have no idea how cmpliteral works now.
0:20:26
Bike
my fix didn't work. i tried reversing the order of run-all calls, and it went from a segfault to a null pointer reference
0:20:32
Bike
i don't know how that's even possible
0:23:24
drmeister
What? You told me how to write it!
0:28:05
drmeister
Could you paste that minimal example?
0:30:12
Bike
file with (defun foo () (cons 'special-variable (load-time-value #.(find-class 'special-variable))))
0:34:31
drmeister
Is that what you meant? 'special-variable' isn't a class so the find-class fails.
0:34:46
Bike
er, no, sorry, you have to define a special variable class. empty is fine.
0:34:55
Bike
(defclass special-variable () ())
0:35:08
drmeister
https://www.irccloud.com/pastebin/NhjOXb7l/
0:35:25
Bike
No, it can't be in the file
0:35:31
drmeister
No - I have to do it outside of the file - right
0:37:27
Bike
so the run-all should include like (setf [special-variable] (intern "SPECIAL-VARIABLE" "CL-USER")), then (setf [ltv] (find-class [special-variable]))
0:37:45
stassats
(load-time-value #.(find-class 'special-variable)) sounds a bit superfluous
0:38:58
drmeister
stassats: We have a problem cropping up because of how I chop up the RUN-ALL function to keep it short enough for LLVM to compile it quickly.
0:39:02
stassats
but do consider (load-time-value (find-class 'foo)) and (setf (find-class 'foo) nil) (defclass foo () ())
0:40:29
Bike
you... might have a point there
0:40:44
Bike
(but this should definitely not cause a segfault)
0:41:38
Bike
that does mean i can edit my thing so that it will probably not break, though
0:43:24
drmeister
Hmm, it works fine in cclasp - were you aware of that?
0:43:39
Bike
i don't think it worked fine for me
0:44:12
drmeister
It crashes in bclasp with "NULL pointer error"
0:44:18
drmeister
It crashes in bclasp with "NULL pointer error"
Tuesday, 27th of June 2017, 1:41:49 UTC