freenode/#clasp - IRC Chatlog
Search
7:57:14
kpoeck
::notify merelystudent the missing symbol is defined here: https://github.com/clasp-developers/clasp/blob/master/src/gctools/threadlocal.cc#L18
7:59:03
kpoeck
::notify merelystudent So I wonder whether these 2 line should go to the header file merelystudent (and I forgot to tell you than while the demo works on macosx I get the same error on Ubuntu (and I believe selwyn as well)
11:16:21
merelystudent
@kpoeck, thank you. Which header file do you think they should go to? Plus, I notice that my_thread_low_level is not even used in threadlocal.cc
11:16:21
Colleen
merelystudent: kpoeck said 3 hours, 19 minutes ago: Your local config on linux is fine
11:16:21
Colleen
merelystudent: kpoeck said 3 hours, 19 minutes ago: the missing symbol is defined here: https://github.com/clasp-developers/clasp/blob/master/src/gctools/threadlocal.cc#L18
11:16:21
Colleen
merelystudent: kpoeck said 3 hours, 17 minutes ago: So I wonder whether these 2 line should go to the header file merelystudent (and I forgot to tell you than while the demo works on macosx I get the same error on Ubuntu (and I believe selwyn as well)
11:18:41
merelystudent
@kpoeck, nvm about my comment about threadlocal.cc not using it. This makes a stronger case for it to be in a header file
11:41:47
Bike
you shouldn't copy those lines out of threadlocal.cc into a header. in a header they should probably have extern
11:42:05
Bike
the declaration of my_thread seems to be in memoryManagement.h instead of threadlocal.fwd.h? bizarre
12:09:42
kpoeck
it compiles fine, but at startup says (copying this from irc) error: Symbols not found: { _ZTH19my_thread_low_level, __emutls_v.my_thread_low_level }
12:10:20
kpoeck
In memoryManagement.h there is extern THREAD_LOCAL core::ThreadLocalState *my_thread;
12:11:45
Bike
a linker error won't be solved by putting things in headers. but i don't know why definitions from threadlocal.cc wouldn't be around
12:41:33
yitzi
drmeister: I think that amber is still jacked up in my latest image. I'm giving up on the conda install and trying a source install.
12:44:20
drmeister
I'm getting errors about not finding cl-jupyter-user, something that we had in the old cl-jupyter but I don't expect in common-lisp-jupyter. What did you do about that?
12:46:47
yitzi
Oh. There is no cl-jupyter-user. The common lisp kernel runs in cl-user just like regular lisp. I set the cando kernel up to run in cando-user. There is probably some old reference to cl-jupyter-user. I'll look.
12:54:29
yitzi
drmeister: There are references cl-jupyter outside of "cando-jupyter" I can fix them all, but do you have an example so I can make sure it works?
12:57:16
drmeister
yitzi: I think we need to switch to common-lisp-jupyter now and fix the problems in clasp/cando in the source code.
15:57:34
kpoeck
cracauer The buildbot does not seem like subprocess.run in wscript, which python is installed on this machine?
16:11:48
kpoeck
drmeister on my mac it seem to work with 'import subprocess' instead of 'from subprocess import call'
18:10:31
drmeister
There is a cytoscape widget that works on jupyter notebooks but not lab yet (I think).
18:18:03
yitzi
Its a quick port of https://github.com/QuantStack/ipycytoscape/tree/master/ipycytoscape
18:19:17
yitzi
Anyways, didn't mean to derail the more important stuff, just thought you might like that.
18:19:39
yonkunas
Yes the image above posted in static obvi. the port is fully functional. This is going to give us a big edge.
18:21:15
yonkunas
yitzi: this is important, drmeister and I will be continue work on a function called LOMAP soon that will use this directly.
18:27:27
karlosz
Bike: did you get a chance to try using setjmp to translate some catch/unwinds with the analysis i wrote?
18:28:28
Bike
no, i've been working on something else.. i could switch tho. is your analysis in a branch or anything?
18:35:21
Bike
"How do I use llvm.eh.sjlj.setjmp and llvm.eh.sjlj.longjmp in LLVM? When I call *.longjmp it appears to just crash."
18:38:44
karlosz
is there a canoncial way to check for a constant input? i'm not sure if i should be looking at just load-time-value-input, an immediate, or something else