freenode/#clasp - IRC Chatlog
Search
13:59:43
kpoeck
beach quicklisp takes flexichain from https://common-lisp.net/project/flexichain/download/flexichain_latest.tgz
14:00:02
kpoeck
See https://github.com/quicklisp/quicklisp-projects/blob/master/projects/flexichain/source.txt
14:01:29
kpoeck
or should the source for quicklisp be switched to https://github.com/robert-strandh/Flexichain ?
14:21:14
kpoeck
So cold a ask Zach for quicklisp to switch to https://github.com/robert-strandh/Flexichain ?
14:34:34
kpoeck
Just quicklisp does not get it sources for flexichain from https://github.com/robert-strandh/Flexichain
14:35:29
kpoeck
Took me some time to find out, obviously for every project one can configure where source are coming from, for flexichain this is configure in https://github.com/quicklisp/quicklisp-projects/blob/master/projects/flexichain/source.txt
14:37:53
kpoeck
and loading mcclim I saw a message No support for weak pointers in this implementation. Things may get big and slow.
14:40:38
kpoeck
frgo put clasp definitions for clasp in https://github.com/trivial-garbage/trivial-garbage/blob/master/trivial-garbage.lisp#L126 and following
14:51:22
Bike
i'll have to look over all these portability libraries and make sure everything's exported from ext
14:54:34
kpoeck
Did create and quicklisp issue to change the source for flexichain https://github.com/quicklisp/quicklisp-projects/issues/1804
14:57:46
drmeister
Bike: can we bypass unwind_find_fde? Otherwise we run into the same contention problem don’t we?
15:07:44
Bike
based on the sources there's code paths where it uses atomic operations instead of locks. maybe we can figure out how to do that.
15:21:03
drmeister
I know people in the llvm community that might answer questions about exception handling implementation on Linux. Gather questions and we can ask them.
15:23:44
Bike
i don't know that i have any questions. it seems likely to me that they use locks because C++ code has never had this problem before
15:26:01
Bike
i mean i might have questions about exception handling more generally in the course of making the custom personality work, but that's different
15:36:15
Bike
C++ exception handling on linux sucks apparently with multithreading. We do a lot of that right now so we get poor parallel performance in code that unwinds a lot. they're not likely to fix anything on their end, but once cleavir2 is in we'll unwind a lot less and it will be less of a problem. in the meantime, it just means we don't get much of a speed improvement.
15:39:51
Bike
in the meantime i can also get the personality working, which should save some time, though not reduce contention.
16:46:47
kpoeck
do we have source-location information for constant or variables? EXT:*SOURCE-LOCATION-KINDS* -> (:CLASS :METHOD :FUNCTION :COMPILER-MACRO :METHOD-COMBINATION :TYPE :SETF-EXPANDER) doesn't seem to fit constant or variables
17:50:04
drmeister
What is you mental timeline for cleavir2 integration? Can we implement the unwinding codenow?
18:02:01
drmeister
If it's the same way that beach is going to organize it - then could we not organize ours the same way and start using it?
18:56:47
drmeister
What would it take to implement cleaver2-like stack unwinding. You have the hir instructions already don’t you?
19:08:37
Bike
also i'm not sure beach has finished working out all the details of local exits out of unwind-protects, and stuff.