libera/#clasp - IRC Chatlog
Search
1:50:45
drmeister
I found a way to debug the slime problem: error in process filter: Invalid rpc: Thread not found: 1
2:42:30
drmeister
The sly people said that clasp's weak pointers have a problem - that they get zeroed out too early.
2:43:08
drmeister
Weak tables are different. These are regular hash tables with weak-pointers as values.
2:52:21
drmeister
What leads you to believe that our weak pointers are broken. I'm not challenging - I'm looking for a reproducer so I can dive in and fix them.
3:26:32
Bike
the weak tables code is really messy and i think it's broken or near broken, and i figure that means weak pointers probably are too
11:53:13
yitzi
drmeister Bike: I've hit a wall on both Arch and Ubuntu during build. I am getting this https://plaster.tymoon.eu/view/2526#2526
12:49:26
Bike
that's not in std::random but i guess i can write out the mersenne_twister_engine specialization
12:50:59
Bike
or i can just switch it to mt19937 and see what happens. we probably don't have any code lying around that depends specifically on mersenne twister parameters
13:04:05
Bike
"The high_resolution_clock is not implemented consistently across different standard library implementations, and its use should be avoided." c++ why are you like this
13:09:32
yitzi
There is several different references to libunwind in the linking comand https://plaster.tymoon.eu/view/2527#2527
13:12:27
yitzi
I also don't see --unwindlib=libunwind or -rtlib=compiler-rt ... maybe they got shoved into an environment variable?