1:34:24drmeisterBike: There is an issue though - I actually started trying to debug the static analyzer problem just using bytecode. When I `ninja -C build load_cclasp-boehm` it dies trying to unwind from the first error.
1:34:34drmeisterMaybe your funwind_protect stuff will fix that.
1:34:42drmeisterDo you have a way to get around the funwind_protect yet?
1:34:55drmeisterOr karlosz was talking about optimizations that would eliminate the need?
1:35:25Bikei mentioned it earlier, but i have something that should be cheaper than the unwind protect
1:35:36Bikenot as cheap as just not doing dynamic environment stuff, though
1:36:01Bikei'm not sure how to do it that way correctly. i think the unwinder needs to know about bytecode frames some way or another.
3:21:30drmeisterI'm building cclsp-boehmprecise in my local vm branch now that I merged vm-boehm into it and once it's done I'll push.
3:38:36Bikenice, and the analyzer works and all? ok. tomorrow i can merge in all the code deleting i've been doing, and the funwind protect replacement
3:46:52drmeisterBuilding cclasp-boehmprecise builds like vclasp-boehmprecise did. It loads everything in as bytecode and then compile-files everything.
3:47:42drmeisterI have turned on the funwind-protect thing in bytecode_call so that I don't have to worry about VM issues.
3:57:38drmeisterThere is still a problem with the bytecode compiled code. It I `ninja -C build load_cclasp-boehmprecise` and then evaluate `(error "foo")` and then use `:r1` restart - it segfaults.
3:58:26drmeisterThis is where it does `sjlj_continue_unwinding` when the dynamic environment that it is unwinding to has already been removed from the dynenv stack.
3:59:51drmeisterThe funwind_protect is enabled in bytecode_call - so that's not the issue.
4:00:26drmeisterLet's look at it tomorrow. It doesn't break anything other than the idea of doing everything in bytecode.