cfbolz changed the topic of #pypy to: PyPy, the flexible snake (IRC logs: https://quodlibet.duckdns.org/irc/pypy/latest.log.html#irc-end ) | use cffi for calling C | if a pep adds a mere 25-30 [C-API] functions or so, it's a drop in the ocean (cough) - Armin
<arigo>
I can now build a full win64 pypy, but I guess there is a JIT bug somewhere: it crashes by getting a KeyboardInterrupt(??) after running for ~1 minute; I'm not pressing Ctrl-C at all
<arigo>
but not if run with "pypy-c.exe --jit off"
<arigo>
it's because some variables are declared "long" in signals.c
<arigo>
which is 4 bytes on win64, but the JIT expects them to be 8 bytes
<mattip>
wow, that is great progress
<arigo>
so the JIT overwrites an unrelated variable by writing to one variable
<arigo>
VCs debugger kind of works nicely after a bit of the usual fight
<mattip>
if there is any kind of consistent way to make the debugger work, it would be nice to document it somewhere
<arigo>
ah, one variable is the decreasing counter (when it reaches -1 we check for signals), and next to it is a bitmask holding which signals occurred. The overflow from 0 to -1 in the first variable makes the second variable also switch from 0 to -1... boom, we suddenly have "received" all signals
Olorin has quit [Read error: Connection reset by peer]
Alex_Gaynor has quit [Read error: Connection reset by peer]
Guest68750 has quit [Read error: Connection reset by peer]
fijal has quit [Read error: Connection reset by peer]
<arigo>
cool, now translation seems to run for more than 1 minute
<cfbolz>
arigo: I would have expected way more problems from long=32
<arigo>
I guess nulano fixed the problems that show up when translating. now he's left with all the other modules, including cpyext, _cppyy and micronumpy (although the latter should probably be removed)
<arigo>
it's not really working, so far even ssl is not included (although unsure why, given that it's a cffi module, but I guess I didn't install X from random internet source Y first)