cfbolz changed the topic of #pypy to: PyPy, the flexible snake (IRC logs: https://botbot.me/freenode/pypy/ ) | use cffi for calling C | mac OS and Fedora are not Windows
<njs>
on linux, gettimeofday and clock_gettime are already optimized C code that runs in your process and does all the system-specific conversions for you
<njs>
so trying to parse /proc/cpuinfo may not make much sense
<fijal>
njs: it's still a syscall though right?
<njs>
nope
<njs>
Spooky Magic™
<fijal>
heh ok
<fijal>
then it has changed indeed
<njs>
the kernel creates an ELF binary in memory, that contains a data page that is actually the real page where the kernel's internal timekeeping stuff stores its metadata, and also some code pages that know how to combine rdtsc and the metadata to implement some of the key timekeeping "syscalls"
<njs>
then it maps this ELF binary into processes, and ld.so loads it into your binary
<njs>
(they used to do it by just mapping a raw page of code into every process at a known address, and glibc's implementation of gettimeofday would jump off to this random hard-coded address, but that turned out to cause security problems so now it's an actual ELF binary that can participate in address-space layout randomization)
redj has quit [Read error: Connection reset by peer]
alexband_ has quit [Ping timeout: 250 seconds]
arigo has joined #pypy
arigato has quit [Quit: Leaving]
arigo is now known as arigato
antocuni has joined #pypy
moei has joined #pypy
antocuni has quit [Ping timeout: 240 seconds]
dddddd has joined #pypy
lritter has joined #pypy
forgottenone has quit [Quit: Konversation terminated!]
krono has joined #pypy
<krono>
Hi PyPy folks again!
<krono>
I got 1.5 questions:
<krono>
1.5) if not, can I alter the inheritance chain programmatically? (Eg, Make C the A and A a subclass of C)?
<krono>
1) can I alter "inherited" _immutable_fileds_? (eg A has ['foo?'] and B(A) should have ['foo']
fryguybob has joined #pypy
vstinner has joined #pypy
dddddd has quit [Ping timeout: 252 seconds]
<vstinner>
arigato: hello. i like your "typedef uintptr_t PyHandle;" idea :) but would it make sense to start with "typedef PyObject* PyHandle;" in a first milestone?
dddddd has joined #pypy
antocuni has joined #pypy
<cfbolz>
krono: if the base class has an immutable field foo, the subclass will always have the same immutable field foo
<krono>
it has a quasi-immutable
<krono>
could someone refresh my memory concerning the difference between _immutable_ and _immutable_fields_?
dmalcolm has joined #pypy
dmalcolm has quit [Ping timeout: 250 seconds]
tayfun26 has quit [Quit: tayfun26]
jcea has joined #pypy
dmalcolm has joined #pypy
redj has joined #pypy
dmalcolm has quit [Ping timeout: 268 seconds]
forgottenone has joined #pypy
dmalcolm has joined #pypy
jacob22__ has quit [Ping timeout: 244 seconds]
dmalcolm has quit [Ping timeout: 272 seconds]
dmalcolm has joined #pypy
vstinner has quit [Quit: leaving]
antocuni has quit [Ping timeout: 268 seconds]
<krono>
(err, is there a replacement for botbot.me, yet?)