These are chat archives for dropbox/pyston

2nd
May 2015
Marius Wachtler
@undingen
May 02 2015 13:03
googles protobuf lib contains a python and a capi implementation. Do we need to support the capi implementation? (It think we would need to have to patch the library if we want to use it)
Things which we would need to patch would be ref count accesses and they are directly accessing the fields of cpython objects (don't use the macros/functions to access the value). I'm not sure yet in how many places would need to be changed.
Marius Wachtler
@undingen
May 02 2015 13:22
it looks like the default is to use the python implementation and the cpp implementation will only get used when a special environment variable is set... so it's probably fine to skip getting the c implementation to work for now.
Marius Wachtler
@undingen
May 02 2015 14:30
man we should have put an assert inside intLShiftInt() in order to catch the case when an int overflows and which is not yet implemented.... took me a while to figure out while protobuf does not work...
Travis Hance
@tjhance
May 02 2015 20:32
"Versions 1.7.0 and later of git contain an annoying change in the behavior of git submodule.
Submodules are now regarded as dirty if they have any modified files or untracked files, whereas previously it would only be the case if HEAD in the submodule pointed to the wrong commit."
Sigh.
Travis Hance
@tjhance
May 02 2015 22:41
you can… resize tuples
wat
wat
"You can also think of it as creating a new tuple object and destroying the old one, only more efficiently."
ah, hm
(comment from cpython) /* Empty tuples are often shared, so we should never resize them in-place even if we do own the only (current) reference */
… what
what does this even mean?
do empty tuples break cpython’s garbage collection abstraction??
Travis Hance
@tjhance
May 02 2015 22:50
what is this supposed to do if the passed object whose type is a subclass of tuple?
a CAPI class, say
oh, cpython disallows it completely