fijal

23 ms·

241.
by fijal 12y ago
It's mostly companies (in the terms of money donated, not the number of donors). We mentioned them on the blog post if they want to (a lot want to stay anonymous)
242.
by fijal 12y ago
PyPy dev here. I would be interested to explore the ideas what can be done to improve our funding situation. We seem to have moderate success in those fundraisers, but nothing REALLY huge.
243.
by fijal 12y ago
It's documented in PyPy differences, but generally open('x').read() does not guarantee that a file is closed. In CPython it's not a problem due to refcounting, but in pypy, the file might be closed at some later stage. T
244.
by fijal 12y ago
I tried talking to them, at the time they were sold on using JVM (noone got fired for using JVM for your VM). Too much hype, too little cold thought
245.
by fijal 12y ago
er, please don't reuse a well known term (peer to peer) to mean something completely different, notably a centralized company
246.
by fijal 12y ago
Bar does only shuffle the problem around to make it "someones else" problem. Maybe having a bench not be a cool option is an answer to those problems?
247.
by fijal 12y ago
this is really cool. The "click to ride" took me a while to find, some introduction video maybe would be cool? EDIT: and yes, the grey pane with no indication it's loading is annoying
248.
by fijal 12y ago
If you want to donate to a project (e.g. twisted, wine, whatever), 90% goes to the project and 10% goes to the SFC. The 10% is mostly covering operational costs as the SFC is incredibly helpful. We (PyPy) are working with them for quite a
249.
by fijal 12y ago
We can't really be creating accounts everywhere for minor donations. I support your sentiment, but the official PyPy bookkeeping has to be done in a proper way via the Software Freedom Conservancy, so going through all those services f
250.
by fijal 12y ago
here is mine - https://gratipay.com/fijal/ - not exactly killing it
251.
by fijal 12y ago
While we're here, please donate to PyPy STM, it's purely a crowdfunded effort http://pypy.org/tmdonate2.html
252.
by fijal 12y ago
"Node.js runs JavaScript faster than anything runs Python, PyPy included" that's a [citation needed] right here. Do you have some facts that I don't happen to have? Please share. That said, I don't care about a recu
253.
by fijal 12y ago
how do you make an immutable cache that you're populating all the time? (same goes for the graph)
254.
by fijal 12y ago
The problem with multiple processes is the "share nothing" model. It works for some problems, but it blatantly fails for a whole variety of other problems. STM tries to address those problems where "share nothing" does n
255.
by fijal 12y ago
PyPy does not really try to do that. CPython is great for a lot of things, small, easy to install on a myriad of platforms and is, after all, Python. PyPy instead tries to push what's possible to do with Python as a language and with d
256.
by fijal 12y ago
"In practice nodejs might be faster currently just because V8 is very good" and I presume you claim PyPy is not so good? Well if so, I would really like to say [citation needed], especially for workloads and not say computer langu
257.
by fijal 12y ago
I think your comment is moderately off-topic. This is really not about tornado at all, rather than you can take medium-sized program and get STM in PyPy not to crash
258.
Tornado Without a GIL on PyPy STM(morepypy.blogspot.com)
160 points by fijal 12y ago|65 comments
259.
by fijal 12y ago
If google recognizes you as an open source contributor who does something significant, you get couple hundred $$$ worth of credit in their store (I wish I were joking)
260.
by fijal 12y ago
cffi is mostly our answer. The cpyext layer (the layer that emulates CPython C API) is there too, but it's prone to be slow due to the need to emulate some other model of a VM. It's possible to do a better job, but just noone real
261.
by fijal 12y ago
undodb gdb (a commercial product) works a lot faster than gdb (which is so slow it's unusable). We used it with success to debug programs as big as pypy
262.
by fijal 12y ago
The same idea has been done for psyco (tracelets there are called basic-block-at-a-time compilation). PyPy uses a tracing JIT instead to achieve better performance in cases where psyco was too hard.
263.
by fijal 12y ago
Question to the authors: Any reason why hippyvm is not included for comparison in the paper? It does usually outperform hhvm on those benchmarks (but not on real world use cases which is maybe a good reason to include real world use cases m
264.
by fijal 12y ago
note that, confusingly enough, "tracelets" have nothing to do with "tracing JITs".
265.
by fijal 12y ago
PyPy people also happily encourage to just use python and not rewrite anything to C (or not rewrite most stuff). Those loops should be really really fast on PyPy btw, written in pure python (with numpy arrays)
266.
by fijal 12y ago
RPython is not a language for normal usage and mostly because e.g. error messages are obscure and hard to read. There is some basic support for generators these days, however, unless you're writing another language, we strongly encoura
267.
by fijal 12y ago
I would like to swap "people" in your sentence for "americans". It's a cultural thing, not a human thing.
268.
by fijal 12y ago
There is lsprofcalltree which will use normal cProfile and give you output suitable for kcachegrind (which is awesome)
269.
by fijal 12y ago
It's not impossible . It's "easily" done if you have a JIT, but even without a JIT, you can inline the call on the bytecode level using similar techniques - you have to be able to rebuild the chain if someone asks for i
270.
by fijal 12y ago
You can either use numpy ints in pypy (I think they're nicely optimized) or there is __pypy__.intop module that does something. Come to #pypy on freenode we'll sort you out ;-)

More ›