I have been working on the array module the past days. The current version already passes all compliance tests when run on top of CPython. But the integration with PyPy still poses some potentially hard problems. They’re mostly caused by the tight coupling of the array module to C data types. E.g., how can I […]
I have a reputation for overestimating the time it takes to accomplish a given task. This has led to interesting discussions with software project managers in the past: “You told me you’d finish this in 2 days, so I assumed it’d take you 4 days, and now you’re telling me you completed it in 1 […]
When I was talking about not walking in the park, I meant exactly the kind of thing that has been holding me up yesterday and today: A bug related to character sets, manifesting itself on one machine but not on the other, for seemingly random reasons.
Character sets are things like [a-cg-j] in a regular expression. […]
I’m making progress much faster than I thought. Blame Canada, er, blame the unspectacular weather this week, keeping my head cool and the urge to go out low. I am now down to just 6 (of totally 51) regex unit tests from CPython 2.4 failing. Of these failures, 4 seem to be unicode- or locale-related, […]
I’ve been working a bit on test infrastructure today. I can now
run my own tests on CPython using _sre.py,
run my own tests on CPython using _sre.c (crosschecking that my tests are actually correct),
run the CPython re tests on CPython using _sre.py,
run my own tests on PyPy using _sre.py,
and run the CPython re tests on PyPy […]
The Python regex implementation compiles regex patterns to an intermediate bytecode form. Since what I’m writing is basically the interpreter for this bytecode, I’ve spent quite some time trying to make sense of numeric bytecode representation produced by sre. It’s not that hard, but you really loose track very quickly, mentally parsing a sequence like […]
The university project I talked about was postponed until October, so I could already pick up SoC again on Friday. That day I implemented pretty much all regex syntax that doesn’t require interpreting a subexpression, among them most categories like \d (matches all digits) and character sets (e.g., [a-d] to match all letters from […]
This is part 2 of my informal sprint report. Here’s part 1.
On the third sprint day in Göteborg I finally kicked off my actual SoC project, rewriting the _sre and array modules in Python. I sat together with Armin to discuss some details. We agreed on the basic requirement that will be used to evaluate […]
I’m trying very hard to not work on my SoC project this week (blogging about it here instead, is just a substitute for the real thing). I know my working habits too well. Would I just casually start coding again, I would quickly get so immersed that everything else that needs to be done this […]
I’ve never heard anyone actually utter the term, but there was supposedly a talk about it at EuroPython: sprint-driven development. This is one of the areas where PyPy is truly innovative. The core developers and other interested parties, although strewn all over Europe, meet pretty regulary for a few days of joint, intense development. This […]