News for hackers. Stuff that matters.
Re:On the Complexity and Performance of Parsing with Derivat... by chubot on Saturday June 25, 2016 @06:04PM (#11978209)
Your comments betray a lot of ignorance, so I will just point out the glaring inaccuracies in #1. Here are two different expression parsers that are likely installed on machines you use, and they absolutely make a function call for each level of precedence (eval1, eval2, ..., exp1, exp2, ...)

http://code.metager.de/source/xref/gnu/coreutils/src/expr.c

https://opensource.apple.com/source/bash/bash-76.1/bash/expr...

My point is not that this is BAD -- it probably doesn't matter on modern machines. I'm just saying there are choices to make. Parsing requires some knowledge and choices -- that's all. You don't just "type out the code".

#3 is not a refutation, because all the languages I mentioned are extraordinarily common. The answer to "parsing real languages is hard" is not "don't parse real languages".

I can also tell by #4 that you don't have any background in security. It's not about YOUR code per se. It's about code that is deployed and that we all rely on.

Re:On the Complexity and Performance of Parsing with Derivat... by SamReidHughes on Saturday June 25, 2016 @06:40PM (#11978339)
1. Did I not say there might be some living strawmen out there?

3. Did I say we shouldn't parse real languages? No, I said the problem was that it's not straightforward to combine two languages into a new one. The problem's not at the level of parsers.

4. Did I not say it wasn't evenly distributed?