I've implemented a new HTTP/1.1 request and response parser by hand

http://four.livejournal.com/1033160.html

by fogus • 17 years ago

79 38 17 years ago

four.livejournal.com

viraptor5• 17 years ago

For a couple of reasons this code doesn't look like a very reliable one...

Almost identical cases don't reuse code (not even a define). There are also sections like "`if (usual[ch >> 5] & (1 << (ch & 0x1f))) break;`" without any comments. The code hardcodes the http methods for some reason so that the detection code spans ~130 lines (454..580). Looks like it will accept "HXXX/1.1" if strict checking is off. This double check is... interesting:

    if (!parser->FOR##_mark) return 0; \
    assert(parser->FOR##_mark); \

Sure - speed++, but at what cost? Otherwise... cool code - I like the MARK / CALLBACK macros.

jerf• 17 years ago

"I can make it arbitrarily fast if I don't actually have to make it work": http://blogs.msdn.com/larryosterman/archive/2009/09/29/i-can...

nostrademons3• 17 years ago

I'm disappointed that the student who hardcoded the results was disqualified. You should get points for finding bugs in your professors' specification.

viraptor5• 17 years ago

After reading that, I wonder if trie-s (i.e. judy arrays) would beat hash tables in this exercise... (and I also think that hardcoding the results wasn't cheating)

gruseom• 17 years ago

I've seen that line referenced a couple of times here recently. I'm pretty sure it derives from Jerry Weinberg's Psychology of Computer Programming, which contains a story about some programmers defending an incorrect program on the grounds that it would process more cards per minute, to which Weinberg replied, "But your code doesn't work. If mine doesn't have to work I can do a million cards a minute!"

One of my favorite software stories.

ryah2• 17 years ago

> Sure - speed++, but at what cost?

it only has to be written once. i think it's an important enough problem to warrant such code - definitely could use a few more macros though

snorkel• 17 years ago

Clever use of ## in preprocessor directives. http://en.wikipedia.org/wiki/C_preprocessor

rglullis• 17 years ago

I need to dig through my old computer and find my college lab assignments where we built a basic HTTP/1.1 server. For the request parser, we had to create a lex file for a grammar that include actions for all of the verbs, and also be robust to accept non-standard verbs, i.e, return a 400 code.

I doubt that my yacc'd program would be only 124 bytes in size, but it would be interesting to get that old code and compare the results.

brianobush• 17 years ago

At my company we rolled our own too. The hard part was tracking down all the brain-dead servers that seemingly worked fine with browsers, but were off spec.

jacobolus• 17 years ago

Are there any good sets of HTTP compatibility tests which test deviations from the spec which show up in existing servers/browsers?

wmf• 17 years ago

http://coad.measurement-factory.com/

Warning: not free.

xal• 17 years ago

It has also been merged into node. This thing is starting to be a case study for optimized C network servers.

pquerna• 17 years ago

I thought it was a pretty cool implementation (zero memory alloc is pretty hard most of the time), though I was disappointed how HTTP methods were hard coded in, so adding a new HTTP method (like PROPFIND?) would be adding a ton of code, though I think you could make a few macros for doing HTTP methods and keep the rest of it.

davisp• 17 years ago

Hacked together an on_method callback last night for the Python bindings I'm writing:

http://github.com/davisp/http-parser/commit/50e54f95fd4c2eac...

bumblebird4• 17 years ago

By hand??? As opposed to what exactly? Isn't most code written by hand?

ionfish• 17 years ago

As opposed to using Ragel, as you'd have known if you'd read the link.

bumblebird4• 17 years ago

The title is "I've implemented a new HTTP/1.1 request and response parser by hand"

My question was, how do you implement a new HTTP/1.1 request and response parser if it's not by hand. Isn't most code written using hands?

I've written one too... big whoop.

Meh anyway...

I've implemented a new HTTP/1.1 request and response parser by hand

Related Stories