Permalink
Commits on Jan 19, 2011
  1. improve execute2 doc

    ry committed Jan 19, 2011
Commits on Dec 6, 2010
  1. event stream

    ry committed Aug 16, 2010
Commits on Nov 22, 2010
  1. Add support for "M-SEARCH" and "NOTIFY" request methods.

    TooTallNate authored and ry committed Oct 12, 2010
    Allow a request path of "*" (for SSDP requests).
Commits on Nov 11, 2010
  1. Set http_major when a request omits the HTTP version

    TooTallNate authored and ry committed Oct 18, 2010
    I.E. "GET /" in telnet
  2. Fix build issues using mingw32 on windows

    tmm1 authored and ry committed Oct 27, 2010
  3. Remove acceptable_header array

    Cliff Frey authored and ry committed Nov 10, 2010
    This was not necessary, as it was just being used as a downcase
    function.
Commits on Nov 5, 2010
  1. Add non-ascii in status line test

    ry committed Nov 5, 2010
    from Ben Noordhuis
Commits on Oct 20, 2010
  1. readme typo (thanks tmm1)

    ry committed Oct 20, 2010
Commits on Oct 12, 2010
Commits on Sep 12, 2010
  1. avoid assertion failure in error case

    Cliff Frey authored and ry committed Sep 11, 2010
    Without this change, it is possible to get an assertion to fail by
    continuing to call http_parser_execute after it has returned an error.
    Specifically, the parser could be called with parser->state ==
    s_chunk_size_almost_done and parser->flags & F_CHUNKED set.  Then,
    F_CHUNKED could have been cleared, and an error could be hit.  In this
    case, the parser would have returned with F_CHUNKED clear, but
    parser->state == s_chunk_size_almost_done, resulting in an assertion
    failure on the next call.
    
    There are alternate solutions possible, including just saving all of
    the fields (state included) on error.
    
    I didn't add a test case because this is a bit annoying to test, but I
    can add one if necesssary.
Commits on Aug 27, 2010
Commits on Aug 18, 2010
  1. save 8 bytes in http_parser structure

    Cliff Frey authored and ry committed Aug 18, 2010
    nread can never be larger than HTTP_MAX_HEADER_SIZE
  2. Array type cleanups. Also save space

    Cliff Frey authored and ry committed Aug 18, 2010
    acceptable_header[x] is always assigned to a variable of type char, so
    the 'unsigned' is unnecessary.
    
    The other arrays can be of type int8_t/uint8_t to save space.
  3. fixes for architectures with signed char default

    Cliff Frey authored and ry committed Aug 18, 2010
    This could have resulted in memory before the normal_url_char array
    being read on architectures with signed char default.
Commits on Jul 31, 2010
  1. Changes to compile with clang

    ry committed Jul 31, 2010
  2. Use lookup tables of my own.

    ry committed Jul 31, 2010
Commits on Jul 30, 2010
  1. Add CONTRIBUTIONS file

    ry committed Jul 30, 2010
  2. Update size of struct in README

    ry committed Jul 30, 2010
Commits on Jul 26, 2010
  1. typo

    ry committed Jul 26, 2010
  2. Support long messages

    ry committed Jul 26, 2010
Commits on Jul 24, 2010
  1. Allow spaces in header fields

    ry committed Jul 24, 2010
  2. Fix long line

    ry committed Jul 24, 2010
Commits on Jul 19, 2010
  1. redundant upgrade flag check

    syngenio authored and ry committed Jul 18, 2010
Commits on Jul 17, 2010
  1. Upgrade on CONNECT method

    sgala authored and ry committed Jul 17, 2010
Commits on Jul 6, 2010
  1. test: always try and break every testcase up into two submessages

    Cliff Frey authored and ry committed Jul 6, 2010
    This is just another way that would have caught the bug introduced in
    076fa15 and fixed by
    03b8eaa.
  2. fix http_parser_init to initialize flags correctly

    Cliff Frey authored and ry committed Jul 6, 2010
    Yay valgrind testing
    
    I don't believe that this actually mattered at all, because state was
    initialized correctly, and flags would be set to 0 almost immediately
    anyways.
  3. test.c: get it to work with valgrind by using realloc less

    Cliff Frey authored and ry committed Jul 6, 2010
    For some reason valgrind would rapidly run out of memory on my machine
    without this.