Tatsuhiro Tsujikawa
5d788e9753
nghttp2_hd: Move pointer variables before size_t
2014-01-12 19:28:57 +09:00
Tatsuhiro Tsujikawa
a85a11c1d9
Update priority for the stream to get response only
2014-01-09 23:30:45 +09:00
Tatsuhiro Tsujikawa
5aa487c5ba
nghttp2_stream: flags is now nghttp2_stream_flags
...
We use this flag to know whether it is pushed or not
2014-01-09 23:01:02 +09:00
Tatsuhiro Tsujikawa
45837a2cfa
Ensure PRIORITY frame reception/transmission rule
2014-01-09 22:06:38 +09:00
Tatsuhiro Tsujikawa
95b0b6cb79
Update doc
2014-01-09 21:42:44 +09:00
Tatsuhiro Tsujikawa
fe1a39ed10
Allow client side to send PUSH_PROMISE
2014-01-09 21:39:29 +09:00
Tatsuhiro Tsujikawa
dacfecf41a
lib, src, hdtest: Ensure that config.h is included
2014-01-08 23:30:02 +09:00
Tatsuhiro Tsujikawa
63e7b21b35
nghttp2_hd: Fix lower bound search in static table
2014-01-08 21:32:39 +09:00
Tatsuhiro Tsujikawa
0fb4d32d6a
nghttp2_hd: Binary search static table
2014-01-07 23:41:09 +09:00
Tatsuhiro Tsujikawa
468d0fccd3
nghttp2_hd: Use hash to speed up name/value pair search
2014-01-07 22:43:34 +09:00
Tatsuhiro Tsujikawa
d703197a50
nghttp2_hd: Mitigate timing attack
2014-01-07 21:51:48 +09:00
Tatsuhiro Tsujikawa
9eb031ce83
nghttp2_hd: Avoid nghttp2_hd_huff_decode_count
...
Huffman decoding is costly. It is faster to do geometric realloc
than calling nghttp2_hd_huff_decode_count to know the length in
advance.
2014-01-07 21:44:56 +09:00
Tatsuhiro Tsujikawa
4ac1652c70
nghttp2_hd: Remove invalid nghttp2_downcase call
2014-01-07 21:18:45 +09:00
Tatsuhiro Tsujikawa
295a01ef91
Reformat valid character table for header field name/value pair
2014-01-04 11:19:07 +09:00
Tatsuhiro Tsujikawa
888792a949
nghttp2_hd: Allow HTAB and obs-text characters for field-content
2014-01-03 22:46:40 +09:00
Tatsuhiro Tsujikawa
74d82aac56
Update doc
2014-01-02 01:00:11 +09:00
Tatsuhiro Tsujikawa
9cb8754d09
Rename nghttp2_session_fail_session as nghttp2_session_terminate_session
2013-12-26 00:23:07 +09:00
Tatsuhiro Tsujikawa
dfcdea894b
Move header name/value pair validation to on_*_received functions
2013-12-25 23:38:55 +09:00
Tatsuhiro Tsujikawa
779366640a
Update doc
2013-12-24 19:27:03 +09:00
Tatsuhiro Tsujikawa
794633f894
Add nghttp2_session_set_stream_user_data API function
2013-12-23 17:40:34 +09:00
Tatsuhiro Tsujikawa
2c1739a24e
Add comment to category in HEADERS and reorder
2013-12-22 19:23:37 +09:00
Tatsuhiro Tsujikawa
7c4dbb6ffc
nghttp2_hd: Don't malloc if huffman encoded string has 0 length
2013-12-19 23:19:14 +09:00
Tatsuhiro Tsujikawa
7a9eca1f7d
nghttp2_hd: Simplify huffman encode
2013-12-19 23:02:48 +09:00
Tatsuhiro Tsujikawa
5ca0cfa018
nghttp2_hd: Search hd table once
2013-12-18 21:45:15 +09:00
Tatsuhiro Tsujikawa
12144acf4d
nghttp2_hd: Check last byte of huffman encoded byte string strictly
2013-12-18 21:27:16 +09:00
Tatsuhiro Tsujikawa
01266182e9
nghttp2_hd: Match deflate_hd_tablelen to hd_table.len on decoder side
2013-12-17 21:39:21 +09:00
Tatsuhiro Tsujikawa
595493a0ec
Add debug output in nghttp2_hd
2013-12-16 23:54:20 +09:00
Tatsuhiro Tsujikawa
650f63c728
Fail session on the reception of DATA in closed stream or illegal state
2013-12-15 16:23:01 +09:00
Tatsuhiro Tsujikawa
65e54ca7b0
Add nghttp2_free()
2013-12-14 23:49:16 +09:00
Tatsuhiro Tsujikawa
bc7473182c
Update doc
2013-12-14 19:18:14 +09:00
Tatsuhiro Tsujikawa
0867b57c39
Update doc
2013-12-14 18:59:09 +09:00
Tatsuhiro Tsujikawa
e61876ccbd
nghttp2_hd: Provide dedicated function to enable no reference set feature
2013-12-14 18:48:41 +09:00
Tatsuhiro Tsujikawa
41104f7b63
Abandon DATA frame priority adjustment (again)
...
We tried several times about this subject, but for the current
HTTP/2.0 priority scheme, we think it is best to serve the highest
priroty streams first (interleaving streams if there are several
higest ones). There are an issue when aggregating several frontend
connections to one connection in backend, but it is HTTP/2.0
spec issue, rather than implementation.
2013-12-14 17:02:59 +09:00
Tatsuhiro Tsujikawa
364501a2cd
Get rid of magic value NGHTTP2_PRI_DECAY
...
For now we just double the priority value on each DATA frame
transmission. If priority is 0, it becomes to 1 and goes all
the way to (1 << 30) - 1.
2013-12-13 22:26:09 +09:00
Tatsuhiro Tsujikawa
bcee1b27a7
Update doc
2013-12-12 00:30:15 +09:00
Tatsuhiro Tsujikawa
e17f888944
Lower priority by constant value
2013-12-11 23:20:59 +09:00
Tatsuhiro Tsujikawa
9d49f9a356
Add copyright notice
2013-12-09 23:25:59 +09:00
Tatsuhiro Tsujikawa
6c77cec270
Remove nghttp2_submit_* API functions which has char **nv parameter
...
The nghttp2_submit_{request,response}2 functions are renamed as
nghttp2_submit_{request, response}.
2013-12-08 21:39:43 +09:00
Tatsuhiro Tsujikawa
d6212a6055
Fix doc
2013-12-08 17:35:14 +09:00
Tatsuhiro Tsujikawa
d773b42a2d
Ensure that pri goes to the lowest value
2013-12-07 16:14:39 +09:00
Tatsuhiro Tsujikawa
6ea91e57e0
Adjust struct/class alignment
2013-12-06 23:17:38 +09:00
Tatsuhiro Tsujikawa
84d1fffb3f
Announce HTTP-draft-09/2.0
2013-12-05 23:23:39 +09:00
Tatsuhiro Tsujikawa
1dea4e154b
Reintroduce priority adjustment for DATA frame
...
This mechanism existed but was deleted. We bring it back in order to
prevent lower priority streams from starving.
2013-12-05 23:12:18 +09:00
Tatsuhiro Tsujikawa
3fde4c7669
nghttp2_map: Code cleanup
2013-12-05 19:26:16 +09:00
Tatsuhiro Tsujikawa
a3c888d7d1
nghttp2_map: Implement hash table
2013-12-05 00:53:03 +09:00
Tatsuhiro Tsujikawa
d0d0009a50
Use largest valid stream ID which passed to callback as last-stream-ID
...
Previously we use largest stream ID received so far as last-stream-ID,
and it is irrevant that it is passed to the callback (thus upper layer).
Now the stream ID which is passed to callback is eligible to
last-stream-ID.
2013-11-28 23:26:34 +09:00
Tatsuhiro Tsujikawa
2a83fc7559
Add nghttp2_submit_response2
2013-11-28 21:35:48 +09:00
Tatsuhiro Tsujikawa
e57b3a2139
Set protocol ID to draft-08 temporarily
2013-11-21 21:58:45 +09:00
Tatsuhiro Tsujikawa
bc21edf5b6
Interleave streams with the same priority
2013-11-19 23:59:10 +09:00
Tatsuhiro Tsujikawa
e5e4fc9d14
nghttp2_hd: Define some headers not for indexing based on our research
2013-11-19 21:41:02 +09:00