Commit message (Collapse) | Author | Age | |
---|---|---|---|
* | Always ref writable streams | 2016-03-02 | |
| | | | | | | | | We suffered a bug whereby doing a follow-up write to another write could resurrect a deleted stream, causing all sorts of crash. Fix: when a stream becomes writable (vs when we start writing) take a ref on the stream, and only relinquish it once we're done writing. | ||
* | Added back trailing newlines | 2016-01-14 | |
| | |||
* | Update copyrights | 2016-01-13 | |
| | |||
* | Fixed a couple of remaining merge issues | 2016-01-08 | |
| | |||
* | Merged from master and resolved merge conflicts | 2016-01-08 | |
|\ | |||
* | | Expose core metadata validation functions in public headers | 2016-01-06 | |
| | | |||
| * | Eliminate gpr_ int types - and insist on C99 variants instead | 2015-12-22 | |
| | | |||
| * | Allow . in header names | 2015-12-15 | |
|/ | |||
* | Re-legalize comma | 2015-12-09 | |
| | |||
* | Merge pull request #4175 from yang-g/tcp_refactor | 2015-12-03 | |
|\ | | | | | Refactor security connector and handshake | ||
* | | Fix HPACK encoder test | 2015-12-02 | |
| | | |||
| * | merge with head and resolve conflict | 2015-12-02 | |
| |\ | |/ |/| | |||
* | | Make metadata leaks a warning, and not a fatal error | 2015-11-23 | |
| | | | | | | | | There's still a bug on Windows that will take some time to track down | ||
* | | Fixes for profiling | 2015-11-21 | |
| | | |||
* | | Remove metadata context entirely | 2015-11-20 | |
| | | |||
* | | Remove metadata context from the channel stack | 2015-11-19 | |
| | | |||
* | | Add commentary | 2015-11-19 | |
| | | |||
* | | clang-format | 2015-11-19 | |
| | | |||
* | | Use static metadata table in metadata.c | 2015-11-19 | |
|/ | |||
* | stream_op cleanup: transport changes | 2015-11-02 | |
| | |||
* | Fix race conditions | 2015-10-10 | |
| | |||
* | Make getting metadata user data a lock free operation | 2015-10-09 | |
| | |||
* | Make metadata unref atomic | 2015-10-09 | |
| | | | | | | We used to need to lock the metadata context to unref an mdelem. This change makes it possible to lock only when the mdelem refcount would reach zero. | ||
* | Remove metadata canonicalization, and fix some bugs | 2015-09-25 | |
| | |||
* | Enable -Wconversion | 2015-09-10 | |
| | |||
* | Core compiles with -Wsign-conversion | 2015-09-10 | |
| | |||
* | Outlaw comma in non-binary metadata values | 2015-08-27 | |
| | |||
* | Fix character classes to updated spec | 2015-08-21 | |
| | |||
* | Outlaw illegal metadata characters | 2015-08-21 | |
| | |||
* | clang-format all source | 2015-08-18 | |
| | |||
* | Properly send GRPC_STATUS_UNAUTHENTICATED from server auth failures | 2015-07-31 | |
| | |||
* | Addressing review comments | 2015-07-22 | |
| | |||
* | Canonicalize metadata keys in core | 2015-07-22 | |
| | |||
* | Merge pull request #2289 from ctiller/footprints-on-the-sands-of-time | 2015-07-10 | |
|\ | | | | | Introduce multiple clocks to GPR | ||
* | | Fix TSAN reported race | 2015-07-10 | |
| | | |||
| * | Merge github.com:grpc/grpc into footprints-on-the-sands-of-time | 2015-07-09 | |
| |\ | |/ |/| | |||
* | | Add metadata refcount debugging | 2015-07-06 | |
| | | |||
| * | Introduce multiple clocks to GPR | 2015-07-01 | |
|/ | |||
* | Remove the ability to specify a string on cancel: it was broken and unused; ↵ | 2015-06-30 | |
| | | | | will restore later | ||
* | Fix some asan issues | 2015-06-08 | |
| | |||
* | Validate that headers contain legal bytes | 2015-05-06 | |
| | |||
* | Fix remaining allocation bugs | 2015-04-17 | |
| | |||
* | Merge with other metadata changes | 2015-04-14 | |
| | |||
* | Merge github.com:grpc/grpc into fast-unref | 2015-04-14 | |
|\ | |||
* | | Batch unref metadata in HTTP/2 stream encoder | 2015-04-14 | |
| | | | | | | | | | | Moves us from one metadata lock per metadata element to one metadata lock per HTTP/2 frame output. | ||
| * | Fix Typo | 2015-04-13 | |
| | | | | | | | | | | ... and this, kids, is why you should always compile in debug before pushing. | ||
| * | Change barriers to protect the innocent | 2015-04-13 | |
| | | |||
| * | Eliminate channel-wide lock for grpc_mdelem_ref. | 2015-04-10 | |
|/ | | | | We only need to lock on the initial ref from garbage to atomically change mdtab_free. | ||
* | Eliminate a redundant unlock-lock pair since this code is only invoked from | 2015-04-02 | |
| | | | | one place | ||
* | Fix use-after-free. | 2015-02-18 | |
| | | | | | | | Transport and channel have different lifetimes, but share a metadata context. Make the metadata context ref counted, and have transport take a ref. |