summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2015-12-30Add missing upb_fielddef_fullname implementationMattia Barbon
2015-10-21Merge pull request #43 from haberman/presenceflagJoshua Haberman
Add flag to MessageDef for whether fields have presence.
2015-10-21Addressed PR comments.Josh Haberman
2015-10-15Properly populate field presence flag on messages according to syntax flag.Josh Haberman
2015-10-14Add flag to MessageDef for whether fields have presence.Josh Haberman
2015-08-19Merge pull request #38 from haberman/decoderfix2Joshua Haberman
Added lots of decoder tests and fixed lots of bugs.
2015-08-19Refer to a shared definition for max buffered bytes.Josh Haberman
2015-08-14Added lots of decoder tests and fixed lots of bugs.Josh Haberman
2015-08-13Merge pull request #37 from haberman/google-internalJoshua Haberman
Fix from google3: memory leak caught by heapcheck.
2015-08-13Fix from google3: memory leak caught by heapcheck.Josh Haberman
2015-08-12Merge pull request #36 from haberman/decoderfixJoshua Haberman
Decoder fix: skipped data at end of submessage.
2015-08-11More decoder fixes, and slightly changed parse call semantics.Josh Haberman
Prior to this change, if an error was returned, it would be guaranteed to always return a short byte count. Now the two concepts are a bit more orthogonal. There are cases where the entire input is consumed even though an error was encountered.
2015-07-31Decoder fix: skipped data at end of submessage.Josh Haberman
2015-07-31Merge pull request #34 from haberman/parserskipJoshua Haberman
Changed parser semantics around skipping.
2015-07-30Fixed PR for JIT-enabled builds.Josh Haberman
2015-07-30Changed parser semantics around skipping.Josh Haberman
Prior to this change: parse(buf, len) -> len + N ...would indicate that the next N bytes of the input are not needed, *and* would advance the decoding position by this much. After this change: parse(buf, len) -> len + N parse(NULL, N) -> N ...can be used to achieve the same thing. But skipping the N bytes is not explicitly performed by the user. A user that doesn't want/need to skip can just say: parsed = parse(buf, len); if (parsed < len) { // Handle suspend, advance stream by "parsed". } else { // Stream was advanced by "len" (even if parsed > len). } Updated unit tests to test this new behavior, and refactored test utility code a bit to support it.
2015-07-15Merge pull request #33 from haberman/jsonerrJoshua Haberman
Fixed JSON parser error reporting.
2015-07-15Clear JSON parser status in reset.Josh Haberman
2015-07-15Updated JSON parser size.Josh Haberman
2015-07-15Fixed JSON parser error reporting.Josh Haberman
This was previously broken -- it would try to set the status object on the parser, but the pointer was never initialized. Also it didn't report errors properly to the environment object.
2015-07-07Also remove license headers from Lua.Josh Haberman
2015-07-07Updated license headers for Ragel source also.Josh Haberman
2015-07-07Brought into compliance with Google open-source policies.Josh Haberman
- removed myself from Author headers in source files. - removed copyright notices from source file headers. - added CONTRIBUTING.md
2015-07-07Merge pull request #32 from haberman/third_partyJoshua Haberman
Moved lunit to third_party for Google compliance.
2015-07-07Moved lunit to third_party for Google compliance.Josh Haberman
2015-07-07Merge pull request #31 from haberman/third_partyJoshua Haberman
Moved DynASM to third_party to comply with Google policy.
2015-07-07Moved DynASM to third_party to comply with Google policy.Josh Haberman
2015-06-22Merge pull request #30 from haberman/encoderfixJoshua Haberman
Fixed some bad bugs in upb_env.
2015-06-22Fixed some bad bugs in upb_env.Josh Haberman
Also added a unit test for upb_encoder that demonstrates the bugs and the fix.
2015-06-15Merge pull request #29 from haberman/symtabfixJoshua Haberman
Fix for stack overflow for cyclic defs.
2015-06-15Addressed code review comment and clarified comments.Josh Haberman
2015-06-12Fix for stack overflow for cyclic defs.Josh Haberman
Fixes this bug that came up in the Ruby extension: https://github.com/google/protobuf/issues/425
2015-06-12Fix amalgamation by changing include style for structdef.int.h.Josh Haberman
2015-06-06Parameterize AR properly in the Makefile.Josh Haberman
2015-06-06Fix for va_copy.Josh Haberman
2015-06-04Fixes from Google-internal.Josh Haberman
2015-06-03Merge branch 'master' into google-internalJosh Haberman
2015-06-03Sync from Google-internal development.Josh Haberman
2015-06-03Merge pull request #27 from haberman/c89Joshua Haberman
Ported upb to C89, for greater portability.
2015-06-02Ported upb to C89, for greater portability.Josh Haberman
A large part of this change contains surface-level porting, like moving variable declarations to the top of the block. However there are a few more substantial things too: - moved internal-only struct definitions to a separate file (structdefs.int.h), for greater encapsulation and ABI compatibility. - removed the UPB_UPCAST macro, since it requires access to the internal-only struct definitions. Replaced uses with calls to inline, type-safe casting functions. - removed the UPB_DEFINE_CLASS/UPB_DEFINE_STRUCT macros. Class and struct definitions are now more explicit -- you get to see the actual class/struct keywords in the source. The casting convenience functions have been moved into UPB_DECLARE_DERIVED_TYPE() and UPB_DECLARE_DERIVED_TYPE2(). - the new way that we duplicate base methods in derived types is also more convenient and requires less duplication. It is also less greppable, but hopefully that is not too big a problem. Compiler flags (-std=c89 -pedantic) should help to rigorously enforce that the code is free of C99-isms. A few functions are not available in C89 (strtoll). There are temporary, hacky solutions in place.
2015-05-28Update Travis and Coveralls URLs for repo switch.Josh Haberman
2015-05-28Merge pull request #21 from google/tablestringsJoshua Haberman
Restructure tables for C89 port and smaller size.
2015-05-19Merge pull request #22 from martinmaly/upbcJoshua Haberman
Generating enums and intentries only if the tables are non-empty.
2015-05-18Generating enums and intentries only if the tables are non-empty.Martin Maly
2015-05-18Updates from code review comments.Josh Haberman
2015-05-17Restructure tables for C89 port and smaller size.Josh Haberman
Changes the data layout of tables slightly so that string keys are prefixed with their size, rather than the size being inline in the table itself. This has a few benefits: 1. inttables shrink a bit, because there is no longer a wasted and unused size field sitting in them. 2. This avoids the need to have a union in the table. This is important for an impending C89 port of upb, since C89 has literally no way of statically initializing a non-first union member.
2015-05-16JIT bugfix: align stack properly on getbytes_slow.Josh Haberman
2015-05-16Upload failing artifacts to S3 for debugging purposes.Josh Haberman
2015-05-16Enabled asserts() and verbosity for most Travis builds.Josh Haberman
Also added a separate ndebug build for testing that -DNDEBUG builds still work. Also disabled reference debugging by default, since it requires either a global lock or -DUPB_THREAD_UNSAFE.
2015-05-15Merge pull request #19 from cfallin/upb-seededalloc-fixJoshua Haberman
Bugfix: seeded_alloc() should not realloc() user-provided memory.
generated by cgit on debian on lair
contact matthew@masot.net with questions or feedback