summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2011-08-19Benchmark to parse into proto2 messages.Joshua Haberman
2011-08-17x86 JIT: add callback specializations for a 10% speedup when parsing to struct.Joshua Haberman
2011-08-17Benchmark JIT vs no JIT without forcing a whole-project recompile.Joshua Haberman
2011-08-12Python: fleshed out accessors.Joshua Haberman
2011-08-12Python: basic SymbolTable support and empty accessors.Joshua Haberman
2011-08-11Python: fixed object cache and fleshed out MessageDef a bit.Joshua Haberman
2011-08-10Fixed overzealous assert().Joshua Haberman
2011-08-10Fixed a few memory leaks and Makefile tweaks.Joshua Haberman
2011-08-10Fixed off-by-one stack allocation error.Joshua Haberman
2011-08-10Python: Implemented weak-ref'd object cache.Joshua Haberman
Also some preliminary work on upb.MessageDef.
2011-08-09Makefile target for running Python tests.Joshua Haberman
2011-08-09More Python upb.FieldDef functionality and tests.Joshua Haberman
2011-08-08Begin port of Python extension to new APIs.Joshua Haberman
2011-08-04More work on Lua extension, and consequent core refactoring.Joshua Haberman
2011-07-31More work on Lua extension.Joshua Haberman
2011-07-31Fixed "make descriptorgen".Joshua Haberman
2011-07-29Significant work on Lua extension.Joshua Haberman
Also changes in core library to accommodate.
2011-07-19Added comment about how to disassemble JIT output.Joshua Haberman
2011-07-16Get rid of upb_symtabtxn.Joshua Haberman
This type was nothing but a map of defs. We can as easily just pass an array of defs into upb_symtab_add().
2011-07-15Make Lua extension build again.Joshua Haberman
2011-07-15Directory restructure.Joshua Haberman
Includes are now via upb/foo.h. Files specific to the protobuf format are now in upb/pb (the core library is concerned with message definitions, handlers, and byte streams, but knows nothing about any particular serializationf format).
2011-07-14Major refactoring: upb_string is gone in favor of upb_strref.Joshua Haberman
2011-06-17Major refactoring: abandon upb_msg, add upb_accessors.Joshua Haberman
Next on the chopping block is upb_string.
2011-05-21Make all handlers objects refcounted.Joshua Haberman
I'm realizing that basically all upb objects will need to be refcounted to be sharable across languages, but *not* messages which are on their way out so we can get out of the business of data representations. Things which must be refcounted: - encoders, decoders - handlers objects - defs
2011-05-21Add proof-of-concept C++ wrapper header.Joshua Haberman
2011-05-20Add startseq/endseq handlers.Joshua Haberman
Startseq/endseq handlers are called at the beginning and end of a sequence of repeated values. Protobuf does not really have direct support for this (repeated primitive fields do not delimit "begin" and "end" of the sequence) but we can infer them from the bytestream. The benefit of supporting them explicitly is that they get their own stack frame and closure, so we can avoid having to find the array's address over and over and deciding if we need to initialize it. This will also pave the way for better support of JSON, which does have explicit "startseq/endseq" markers: [].
2011-05-20JIT: factor out push/pop frame into macros.Josh Haberman
2011-05-19Change dispatcher error handling model.Joshua Haberman
Now the dispatcher will call error handlers instaed of returning statuses that the caller has to constantly check.
2011-05-10Fix JIT for new interface.Joshua Haberman
2011-05-10Fix aliasing violation.Joshua Haberman
2011-05-10Vastly improved/simplified the upb_handlers API.Joshua Haberman
2011-05-08More renaming.Joshua Haberman
2011-05-08Documentation, some type renaming, nix unknown handler for now.Joshua Haberman
2011-05-08Split upb_stream -> upb_bytestream/upb_handlers.Joshua Haberman
2011-05-08Decoder redesign in preparation for packed fields and start/endseq.Joshua Haberman
2011-05-06Refactor varint encoding/decoding.Joshua Haberman
2011-04-25Revise/clarify comment about clear() implementation.Joshua Haberman
2011-04-15Calculate and print string sizes in test messages.Joshua Haberman
2011-04-15Merge branch 'master' of git@github.com:haberman/upbJoshua Haberman
2011-04-14Fix GDB JIT registration on Linux.Joshua Haberman
2011-04-08Small change to make Lua extension compile again.Joshua Haberman
2011-04-06JIT now works on OS X!Joshua Haberman
Had to fix the case where the jump table was allocated in a >4GB address.
2011-04-06Disable GDB JIT interface on OS X.Josh Haberman
OS X doesn't use ELF, so our little trick doesn't work there.
2011-04-05Update README and fix small decoder issue.Joshua Haberman
2011-04-04Speed up parsetostruct by using type-specialized callbacks.Joshua Haberman
2011-04-02JIT passes all tests!Joshua Haberman
2011-04-01Whoops, add src/jit_debug_elf_file.sJoshua Haberman
2011-04-01First rough version of the JIT.Joshua Haberman
It can successfully parse SpeedMessage1. Preliminary results: 750MB/s on Core2 2.4GHz. This number is 2.5x proto2. This isn't apples-to-apples, because proto2 is parsing to a struct and we are just doing stream parsing, but for apps that are currently using proto2, this is the improvement they would see if they could move to stream-based processing. Unfortunately perf-regression-test.py is broken, and I'm not 100% sure why. It would be nice to fix it first (to ensure that there are no performance regressions for the table-based decoder) but I'm really impatient to get the JIT checked in.
2011-03-28Switch to non-branching varint decoder.Joshua Haberman
2011-03-26Small semantics changes in the decoder.Josh Haberman
Simplified some of the semantics around the decoder's data structures, in anticipation of sharing them between the regular C decoder and a JIT-ted decoder.
generated by cgit on debian on lair
contact matthew@masot.net with questions or feedback