c619a3da2a
Fix syntax of function definition in shell script so that it hopefully works on Solaris.
15 years ago
d2fcbba230
Improve zip/jar support:
...
- Don't support par or war since par is not a standard format and outputting to
war doesn't make sense.
- Add boilerplate manifest when outputting to Jar.
- Remove "XXX check conversion" comments because I don't care.
- Add a test.
15 years ago
15604083c7
Move string representations back upb.h -> upb_data.h.
15 years ago
fa5710f1ca
upb_string* -> upb_strptr, to follow aliasing rules.
15 years ago
ece08710a6
Bugfixes: descriptorgen works without leaks!
15 years ago
2fdc9df97e
Other than a couple memory leaks, "make descriptorgen" works again!
15 years ago
c33d51581a
upbc compiles and links! But probably doesn't work yet.
15 years ago
f6cc21bc2d
Doesn't compile, but all the code to be adapted is there.
15 years ago
ab0f2b8bba
Moved string library functions to .c instead of inlined, because they're not that small.
15 years ago
3f0d126b2a
Only remaining undefined symbol is upb_msg_parsestr.
15 years ago
cc39625767
Getting closer, only a few functions undefined now.
15 years ago
afe911ea00
Removed upb_mm.
15 years ago
8978661a76
upbc compiles, but doesn't link due to unimplemented stuff in upb_data.
15 years ago
e5debfa1c9
More incremental work; ported some of upbc.
15 years ago
5066322074
Allow protoc to output directly to zip/jar/par/war files.
15 years ago
602e093504
Core compiles, though with warnings, and is still incomplete.
15 years ago
5f12164f54
Refactor the way output is handled in CommandLineInterface -- now it will be stored in-memory until all code generators have completed, then dumped to disk all at once. While this means that protoc uses more memory, the code is much simpler, and handles insertions much faster. Also, this made it easier to implement a useful feature: insertions will be indented to match the insertion point line. Therefore, when inserting into Python code, you don't have to figure out how much to indent your inserted code. The refactoring should also make it easier to implement output-to-jar at some point.
15 years ago
46ed74e8d4
Actually, that last revision can be simpler -- we don't need to parse strings at all, as simply entering 1e1000 as a float literal in Python will be evaluated as infinity.
15 years ago
d0047c43d9
In Python, avoid relying on float('inf') and float('nan') as these don't work on Windows with Python pre-2.6.
15 years ago
f49f7f94c1
More progress, upb_data and upb_def compile (but are incomplete).
15 years ago
6e8b9e4a4a
Make extension identifiers final. This improves security when untrusted code is present in the same class loader. In order to get around initialization ordering issues, I simply made the constructor for extension identifiers take no arguments and deferred initialization to an internalInit() method, which generated code will always call during init.
15 years ago
7f4938b244
Mark plugins as experimental for now.
15 years ago
d6e84b3ed8
When serializing to an ostream, verify that there were no stream-level errors before returning success.
15 years ago
6f12e3e29a
Accept form feeds as whitespace.
15 years ago
f9c5978f5d
Hopefully fix build on Sun Studio.
15 years ago
a230cf5053
More work.
15 years ago
c0ee4d2ed9
Build DLLs on Cygwin and MinGW.
15 years ago
d1ce7a2839
Fix build on MSVC.
15 years ago
7dd113baa8
Added upb_data.*, left out of last commit.
15 years ago
f2a732938e
Fix build on GCC 3.3/Linux.
15 years ago
0abe9c2143
Argh, this line was not supposed to be submitted.
15 years ago
ee7e9420e3
Use unordered_map when available. Changes to stl_hash.m4 provided by Oleg Smolsky.
15 years ago
c241976485
In the midst of a major refactoring.
15 years ago
8ee1474044
Add a code generator insertion point for Java builders.
15 years ago
684d45b2fe
Fix build on MinGW/Win32 (including implementing Subprocess using CreateProcess()).
15 years ago
bdbb863099
Ensure that 'once' objects are declared using the macro. This is either the third or fourth time I've screwed this up when down-integrating, because our internal code does not require the macro (it's not portable) and on Linux a pthread_once_t that is zero-initialized just happens to work. So, I only discover the problem when I test on Mac, then kick myself for making the same mistake yet again. No more! This time, I have renamed GoogleOnceType to ProtobufOnceType, thus making the type name differ from our internal code. As a result, if you don't update the decls to use the macros, they won't compile. Hah! Take that, future self!
15 years ago
91218afc67
Fix Cygwin build.
15 years ago
5e744ff961
Commit missing files from r251. Sigh.
15 years ago
afd32abb75
Set version to 2.3.0-pre.
15 years ago
83ad01e35f
Point test_plugin at the correct gtest headers.
15 years ago
fccb146e3f
Massive roll-up of changes. See CHANGES.txt.
15 years ago
b221008884
Fix UTF-8 validity checks to not do unaligned reads.
15 years ago
de74779454
fix SONAME in trunk
15 years ago
c6cba2af37
Clarified comment.
15 years ago
dc4246db78
Refined interface of upb_symtab.
15 years ago
0a6fc5fad3
Truly fixed type cyclic refcounting.
15 years ago
e15f834a91
Circular references truly work now, along with a test.
...
One simplification to come.
15 years ago
08b4a91204
Add a test for circularly-linked descriptors.
...
The test currently triggers valgrind-detected memory errors.
15 years ago
651c92ab33
Scheme for collecting circular refs.
...
"make descriptorgen" is now valgrind-clean again.
15 years ago
18291eedc3
Make defs refcounted, rename upb_context->upbsymtab.
...
There is currently a memory leak when type definitions
form cycles. This will need to be dealt with.
15 years ago
kenton@google.com
Joshua Haberman
temporal