| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
| |
|
| |
|
|
|
|
|
|
|
| |
constructor, and instead create an _internal_only getter
that gets the needed information. This is a workaround for
a deficiency in gcc-4.4 that does not properly support
templated friend classes.
|
|
|
|
| |
As it may be implementation dependent.
|
|\
| |
| | |
Bridge vs2008 hashmaps.
|
| | |
|
|/
|
|
|
| |
This time move the macro inside namespaces; hopefully will work with the
integration script.
|
|\
| |
| | |
updated README
|
| | |
|
| | |
|
| |
| |
| | |
Added relative link to cmake guide.
|
|\ \
| |/
|/| |
Fix bug with silent message corruption in LITE_RUNTIME.
|
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
A protobuf message will be corrupted in the following scenario:
1. Use LITE_RUNTIME.
2. Have an optional enum field following some other field.
3. Update protocol by adding new values to the enum.
4. Have an old client parse and serialize a message having enum field
set to a value the client does not understand.
5. Field preceeding the enum is now corrupted.
The bug is due to the fact that optimized fallthrough in parser code
does not update variablle 'tag' when jumping to the parser code for the
next field.
|
|\ \ |
|
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
proto2 syntax allows the first enum to have a non zero value. This means any
field using that default has a non zero default without having an explicit
default being set. So when deciding what runtime info is needed, don't rely
on an explicit default, always check that the values aren't zero.
Fixes https://github.com/google/protobuf/issues/1453
|
| |\ \
| | | |
| | | | |
Change protobuf CPP proto generator to support the 'lite' option in
|
| | |/
| | |
| | |
| | |
| | |
| | |
| | |
| | |
| | | |
proto3.
Added a couple unit test proto3 files, for arena_lite and lite. Cloned
the proto3_arena_unittest to test some of the basics of generated code
(and to ensure that the generated proto3 test files are used by some
test).
|
|/ / |
|
|\ \
| | |
| | | |
Remove googletest.h header from stringprintf.cc
|
| | |
| | |
| | |
| | |
| | | |
It doesn't seem to be necessary here, and it pulls other testing headers
during compilation of release protobuf.
|
|\ \ \
| | | |
| | | | |
Enum casing in C#
|
| | | | |
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
Overview of changes:
- A new C#-specific command-line option, legacy_enum_values to revert to the old behavior
- When legacy_enum_values isn't specified, we strip the enum name as a prefix, and PascalCase the value name
- A new attribute within the C# code so that we can always tell the original in-proto name
Regenerating the C# code with legacy_enum_values leads to code which still compiles and works - but
there's more still to do.
|
| | | |
| | | |
| | | |
| | | |
| | | |
| | | | |
enum value name
This will make it easier to change the enum value names, as it reduces the number of places they're used.
|
| |_|/
|/| | |
|
|\ \ \
| |/ /
|/| | |
Internal merge
|
| | | |
|
| |\ \ |
|
| |\ \ \ |
|
| | | | |
| | | | |
| | | | |
| | | | | |
For fixing MSVC
|
| | | | | |
|
| | | | | |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
stubs/common.h undefines the GetMessage macro introduced in windows.h
map_test however include stubs/common.h before windows.h is transitively
included. This hack force map_test.cc to include windows.h first, so we
have a chance to undefine the GetMessage macro.
|
| | | | | |
|
| | | | | |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
Referencing data member table_ in static functions break certain
platforms. Change it to refer to the declare type void** instead.
|
| | | | | |
|
| |\ \ \ \
| | | |_|/
| | |/| | |
|
| | | | | |
|
| | | | | |
|
|\ \ \ \ \
| |_|_|_|/
|/| | | | |
Added support for internal_access for C#
|
| | | | | |
|
|\ \ \ \ \
| |/ / / /
|/| | | | |
Sync the manually integrate changes in google3/third_party
|
| | | | | |
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This also renames generate_directories to base_namespace_specified; generating directories is the
immediate *effect* of specifying a base namespace, but with this change the options reflect what has been
specified rather than the effect. (There may be other effects in the future, of course.)
|
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
This should have no behavioral changes at all.
This doesn't strictly enforce an 80-column limit, but removes the most egregious violations.
The indentation in the C# generator code is inconsistent in general, unfortunately - if we have
any good tools that can be trusted to reformat, I'd be happy to apply them.
|
|\ \ \ \ \
| | | | | |
| | | | | | |
Added CLI option internal_access for types (C#)
|
| | |_|_|/
| |/| | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | |
| | | | | |
* `csharp_options`: Added `Options` to encapsulate generator options.
Supported options for now - file_extension, base_namespace
* `{Blah}Generator`: Now accept `Options*` as parameter to constructor
* `csharp_generator.cc`: Parse and populate options
* `Makefile.am`: Added `csharp_options.h`
* `extract_includes.bat.in`: Added `csharp_options.h`
Refactoring code to two commits. This is the first commit
|
|/ / / / |
|