This is a more logical name in parallel with PB_HTYPE_REQUIRED and PB_HTYPE_OPTIONAL.
Warning: This breaks backwards-compatibility of generated .pb.c files.
You will have to regenerate the files and recompile.
This allows slight optimizations if only memory buffer support
(as opposed to stream callbacks) is wanted. On ARM difference
is -12% execution time, -4% code size when enabled.
In the pb_istream_from_buffer and pb_ostream_from_buffer, memcpy was
used to transfer values to the buffer. For the common case of
count = 1-10 bytes, a simple loop is faster.
Main code is now compiled (for tests) with -pedantic -Wextra.
The test programs are not as strictly bound, but this should
improve the chances that atleast the core library compiles with
most compilers without warnings.
Replaced the confusing pb_enc_* functions with new pb_encode_* functions that
have a cleaner interface. Updated documentation.
Got rid of the endian_copy stuff in pb_encode.c, instead using C casts to do it automatically.
This makes the code safer and also reduces binary size by about 5%.
Fixes Issue 6.
Stream state was not copied back from substream in pb_enc_submessage,
which caused garbage output if the stream callback modified the state.
Expanded tests to cover this problem.
Thanks to Paweł Pery for debugging and reporting this problem.
git-svn-id: https://svn.kapsi.fi/jpa/nanopb@1089 e3a754e5-d11d-0410-8d38-ebb782a927b9
Most importantly, callback fields in submessages were being overwritten with garbage, causing segfaults.
Additionally, converted PB_LTYPE_FIXED to PB_LTYPE_FIXED32 and PB_LTYPE_FIXED64. This makes the interface
a bit easier to use, and in addition runs faster.
git-svn-id: https://svn.kapsi.fi/jpa/nanopb@975 e3a754e5-d11d-0410-8d38-ebb782a927b9
