This avoids double initialization when decoding nested submessages.
Fixes an issue with submessage arrays that was present in previous
version of this patch.
Update issue 28
Status: FixedInGit
This reverts commit dc2da0edc5.
Add pb_close_string_substream() for copying back the state.
This makes adding error messages easier in the future, as also
them need to be propagated back from the substream.
This makes the field decoding functions more intuitive to use.
The old interface is still present if you specify NANOPB_INTERNALS.
Update issue 2
Status: FixedInGit
This makes it unnecessary to copy back the state, and also relaxes
the requirements on callbacks (bytes_left will always be valid).
It decreases code size by a few bytes, but may be just slightly slower.
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.
Now the limit of tracked fields is configurable at compile-time using
PB_MAX_REQUIRED_FIELDS. Added related test and updated documentation.
Fixes issue #18.
The size of non-callback bytes-fields was miscalculated, which
caused all following fields in a message to contain garbage.
Previous commit contains a testcase for this.
This fix changes the generated message description. If your protocol uses
bytes-fields, you should regenerate *.pb.c.
On big endian machines, pb_dec_fixed32 and pb_dec_fixed64 were
incorrectly overwriting the result value even if reading failed.
git-svn-id: https://svn.kapsi.fi/jpa/nanopb@1022 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
