Kang-Che Sung (宋岡哲)

@westes Gosh. There're many similarly named "libobj" library out there on the web, so it was not easy to see which one you were talking about. Well, I could make...

@lexxmark If it's me, I would like to have htons/htonl function names preserved. The reason is they're POSIX, even though they're not part of standard (ISO) C (I wish it...

@GitMensch Ideally we should let `tests/tableopts.sh` run *during* the running the autotools, or at least not later than `./configure`. Forcing another step before `autoreconf` would be looking for trouble when...

Excuse me, but why do we need to care about the `memmove` vs. `memcpy` performance difference? The two functions have different purposes and one cannot be substituted for another. If...

@jafl Since you performed the `memmove` test last, it is likely that the shorter time you observed is not from the implementation of `memmove` itself, but because the memory is...

For your info, we used to have a problem with the `BUILT_SOURCES` automake variable. See #144. Now I always prefer explicit dependency tracking instead of that variable.

I think this missed a check that MSVC compiles code only for little-endian machines. Otherwise the semantic would be wrong. `htonl` and `htons` should be no-op when compiled for big-endian...

I would argue that it's part of the design to not process m4preproc macros in Flex.

What is EOB? And what is your use case for needing that thing?

@eqvinox I did not see any problem when reading your patch, but I'm not sure if this thing can be tested, to check if the bug is really fixed, and...