Michael Matz wrote:
On Sat, 15 Oct 2016, David Mertens wrote:
Ah, now I get errors running the preprocessor tests. Here's a
PPTest 14 ...
Yes. The whole commit range around there is a bit shaky. You need some
commits from after that one to get a better working compiler.
That is true. One of those would erase your /usr/share/doc completely
(with "make uninstall"). Current mob should be fully working though.
(Christian would have been the first to have told otherwise).
As to merge pains with your exsymtab I can imagine well and apologize
however I guess it is normal that git can't make it happen fully
automagically always. If you want to suggest some LIBTCCAPI addition
in order to make your life easier, that is to support your extension
in a more modular way then, why not. Thinking about API's is often
a very good thing that can help to improve the code structure.
As to winapi header, well, my opinion is in the comment from
and in the tcc-win32.txt, that is that I'd prefer to have just
the minimal set required to compile win32/examples/hello_win.c.
Of course always someone will find that "with just that more file
it is possible now to compile project xyz".
But it is IMHO not soo difficult for someone to get files from the
mingw w32api pack (rsp. mingw64) and use them in the private project.
Another point is that our files are quite outdated now (from 2009),
also they were from one of the very first mingw-64 beta releases.
Then again in the meantime, people started to fix these and to add
missing functions, whereas originally my intention was to leave them
as unchanged as possible (with exception of _mingw.h which is to
make everything compatible with tcc) such that we could replace them
with newer versions easily, eventually. I don't know whether this
still make sense.
Another though is that we could upload a more complete header pack
as a separate download on savannah (provided that someone is willing
to create it).
Tinycc-devel mailing list