Re: libssh2 master 42aefdb configure.ac: Call zlib zlib and not libz in text but keep option names
On 16.09.2013 13:08, Peter Stuge wrote:
> I understand that it's tempting to compensate for weak build
> environments by abusing the C preprocessor but I think we can
> do a bit better than that.
oh, then I'm glad to learn about better methods.
Also I dont see where we abuse the C preprocessor when we use it to
build code parts conditional - isnt this one if its features?
> We'll obviously have conditions per make tool, like we have the rest
> of the makefiles, but as I wrote in fact there is just nmake to
> consider since all other platforms are hardcoded to use OpenSSL. Easy!
what we have just now dont has to be the end solution for ever; instead
it just reflects what has been done and used by now; if tomorrow someone
ports gcrypt to another platform and then wants to contribute the
resulting changes it should be as easy as it is now.
> I can't take that seriously. The objects.mk thing is pretty awful,
> there's no reason not to take advantage of the few features that
> nmake actually does offer.
again, I'm happy to learn about better methods;
I see it as improvement what I hacked up since before it happened every
now and then that we released a version with broken nmake makefiles -
that should now not happen again;
but then I'm not a nmake expert - if you are then show us how it can be
Received on 2013-09-16