Compile problems with 1.0.6
bob at ti.com
Wed May 30 21:58:02 CEST 2001
On Wed, May 30, 2001 at 06:54:35PM +0200, Florian Weimer wrote:
> Werner Koch <wk at gnupg.org> writes:
> > On Wed, 30 May 2001, Florian Weimer wrote:
> > > Well, that depends on our goals. For example, Emacs has to be
> > > implemented in K&R C in order to have it available on a multitude of
> > > platforms.
> > IMHO this is not so important anymore. At least one version of gcc
> > is running on each platform and this one can be used too bootstrap
> > a development platform.
> I think it's not clear that the latest version of GCC running on such
> platforms supports C99. Current GCC support on some common targets is
> quite bit poor.
There is also the issue that on some platforms it may be preferable to use
the vendor's own compiler instead of gcc. This varies from platform to
platform, but last time I heard anything about this, some vendor compilers
ran significantly faster than gcc on their platforms, because the vendor
has taken significant steps to optimise for their own hardware.
Whilst gcc is a fallback option if the native compiler can't handle it, it
would still be preferable if the source is broadly compatible where possible,
to minimise the number of times people have to switch.
If I recall correctly, the original issue which sparked this was a trailing
comma - if one compiler barfs on this, whereas both are happy when the comma
is not there, then hopefully it is not too difficult to avoid trailing commas
in this situation, and thus keep both compilers happy.
I proffer this suggestion without having contributed any code to gpg, so
feel free to ignore me / shout me down... :-)
Bob Luckin bob at ti.com "Coder, adapt; FTP Ada, redo C"
More information about the Gnupg-devel