[cfe-dev] printf-like function.

[David Blaikie via cfe-dev]

https://clang.llvm.org/get_involved.html discusses the considerations for
proposing extensions to clang - likely this might be best prototyped in a
fork of clang (clang being on github, this should be fairly low-cost) & if
a significant user base forms around it, and/or it's the sort of thing you
could propose for standardization in the relevant bodies (I guess getting a
proposal for the C standard would be the place to start) then it might be
good to discuss upstreaming it into clang proper

On Fri, Mar 20, 2020 at 7:15 AM JF Knudsen via cfe-dev <
cfe-dev at lists.llvm.org> wrote:

> There is no contradictions.
>
> fmt doesn't work with C. I'm also using C++ in freestanding env without
> STL on small devices with no dynamic memory allocation.
>
> I still think that would be a nice addition, with no need to modify
> existing libraries.
>
> On 20.03.20 17:06, Keane, Erich via cfe-dev wrote:
> > What about the fmt library? https://github.com/fmtlib/fmt
> >
> > A version of this was accepted to the c++20 spec, so in addition it'll
> be portable.
> >
> > -----Original Message-----
> > From: cfe-dev <cfe-dev-bounces at lists.llvm.org> On Behalf Of JF Knudsen
> via cfe-dev
> > Sent: Thursday, March 19, 2020 11:53 PM
> > To: cfe-dev at lists.llvm.org
> > Subject: [cfe-dev] printf-like function.
> >
> > Hi all,
> >
> > I'm a printf-like function lover. I always found that better than doing
> something like out << "blabla : " << var1 << ", blabla" << etc.
> >
> > Now, I use a lot of different platforms and to be portable, I'm supposed
> to use PRIxxx constants for portability. For me that half-breaks the
> readability of printf format.
> >
> > What about a new set of functions that'll be specified slightly
> differently :
> >
> > accepting %d for ANY number ? The compiler can modify the format
> litteral at compile time to fit what's needed for that platform, replacing
> %d by %hhd, %hd, %d, %ld, %lld, %hhu etc..
> > Replacing the litteral format at compile time prevents to modify the
> vararg mechanism.
> >
> > Of course, if format is not litteral : no replacement !
> >
> > OR
> >
> > 2bis) some attributes could be passed on to this function so it'll know
> what to expect at runtime. I'm guessing modifying vararg mechanism to add
> attributes might be nearly impossible...
> >
> >
> > I'm interested of what would you all think about that ?
> >
> > Thanks.
> >
> >
> > PS : we could do the same with %s that could accept char*, wchar_t*,
> char16_t*, char32_t*.
> >
> > _______________________________________________
> > cfe-dev mailing list
> > cfe-dev at lists.llvm.org
> > https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev
> > _______________________________________________
> > cfe-dev mailing list
> > cfe-dev at lists.llvm.org
> > https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev
> _______________________________________________
> cfe-dev mailing list
> cfe-dev at lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20200320/50a1b72d/attachment.html>