This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH]: include GMP/MPFR version info in "gcc -v" output


On 2/17/07, Kaveh R. GHAZI <ghazi@caip.rutgers.edu> wrote:
On Sat, 17 Feb 2007, Richard Guenther wrote:

> > Thanks, I'll change it.  BTW the output looks like this:
> >
> >         GNU C version 4.3.0 20070216 (experimental) (sparc-sun-solaris2.10)
> >                 compiled by GNU C version 3.4.6.
> >         GGC heuristics: --param ggc-min-expand=100 --param ggc-min-heapsize=131072
> >         GMP: header version <4.2.1>, library version <4.2.1>
> >         MPFR: header version <2.2.1>, library version <2.2.1>
>
> The <> look inconsistent with the GCC version numbers.  I'd put the
> information next to the compiler used for building gcc, like
>
>    compiled by GNU C version 3.4.6, GMP version 4.2.1, MPFR version 2.2.1.
>
> and print the header/library version only on a mismatch as a warning.
> Richard.


I'm not sure I agree about combining the two. With the combined form, it's not clear to the user that gcc is checking internally that the two versions are identical. They may be left wondering whether gcc is showing the header or lib version or has the smarts to check both.

Well, "compiled by" implies this is a header (compile-time) check. Also I'm not sure anymore why we need gmp / mpfr version but not for example glibc version or binutils version as both are checked for features at gcc compile time as well (and there are known to be broken versions of those as well).

Richard.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]