RE: [Mingw-users] version numbers galore

[Randy W. S. <RandyS@ThePierianSpring.org>]

> -----Original Message-----
> From: min...@li...
> [mailto:min...@li...]On Behalf Of Earnie Boyd
> Sent: Monday, September 09, 2002 2:03 PM
> To: Al Stevens
> Cc: min...@li...
> Subject: Re: [Mingw-users] version numbers galore
>
>
> I may be able to supply something in the MinGW.INI file found in the
> <WINDOWS> directory.  I'll get back to you on that.  Yes, it doesn't
> help currently but maybe in the future.
>
> Earnie.
>
> Al Stevens wrote:
> >
> > > The GCC 3.2 binary package was released before Mingw 2.0, and
> for each of
> > > the -1, -2, -3 updates the compiler itself has not changed,
> so it makes no
> > > sense to rebuild it. AFAIK the displayed timestamp has not
> changed either.
> >
> > Hmm. That tends to emphasize the need for something available at runtime
> > that identifies the distribution being used. The information
> being displayed
> > is ambiguous assuming that -1, -2, and -3 all display the same
> timestamp and
> > assuming that there is anything whatsoever different in the
> updates, which,
> > of course, there would be.
> > ==================
> > gcc -v
> > [snip]
> > gcc version 3.2 (mingw special 20020817-1)
> > ==================
> > Given that three different updates display the same
> > release/version/update/whatever timestamp text, and that no
> other display or
> > recorded data augment it, this display is indeed ambiguous. As any
> > self-righteous configuration manager will tell you. :-)
> >
> > Inasmuch as the "mingw sepcial" text displayed by gcc -v is not
> in the specs
> > file, I assume that the gcc program itself gets changed to display it,
> > unless there's another file somewhere that contains
> port-specific version
> > data. Incidentally, 3.1 doesn't display any such string whereas
> 2.95.2 and
> > 2.95.3-6 do, so something gets changed somewhere. Wherever the
> text comes
> > from, it could include the 2.0.0-3 information. But if I'm the
> only one who
> > needs or wants it, I won't ask the volunteers to burn any valuable time
> > adding it. I'll just get by with whatever is available.
> >
> > But, to continue flailing this deceased equine, I'd think the
> project would
> > need it for its own purposes. If someone downloads, installs, and then
> > deletes the self-extracting exe to make space, and then asks a question,
> > what question do you ask them to know which
> release/update/whatever they are
> > using?
> >
> > (The obvious answer to that question is, "Our users are programmers and
> > smart people; they can figure it out," or, "The question or problem will
> > reveal the version." My users are students, some of whom ask
> questions you
> > wouldn't believe.)
> >
> > Al Stevens
> > Dr. Dobb's Journal
> >

Wouldn't the versions of the individual packages be better? Either way
if gcc is the compiler you should be able to have your ide compile the
code below and get the versions of the packages. Then, if you need to
know the version of the MinGW distribution, you can match these
versions against a table consisting of the versions of packages
included in each of the MinGW distributions.

HTH,
Randy.

#include <stdio.h>
#include "w32api.h"
#include "_mingw.h"

int
main( int argc, char **argv )
{
  printf( "MinGW w32api version %d.%d\n",
          __W32API_MAJOR_VERSION,
          __W32API_MINOR_VERSION
        );

  printf( "MinGW Runtime version %d.%d\n",
          __MINGW32_MAJOR_VERSION,
          __MINGW32_MINOR_VERSION
        );

  printf( "GCC version %d.%d\n",
          __GNUC__,
          __GNUC_MINOR__
        );

  return 0;
}