Re: [gentoo-dev] [PATCH 2/2] eclass/toolchain-funcs: add clang version functions

On Tue, Jul 05, 2016 at 01:08:13AM -0500, Austin English wrote
>
> My goal is clang support parity with gcc. If you are opposed to these
> sort of checks, then why don't we deprecate and remove those functions?
> I want to know why gcc deserves special treatment, either all compilers
> should have easy way to check major/minor/full versions, or none should.
> Obviously I'm not saying gcc should be removed now, but it could
> certainly be marked deprecated so the usage doesn't spread (hopefully)
> further.

  This is reminiscent of the web-browser situation.  I use Pale Moon.
It's feature-compatible with Firefox, but has not gone berserk with the
version numbering.  The current Firefox is 49-point-something.  Stupid
webpages see Pale Moon 26.3.3 and whine about "out-of-date-web-browser"
and kick the user out.  But if the user sets the user agent (i.e. lies
to the webpage) that he's using Firefox 49.1, it works just fine.

  It's not unique to the current FOS world, either.  Some old MS-DOS
applications would only run when the OS reported a certain narrow range
of versions.  When you updated MS-DOS, some older applications would
refuse to run, even though the newer MS-DOS was perfectly capable of
running it.  Things got so bad that Microsoft introduced the SETVER
command https://support.microsoft.com/en-us/kb/96767 to deliberately
lie about the MS-DOS version number, when queried by specific
applications.

   How is the version checking done?  Does the check parse the file name
of the compiler?  Can we get the GCC and CLANG people to agree to a
common command/parameter that returns a compatibility level for
"version-checking"?

-- 
Walter Dnes <waltdnes@××××××××.org>
I don't run "desktop environments"; I run useful applications