This is the mail archive of the 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]: fix PTHREAD_ONCE_INIT when using -ansi

 >  > How about:
 >  > 
 >  > mach = '*-*-solaris2.[1-9]', '*-*-solaris2.[1-9][!0-9]*';
 > Ah, I didn't know "mach" could use commas.  So the rule is must match
 > *all* patterns in a comma list, but *any* pattern among separate
 > "mach" values, right?  That's not documented either... :-)
 >  > That will catch all pre-10 Solari, regardless of any suffix after
 >  > the 5.N gotten from `uname -r`.
 > Ok but I don't see how it matches 5.N, as in solaris2.5.1.  I thought
 > glob patterns are explicitly anchored.  And it doesn't match
 > solaris2.0, although I don't know if gcc works there anymore,
 > configure still supports it.
 > So how about:
 >     mach = '*-*-solaris2.[0-9]';
 >     mach = '*-*-solaris2.[0-9].*';
 > I verified this does the right thing on solaris7, 9 and 10.  I don't
 > have 2.5.1 to test but I think this is correct.

Oh wait, I see my mistake.  The comma form is the same as multiple
"mach" form.  So in the second pattern, the [!0-9] matches the dot
between "5.1" and the trailing asterisk gets the trailing "1".

I think using the [!0-9] is a little non-intuitive, I like the
2.[0-9].* form better, I can use a comma instead of two "mach" lines
if you prefer.

So okay with the aforementioned pattern change?

Kaveh R. Ghazi

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