Bugs for Sorcery are reported here

[bugzilla-daemon AT bugs.sourcemage.org]

http://bugs.sourcemage.org/show_bug.cgi?id=14620





--- Comment #6 from fredrik.carlstrom AT tele2.se  2008-08-07 17:14:02 ---
(In reply to comment #5)

I see! Perhaps @param part-of-version would be better? This would reflect the
fact that there are no checking on the parameter and that checking for 4.3
would also match in 3.4.3 as you correctly noted.

If you look at the example in the original post you will see that the less
strict requirements on the parameter is intensional (but the return statement
should be: returns 0 iff the versions of GCC is at least 4.3.0.) I see now 
that
the opposite would be a problem. use_gcc 4.3.0 would return the wrong value if
gcc -dumpversion somehow could evaluate to 4.3. I do not know if this is the
case, but I can change this if needed, or else stress that trailing zeros
should not be given in the parameter.

There is no risk that the function should (re)match the wrong bits in the 
loop.
This can only happen when the parameter is given without the trailing +.

The reason for the separation of ARG=${ARG#${TEMP1}} from ARG=${ARG#.} is just
for the case that there might be a different number of digit groups in the
parameter and in $(gcc -dumpversion). The construction ${ARG#{TEMP1}.} will 
not
remove the last group of digits.


-- 
Configure bugmail: http://bugs.sourcemage.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.