On Mar 28, 2012, at 2:55 AM, Henri Gomez wrote:

>> If you add --miredebug to the invocation, you will see the
>> result of every pattern match.
>> 
>> The patterns are used to ensure that there aren't any "surprises".
> 
> Here is result with --miredebug :
> 
> 
> Version to package is 3.0.3
> --> mire 0x7fff32c15ae0 ++ 1 mireNew at mire.c:115

This is a created (and refcnt++) of a MIRE_t object. What I tend
to call a "ctor" abusing exisyting terminology.

> <-- mireRegcomp(0x7fff32c15ae0, "^[A-Za-z0-9][A-Za-z0-9+._-]*$") rc 0
> <-- mireRegexec(0x7fff32c15ae0, 0x7fff33010e06[8]) rc 0 mode 2 "cimaven3"

This is a compiled and matched PCRE expression.

> --> mire 0x7fff32c15ae0 -- 1 tagValidate at parsePreamble.c:592
> <-- mireClean(0x7fff32c15ae0)

And this is --refcnt "dtor" aka a free (though the object is recycled through
a memory pool).

> --> mire 0x7fff32c15ae0 ++ 1 mireNew at mire.c:115
> <-- mireRegcomp(0x7fff32c15ae0, "^[A-Za-z0-9+._]+$") rc 0
> error: ^[A-Za-z0-9+._]+$: regexec failed: regexec() failed to match(1)
> <-- mireRegexec(0x7fff32c15ae0, 0x7fff33010e09[5]) rc -2 mode 2 "3.0.3"
> error: line 8: invalid tag value("^[A-Za-z0-9+._]+$") Version: Version: 3.0.3

This is a failed match on the string "3.0.3" which SHOULD have Just Worked.
There's still some chance of symbol pollution if you did not
        make distclean
before building. Slightly quicker is to do
        cd misc
        make clean
and rebuild. Check that "pcre_regcomp" exists in ./lib/librpmmisc.a.

Try the C locale first, then double check the options
following
        --with-pcre=internal
in the %falmouth stanza and make sure those are passed when you build.

> --> mire 0x7fff32c15ae0 -- 1 tagValidate at parsePreamble.c:592
> <-- mireClean(0x7fff32c15ae0)
> error: Package has no %description: cimaven3.x86_64
> 
> rpmbuild command is :
> 
> rpmbuild -bb --define="_topdir $PWD" --define="_tmppath $PWD/TEMP"
> --define="MAVEN_RE 3.0.3"  --miredebug SPECS/cimaven3.spec
> 
> And why such error about description ?
> 

Because rpmbuild proceed "best effort" and there is a cascade
effect when code is run further after a "can't happen" error.

Fix the first problem first.

There are simpler builds than maven for confidence *ahrm* building.

73 de Jeff


> ...
> BuildArch:  noarch
> ...
> %description
> Maven %{maven_rel}
> ______________________________________________________________________
> RPM Package Manager                                    http://rpm5.org
> User Communication List                             rpm-users@rpm5.org

______________________________________________________________________
RPM Package Manager                                    http://rpm5.org
User Communication List                             rpm-users@rpm5.org

Reply via email to