Configure doesn't pick up Hash/Util/FieldHash

I've been trying to get perl5.10.0 working on QNX4.  I'm now
able to compile
and pass or confidently skip almost all of the tests, and
all of the critical
ones.

The only problem I've found which might affect other
platforms is that
Configure fails to locate the Hash/Util/FieldHash extension
when it
runs find_extensions.  This means it doesn't get bound
statically into
the produced perl, and the world ends in fireworks.

I've fixed it using the simple approach of hard-coding it
in, but someone
that enjoys playing with Configure might want to do this
better (chunk
below).

I've also had to mess a bit with Time::Piece to get that to
compile, should
I send those changes to p5p or to the ext module or both?

I'll package up the qnx specific changes in a bit, once I've
got more tests
going...

Alex Gough

########################################
--- Configure.orig      2007-10-07 22:45:30.000000000 +0000
+++ Configure   2007-10-07 23:04:20.000000000 +0000
 -21059,7
+21059,8 
 # Special case:  Add in threads/shared since it is not
picked up by the
 # recursive find above (and adding in general recursive
finding breaks
 # SDBM_File/sdbm).  A.D.  10/25/2001.
-known_extensions="$known_extensions
threads/shared"
+# ajgough: 200710 FieldHash doesn't get found
+known_extensions="$known_extensions threads/shared
Hash/Util/FieldHash"
 set X $known_extensions
 shift
 known_extensions="$*"

On Mon, 8 Oct 2007 00:29:11 +0100, Alex Gough
<alex-p5pearth.li> wrote:

> I've been trying to get perl5.10.0 working on QNX4. 
I'm now able to compile
> and pass or confidently skip almost all of the tests,
and all of the critical
> ones.
> 
> The only problem I've found which might affect other
platforms is that
> Configure fails to locate the Hash/Util/FieldHash
extension when it
> runs find_extensions.  This means it doesn't get bound
statically into
> the produced perl, and the world ends in fireworks.

Thanks, applied with a little twist for the comment part:

Change 32061 by merijnmerijn-nb09 on 2007/10/08 08:23:08

        Subject: [Patch] Configure doesn't pick up
Hash/Util/FieldHash
        From: Alex Gough <alex-p5pearth.li>
        Date: Mon, 8 Oct 2007 00:29:11 +0100
        Message-ID: <20071007232910.GA24241the.earth.li>

Affected files ...

... //depot/perl/Configure#667 edit

Differences ...

==== //depot/perl/Configure#667 (xtext) ====

 -26,7
+26,7 

 # $Id: Head.U,v 3.0.1.9 1997/02/28 15:02:09 ram Exp $
 #
-# Generated on Sun Sep  9 11:36:23 CEST 2007 [metaconfig
3.0 PL70]
+# Generated on Mon Oct  8 10:22:27 CEST 2007 [metaconfig
3.0 PL70]
 # (with additional metaconfig patches by perlbugperl.org)

 cat >c1$$ <<EOF

-21056,10 +21056,12 
 set X
 shift
 eval $find_extensions
-# Special case:  Add in threads/shared since it is not
picked up by the
-# recursive find above (and adding in general recursive
finding breaks
-# SDBM_File/sdbm).  A.D.  10/25/2001.
-known_extensions="$known_extensions
threads/shared"
+# Special case:  Add in modules that nest beyond the first
level.
+# Currently threads/shared and Hash/Util/FieldHash, since
they are
+# not picked up by the recursive find above (and adding in
general
+# recursive finding breaks SDBM_File/sdbm).
+# A.D. 20011025 (SDBM), ajgough 20071008 (FieldHash)
+known_extensions="$known_extensions threads/shared
Hash/Util/FieldHash"
 set X $known_extensions
 shift
 known_extensions="$*"

> I've fixed it using the simple approach of hard-coding
it in, but someone
> that enjoys playing with Configure might want to do
this better (chunk
> below).
> 
> I've also had to mess a bit with Time::Piece to get
that to compile, should
> I send those changes to p5p or to the ext module or
both?
> 
> I'll package up the qnx specific changes in a bit, once
I've got more tests
> going...
> 
> Alex Gough
> 
> ########################################
> --- Configure.orig      2007-10-07 22:45:30.000000000
+0000
> +++ Configure   2007-10-07 23:04:20.000000000 +0000
>  -21059,7 +21059,8 
>  # Special case:  Add in threads/shared since it is not
picked up by the
>  # recursive find above (and adding in general
recursive finding breaks
>  # SDBM_File/sdbm).  A.D.  10/25/2001.
> -known_extensions="$known_extensions
threads/shared"
> +# ajgough: 200710 FieldHash doesn't get found
> +known_extensions="$known_extensions
threads/shared Hash/Util/FieldHash"
>  set X $known_extensions
>  shift
>  known_extensions="$*"
> 


-- 
H.Merijn Brand         Amsterdam Perl Mongers (http://amsterdam.pm.org/
)
using & porting perl 5.6.2, 5.8.x, 5.9.x   on HP-UX
10.20, 11.00, 11.11,
& 11.23, SuSE 10.0 & 10.2, AIX 4.3 & 5.2, and
Cygwin. http://qa.perl.org
http://mirrors.de
velooper.com/hpux/            http://www.test-smoke.org
                        http
://www.goldmark.org/jeff/stupid-disclaimers/

Rafael Garcia-Suarez wrote:
> On 08/10/2007, Alex Gough <alex-p5pearth.li> wrote:
>> I've been trying to get perl5.10.0 working on QNX4.
 I'm now able to
>> compile and pass or confidently skip almost all of
the tests, and
>> all of the critical ones. 
>> 
>> The only problem I've found which might affect
other platforms is
>> that Configure fails to locate the
Hash/Util/FieldHash extension
>> when it 
>> runs find_extensions.  This means it doesn't get
bound statically
>> into 
>> the produced perl, and the world ends in
fireworks.
> 
> Ah, that's also why it doesn't get picked up when
Configuring perl
> with -Uusedl. 

And it's why Hash/Util/FieldHash failed to build statically
on Win32
when the ALL_STATIC option was added in #31784.

I've now applied this patch to win32/FindExt.pm too, so now
Hash/Util/FieldHash can be included in the ALL_STATIC
build.