You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@subversion.apache.org by Peter Samuelson <pe...@p12n.org> on 2010/11/18 20:11:21 UTC
Re: svn commit: r1036534 - /subversion/trunk/Makefile.in
[stsp@apache.org]
> swig-pl: $(swig-pl_DEPS)
> - if test "$(SWIG_PL_DIR)" != "$(SWIG_PL_SRC_DIR)"; then \
> + if test "`$(READLINK) $(SWIG_PL_DIR)`" != "`$(READLINK) $(SWIG_PL_SRC_DIR)`"; then \
Does everyone have pwd -P? Seems to be specified by POSIX. That seems
like a pretty good 'readlink' to me:
if test "$$(cd $(SWIG_PL_DIR); pwd -P)" != "$$(cd $(SWIG_PL_SRC_DIR); pwd -P)"; then
--
Peter Samuelson | org-tld!p12n!peter | http://p12n.org/
Re: svn commit: r1036534 - /subversion/trunk/Makefile.in
Posted by Philip Martin <ph...@wandisco.com>.
Stefan Sperling <st...@elego.de> writes:
> On Thu, Nov 25, 2010 at 11:45:08AM +0200, Daniel Shahaf wrote:
>> Philip Martin wrote on Thu, Nov 25, 2010 at 09:31:39 +0000:
>> > Stefan Sperling <st...@elego.de> writes:
>> >
>> > > On Thu, Nov 18, 2010 at 09:37:21PM +0100, Stefan Sperling wrote:
>> > >> I guess the python libraries really are more reliable in this case.
>> > >
>> > > Point being that we require python anyway to compile Subversion.
>> >
>> > Is that right? I thought it was possible to compile a tarball without
>> > using Python? If so it should be possible to build the Perl bindings on
>> > a platform that doesn't have Python.
>> >
>>
>> I had the same thought.
>
> Yes please. That reasoning makes sense.
> I didn't think of the compile-from-tarball case.
r1039040. The SWIG Perl bindings build on my Linux box.
--
Philip
Re: svn commit: r1036534 - /subversion/trunk/Makefile.in
Posted by Stefan Sperling <st...@elego.de>.
On Thu, Nov 25, 2010 at 11:45:08AM +0200, Daniel Shahaf wrote:
> Philip Martin wrote on Thu, Nov 25, 2010 at 09:31:39 +0000:
> > Stefan Sperling <st...@elego.de> writes:
> >
> > > On Thu, Nov 18, 2010 at 09:37:21PM +0100, Stefan Sperling wrote:
> > >> I guess the python libraries really are more reliable in this case.
> > >
> > > Point being that we require python anyway to compile Subversion.
> >
> > Is that right? I thought it was possible to compile a tarball without
> > using Python? If so it should be possible to build the Perl bindings on
> > a platform that doesn't have Python.
> >
>
> I had the same thought.
Yes please. That reasoning makes sense.
I didn't think of the compile-from-tarball case.
> > We can assume that anyone building the Perl bindings has Perl, so how
> > about using a perl version of the Python command:
> >
> > READLINK_PL=$(PERL) -e 'use Cwd; print Cwd::realpath($ARGV[0])'
>
> Need to double the second $ sign. (or s/$ARGV[0]/shift/)
Re: svn commit: r1036534 - /subversion/trunk/Makefile.in
Posted by Daniel Shahaf <d....@daniel.shahaf.name>.
Philip Martin wrote on Thu, Nov 25, 2010 at 09:31:39 +0000:
> Stefan Sperling <st...@elego.de> writes:
>
> > On Thu, Nov 18, 2010 at 09:37:21PM +0100, Stefan Sperling wrote:
> >> I guess the python libraries really are more reliable in this case.
> >
> > Point being that we require python anyway to compile Subversion.
>
> Is that right? I thought it was possible to compile a tarball without
> using Python? If so it should be possible to build the Perl bindings on
> a platform that doesn't have Python.
>
I had the same thought.
> We can assume that anyone building the Perl bindings has Perl, so how
> about using a perl version of the Python command:
>
> READLINK_PL=$(PERL) -e 'use Cwd; print Cwd::realpath($ARGV[0])'
Need to double the second $ sign. (or s/$ARGV[0]/shift/)
Re: svn commit: r1036534 - /subversion/trunk/Makefile.in
Posted by Philip Martin <ph...@wandisco.com>.
Stefan Sperling <st...@elego.de> writes:
> On Thu, Nov 18, 2010 at 09:37:21PM +0100, Stefan Sperling wrote:
>> I guess the python libraries really are more reliable in this case.
>
> Point being that we require python anyway to compile Subversion.
Is that right? I thought it was possible to compile a tarball without
using Python? If so it should be possible to build the Perl bindings on
a platform that doesn't have Python.
We can assume that anyone building the Perl bindings has Perl, so how
about using a perl version of the Python command:
READLINK_PL=$(PERL) -e 'use Cwd; print Cwd::realpath($ARGV[0])'
--
Philip
Re: svn commit: r1036534 - /subversion/trunk/Makefile.in
Posted by Stefan Sperling <st...@elego.de>.
On Thu, Nov 18, 2010 at 09:37:21PM +0100, Stefan Sperling wrote:
> I guess the python libraries really are more reliable in this case.
Point being that we require python anyway to compile Subversion.
Stefan
Re: svn commit: r1036534 - /subversion/trunk/Makefile.in
Posted by Stefan Sperling <st...@elego.de>.
On Thu, Nov 18, 2010 at 02:11:21PM -0600, Peter Samuelson wrote:
>
> [stsp@apache.org]
> > swig-pl: $(swig-pl_DEPS)
> > - if test "$(SWIG_PL_DIR)" != "$(SWIG_PL_SRC_DIR)"; then \
> > + if test "`$(READLINK) $(SWIG_PL_DIR)`" != "`$(READLINK) $(SWIG_PL_SRC_DIR)`"; then \
>
> Does everyone have pwd -P? Seems to be specified by POSIX. That seems
> like a pretty good 'readlink' to me:
>
> if test "$$(cd $(SWIG_PL_DIR); pwd -P)" != "$$(cd $(SWIG_PL_SRC_DIR); pwd -P)"; then
I considered that, too. But I don't know if everyone has that.
OpenBSD has it implemented but it's intenionally not documented.
The pwd source says:
"The POSIX 1003.2B/D9 document has an optional -P flag"
Note the word "optional".
I guess the python libraries really are more reliable in this case.
Stefan