[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Building and Testing Python, Perl, and Ruby Bindings

From: Paul Burba <paulb_at_softlanding.com>
Date: 2006-07-26 15:42:01 CEST

"D.J. Heap" <djheap@gmail.com> wrote on 07/24/2006 04:44:16 PM:

> On 7/24/06, Paul Burba <paulb@softlanding.com> wrote:
> [snip]
> >
> > Thanks for taking a look, unfortunately the problems you found with
> > $(top_srcdir), #error, and #warning appear unrelated to my swig
> > difficulties. I tried 1.3.25, but building with VC++6.0 gives the
same
> > errors:
> >
> > Found installed SWIG version 1.3.25
> > Using SWIG library directory C:\SVN\SWIG\Lib
> > Found ML, ZLib build will use ASM sources
> > Found neon version 0.25.5
> > Configuring apr-util library...
> > perl "..\httpd-2.0.58\srclib\apr-util\build\w32locatedb.pl" dll
> > "db4-win32\include" "db4-win32\lib"
> > Using BDB version 4.4.20
> [snip]
> > The file
C:\SVN\src-trunk\build\win32\msvc-dsp\swig_ruby_repos_msvc.dsp
> > has been modified and cannot be loaded as a Developer Studio project.
> > The file C:\SVN\src-trunk\build\win32\msvc-dsp\swig_ruby_wc_msvc.dsp
has
> > been modified and cannot be loaded as a Developer Studio project.
> >
> > I've googled, altavistaed, lycosed, and otherwise yahooed this error
> > message to within an inch of its life and haven't found anything
helpful
> > except some vague discussion of dsp file corruption due to line ending
> > problems.
> >
> > I don't imagine this matters, but I'm running these tests on a VMWare
> > virtual machine. I remain completely baffled at the moment but will
> > continue to look into this.
>
>
> Very strange -- I'm running everything under VMWare Workstation 5.5.1,
> by the way, so that doesn't seem likely to be the problem. I'm using
> VS6 with Service Pack 6 with the latest Platform SDK and not seeing
> any problems aside from the top_srcdir warning anymore.
>
> Actually, I did have to tweak a Python header to comment out a line
> that was defining socklen_t or something like that -- the lastest
> Platform SDK typedef's it which was conflicting with Python's #define.
>
> Anyway, if you figure out anything else please share.

A quick update:

Building 1.4.x r20780 (i.e. RC4) on my regular (non-virtual) machine and
using swig 1.3.25, I was able to build the Python bindings with only the
top_srcdir warning. So now it's a *simple* matter of figuring out how my
VM and regular machines differ :-P I'm putting this aside for a while,
but when I eventually track down the problem I'll post an explanation.

I was also able to build the Perl bindings with only the top_srcdir
warning as well (with r20826 applied). And while I was at it, I looked at
r20190, r20250, and r20444 too. I voted for all but 20250. I'm not
familiar enough with gen-win.py to judge this last one, but I was able to
build and test without any problems so I gave it +0.

Paul B.

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Wed Jul 26 15:50:06 2006

This is an archived mail posted to the Subversion Dev mailing list.