Hi PyLucene developers,

I'm trying to install PyLucene 3.5.0-3 on some Fedora 14 workstations for
one of the researchers in my department.
I carefully followed the instructions here
http://lucene.apache.org/pylucene/documentation/install.html
and I got to the expected bug with setuptools.

I then issued exactly the patch command that your software told me to,
namely:

patch -d /usr/lib/python2.7/site-packages -Nup0 <
/usr/network/tmp/pylucene-3.5.0-3
/jcc/jcc/patches/patch.43.0.6c11
patching file setuptools/extension.py

Hunk #1 succeeded at 32 (offset 4 lines).
patching file setuptools/command/build_ext.py
Hunk #1 FAILED at 85.
Hunk #2 succeeded at 177 (offset 7 lines).
Hunk #3 succeeded at 259 (offset 7 lines).
1 out of 3 hunks FAILED -- saving rejects to file
setuptools/command/build_ext.py.rej
root@sampras:/tmp

As you can see, it failed.  I've attached a typescript, in which I show
that I haven't modified the
standard python-setuptools from the distribution and I show that it's
running setuptools-0.6c11.
I've also attached the file of rejects listed above.

Since you have a pretty sophisticated patch setup, I'm hoping that you'll
be able to solve this pretty
easily.  Please let me know if there's any way I can help.

Thanks for your support,

Scott

--
Scott D. Anderson
Computer Science Department
Wellesley College
[EMAIL PROTECTED]
[EMAIL PROTECTED]

  Hi Scott,

On Tue, 24 Jan 2012, Scott D Anderson wrote:

> I'm trying to install PyLucene 3.5.0-3 on some Fedora 14 workstations for
> one of the researchers in my department.
> I carefully followed the instructions here
> http://lucene.apache.org/pylucene/documentation/install.html
> and I got to the expected bug with setuptools.
>
> I then issued exactly the patch command that your software told me to,
> namely:
>
> patch -d /usr/lib/python2.7/site-packages -Nup0 <
> /usr/network/tmp/pylucene-3.5.0-3
> /jcc/jcc/patches/patch.43.0.6c11
> patching file setuptools/extension.py
>
> Hunk #1 succeeded at 32 (offset 4 lines).
> patching file setuptools/command/build_ext.py
> Hunk #1 FAILED at 85.
> Hunk #2 succeeded at 177 (offset 7 lines).
> Hunk #3 succeeded at 259 (offset 7 lines).
> 1 out of 3 hunks FAILED -- saving rejects to file
> setuptools/command/build_ext.py.rej
> root@sampras:/tmp
>
> As you can see, it failed.  I've attached a typescript, in which I show
> that I haven't modified the

I see no typescript attachment ?

Andi..

> standard python-setuptools from the distribution and I show that it's
> running setuptools-0.6c11.
> I've also attached the file of rejects listed above.
>
> Since you have a pretty sophisticated patch setup, I'm hoping that you'll
> be able to solve this pretty
> easily.  Please let me know if there's any way I can help.
>
> Thanks for your support,
>
> Scott
>
> --
> Scott D. Anderson
> Computer Science Department
> Wellesley College
> [EMAIL PROTECTED]
> [EMAIL PROTECTED]
>

Hi Andi,

On Tue, Jan 24, 2012 at 3:35 PM, Andi Vajda <[EMAIL PROTECTED]> wrote:

>
>  Hi Scott,
>
>
> On Tue, 24 Jan 2012, Scott D Anderson wrote:
>
>  I'm trying to install PyLucene 3.5.0-3 on some Fedora 14 workstations for
>> one of the researchers in my department.
>> I carefully followed the instructions here
>> http://lucene.apache.org/**pylucene/documentation/**install.html<http://lucene.apache.org/pylucene/documentation/install.html>
>> and I got to the expected bug with setuptools.
>>
>> I then issued exactly the patch command that your software told me to,
>> namely:
>>
>> patch -d /usr/lib/python2.7/site-**packages -Nup0 <
>> /usr/network/tmp/pylucene-3.5.**0-3
>> /jcc/jcc/patches/patch.43.0.**6c11
>> patching file setuptools/extension.py
>>
>> Hunk #1 succeeded at 32 (offset 4 lines).
>> patching file setuptools/command/build_ext.**py
>> Hunk #1 FAILED at 85.
>> Hunk #2 succeeded at 177 (offset 7 lines).
>> Hunk #3 succeeded at 259 (offset 7 lines).
>> 1 out of 3 hunks FAILED -- saving rejects to file
>> setuptools/command/build_ext.**py.rej
>> root@sampras:/tmp
>>
>> As you can see, it failed.  I've attached a typescript, in which I show
>> that I haven't modified the
>>
>
> I see no typescript attachment ?
>

Sorry for the confusion.  I named it pylucene-install.log, but that's just
a rename
of the typescript file.

Scott

>
> Andi..
>
>
>  standard python-setuptools from the distribution and I show that it's
>> running setuptools-0.6c11.
>> I've also attached the file of rejects listed above.
>>
>> Since you have a pretty sophisticated patch setup, I'm hoping that you'll
>> be able to solve this pretty
>> easily.  Please let me know if there's any way I can help.
>>
>> Thanks for your support,
>>
>> Scott
>>
>> --
>> Scott D. Anderson
>> Computer Science Department
>> Wellesley College
>> [EMAIL PROTECTED]
>> [EMAIL PROTECTED]
>>
>>
--
Scott D. Anderson
Computer Science Department
Wellesley College
[EMAIL PROTECTED]
[EMAIL PROTECTED]

On Tue, 24 Jan 2012, Scott D Anderson wrote:

> Sorry for the confusion.  I named it pylucene-install.log, but that's just
> a rename
> of the typescript file.

Your messages have no attachments. Maybe they're being filtered out.
Can you include it inline since it's plain text anyway ?

Andi..

>
> Scott
>
>>
>> Andi..
>>
>>
>>  standard python-setuptools from the distribution and I show that it's
>>> running setuptools-0.6c11.
>>> I've also attached the file of rejects listed above.
>>>
>>> Since you have a pretty sophisticated patch setup, I'm hoping that you'll
>>> be able to solve this pretty
>>> easily.  Please let me know if there's any way I can help.
>>>
>>> Thanks for your support,
>>>
>>> Scott
>>>
>>> --
>>> Scott D. Anderson
>>> Computer Science Department
>>> Wellesley College
>>> [EMAIL PROTECTED]
>>> [EMAIL PROTECTED]
>>>
>>>
>
>
> --
> Scott D. Anderson
> Computer Science Department
> Wellesley College
> [EMAIL PROTECTED]
> [EMAIL PROTECTED]
>

On Tue, Jan 24, 2012 at 4:37 PM, Andi Vajda <[EMAIL PROTECTED]> wrote:

>
> On Tue, 24 Jan 2012, Scott D Anderson wrote:
>
>  Sorry for the confusion.  I named it pylucene-install.log, but that's just
>> a rename
>> of the typescript file.
>>
>
> Your messages have no attachments. Maybe they're being filtered out.
> Can you include it inline since it's plain text anyway ?
> Andi..
Sure.  Here's the typescript:

Script started on Tue 24 Jan 2012 01:07:10 PM EST
[root@sampras tmp] cat /etc/issue
Fedora release 14 (Laughlin)
Kernel \r on an \m (\l)

[root@sampras tmp] rpm -qV python-setuptools
[root@sampras tmp] ls -l /usr/lib/python2.7/site-packages/setuptools*
-rw-r--r-- 1 root root  144 Oct 26 22:05
/usr/lib/python2.7/site-packages/setuptools-0.6c11-py2.7.egg-info
-rw-r--r-- 1 root root   34 Oct 26 22:05
/usr/lib/python2.7/site-packages/setuptools.pth

/usr/lib/python2.7/site-packages/setuptools:
total 300
-rw-r--r-- 1 root root  5927 Oct 26 22:04 archive_util.py
-rw-r--r-- 2 root root  5725 Oct 26 22:05 archive_util.pyc
-rw-r--r-- 2 root root  5725 Oct 26 22:05 archive_util.pyo
drwxr-xr-x 2 root root  4096 Jan 24 11:47 command
-rw-r--r-- 1 root root  6209 Oct 26 22:04 depends.py
-rw-r--r-- 2 root root  6749 Oct 26 22:05 depends.pyc
-rw-r--r-- 2 root root  6749 Oct 26 22:05 depends.pyo
-rw-r--r-- 1 root root 30309 Oct 26 22:04 dist.py
-rw-r--r-- 1 root root 30221 Oct 26 22:05 dist.pyc
-rw-r--r-- 1 root root 30163 Oct 26 22:05 dist.pyo
-rw-r--r-- 1 root root  1278 Oct 26 22:04 extension.py
-rw-r--r-- 2 root root  1761 Oct 26 22:05 extension.pyc
-rw-r--r-- 2 root root  1761 Oct 26 22:05 extension.pyo
-rw-r--r-- 1 root root  1278 Oct 26 22:04 extension.py.orig
-rw-r--r-- 1 root root  3744 Oct 26 22:04 __init__.py
-rw-r--r-- 2 root root  4028 Oct 26 22:05 __init__.pyc
-rw-r--r-- 2 root root  4028 Oct 26 22:05 __init__.pyo
-rw-r--r-- 1 root root 30447 Oct 26 22:04 package_index.py
-rw-r--r-- 2 root root 28625 Oct 26 22:05 package_index.pyc
-rw-r--r-- 2 root root 28625 Oct 26 22:05 package_index.pyo
-rw-r--r-- 1 root root  8756 Oct 26 22:04 sandbox.py
-rw-r--r-- 2 root root 11425 Oct 26 22:05 sandbox.pyc
-rw-r--r-- 2 root root 11425 Oct 26 22:05 sandbox.pyo
[root@sampras tmp] patch -d /usr/lib/python2.7/site-packages -Nup0 <
/usr/network/tmp/pylucene-3.5.0-3 /jcc/jcc/patches/patch.43.0.6c11
patching file setuptools/extension.py
Hunk #1 succeeded at 32 (offset 4 lines).
patching file setuptools/command/build_ext.py
Hunk #1 FAILED at 85.
Hunk #2 succeeded at 177 (offset 7 lines).
Hunk #3 succeeded at 259 (offset 7 lines).
1 out of 3 hunks FAILED -- saving rejects to file
setuptools/command/build_ext.py.rej
[root@sampras tmp] exit

Script done on Tue 24 Jan 2012 01:08:32 PM EST

and here's the contents of build_ex.py.rej

--- setuptools/command/build_ext.py    (revision 75864)
+++ setuptools/command/build_ext.py    (working copy)
@@ -85,8 +85,12 @@
         if fullname in self.ext_map:
             ext = self.ext_map[fullname]
             if isinstance(ext,Library):
+                if ext.force_shared and not use_stubs:
+                    _libtype = 'shared'
+                else:
+                    _libtype = libtype
                 fn, ext = os.path.splitext(filename)
-                return self.shlib_compiler.library_filename(fn,libtype)
+                return self.shlib_compiler.library_filename(fn,_libtype)
             elif use_stubs and ext._links_to_dynamic:
                 d,fn = os.path.split(filename)
                 return os.path.join(d,'dl-'+fn)
Thanks for your help.

Scott

>
>
>
>> Scott
>>
>>
>>> Andi..
>>>
>>>
>>>  standard python-setuptools from the distribution and I show that it's
>>>
>>>> running setuptools-0.6c11.
>>>> I've also attached the file of rejects listed above.
>>>>
>>>> Since you have a pretty sophisticated patch setup, I'm hoping that
>>>> you'll
>>>> be able to solve this pretty
>>>> easily.  Please let me know if there's any way I can help.
>>>>
>>>> Thanks for your support,
>>>>
>>>

Scott D. Anderson
Computer Science Department
Wellesley College
[EMAIL PROTECTED]
[EMAIL PROTECTED]

  Hi Scott,

I looked at the data you sent.

I don't have a Fedora 14 system with python 2.7 where I can reproduce this.
I suspect something else modified this file and thus the jcc patch doesn't
apply out of the box.

I can't guess what you have in that file so I don't know what to tell you
off hand. If you send me the original file, I can try to patch it manually.

Have you looked at it ? Does it look difficult for you to do ?

Andi..

On Tue, 24 Jan 2012, Scott D Anderson wrote:

> On Tue, Jan 24, 2012 at 4:37 PM, Andi Vajda <[EMAIL PROTECTED]> wrote:
>
>>
>> On Tue, 24 Jan 2012, Scott D Anderson wrote:
>>
>>  Sorry for the confusion.  I named it pylucene-install.log, but that's just
>>> a rename
>>> of the typescript file.
>>>
>>
>> Your messages have no attachments. Maybe they're being filtered out.
>> Can you include it inline since it's plain text anyway ?
>
>
>> Andi..
>
>
> Sure.  Here's the typescript:
>
> Script started on Tue 24 Jan 2012 01:07:10 PM EST
> [root@sampras tmp] cat /etc/issue
> Fedora release 14 (Laughlin)
> Kernel \r on an \m (\l)
>
> [root@sampras tmp] rpm -qV python-setuptools
> [root@sampras tmp] ls -l /usr/lib/python2.7/site-packages/setuptools*
> -rw-r--r-- 1 root root  144 Oct 26 22:05
> /usr/lib/python2.7/site-packages/setuptools-0.6c11-py2.7.egg-info
> -rw-r--r-- 1 root root   34 Oct 26 22:05
> /usr/lib/python2.7/site-packages/setuptools.pth
>
> /usr/lib/python2.7/site-packages/setuptools:
> total 300
> -rw-r--r-- 1 root root  5927 Oct 26 22:04 archive_util.py
> -rw-r--r-- 2 root root  5725 Oct 26 22:05 archive_util.pyc
> -rw-r--r-- 2 root root  5725 Oct 26 22:05 archive_util.pyo
> drwxr-xr-x 2 root root  4096 Jan 24 11:47 command
> -rw-r--r-- 1 root root  6209 Oct 26 22:04 depends.py
> -rw-r--r-- 2 root root  6749 Oct 26 22:05 depends.pyc
> -rw-r--r-- 2 root root  6749 Oct 26 22:05 depends.pyo
> -rw-r--r-- 1 root root 30309 Oct 26 22:04 dist.py
> -rw-r--r-- 1 root root 30221 Oct 26 22:05 dist.pyc
> -rw-r--r-- 1 root root 30163 Oct 26 22:05 dist.pyo
> -rw-r--r-- 1 root root  1278 Oct 26 22:04 extension.py
> -rw-r--r-- 2 root root  1761 Oct 26 22:05 extension.pyc
> -rw-r--r-- 2 root root  1761 Oct 26 22:05 extension.pyo
> -rw-r--r-- 1 root root  1278 Oct 26 22:04 extension.py.orig
> -rw-r--r-- 1 root root  3744 Oct 26 22:04 __init__.py
> -rw-r--r-- 2 root root  4028 Oct 26 22:05 __init__.pyc
> -rw-r--r-- 2 root root  4028 Oct 26 22:05 __init__.pyo
> -rw-r--r-- 1 root root 30447 Oct 26 22:04 package_index.py
> -rw-r--r-- 2 root root 28625 Oct 26 22:05 package_index.pyc
> -rw-r--r-- 2 root root 28625 Oct 26 22:05 package_index.pyo
> -rw-r--r-- 1 root root  8756 Oct 26 22:04 sandbox.py
> -rw-r--r-- 2 root root 11425 Oct 26 22:05 sandbox.pyc
> -rw-r--r-- 2 root root 11425 Oct 26 22:05 sandbox.pyo
> [root@sampras tmp] patch -d /usr/lib/python2.7/site-packages -Nup0 <
> /usr/network/tmp/pylucene-3.5.0-3 /jcc/jcc/patches/patch.43.0.6c11
> patching file setuptools/extension.py
> Hunk #1 succeeded at 32 (offset 4 lines).
> patching file setuptools/command/build_ext.py
> Hunk #1 FAILED at 85.
> Hunk #2 succeeded at 177 (offset 7 lines).
> Hunk #3 succeeded at 259 (offset 7 lines).
> 1 out of 3 hunks FAILED -- saving rejects to file
> setuptools/command/build_ext.py.rej
> [root@sampras tmp] exit
>
> Script done on Tue 24 Jan 2012 01:08:32 PM EST
>
> and here's the contents of build_ex.py.rej
>
> --- setuptools/command/build_ext.py    (revision 75864)
> +++ setuptools/command/build_ext.py    (working copy)
> @@ -85,8 +85,12 @@
>         if fullname in self.ext_map:
>             ext = self.ext_map[fullname]
>             if isinstance(ext,Library):
> +                if ext.force_shared and not use_stubs:
> +                    _libtype = 'shared'
> +                else:
> +                    _libtype = libtype
>                 fn, ext = os.path.splitext(filename)
> -                return self.shlib_compiler.library_filename(fn,libtype)
> +                return self.shlib_compiler.library_filename(fn,_libtype)

On Tue, Jan 24, 2012 at 10:33 PM, Andi Vajda <[EMAIL PROTECTED]> wrote:

>
>  Hi Scott,
>
> I looked at the data you sent.
>
> I don't have a Fedora 14 system with python 2.7 where I can reproduce this.
> I suspect something else modified this file and thus the jcc patch doesn't
> apply out of the box.
>

I can understand your suspicion, which is why I did the rpm -V to check
whether it's been modified, and it hasn't.
I even uninstalled and re-installed it, to be sure (that wasn't in the
typescript).  I even checked that it's coming from
the fedora repo, not some odd other repo. So unless something is very
strange, there are a lot of Fedora 14 files
out there that the patch doesn't work for.
>
> I can't guess what you have in that file so I don't know what to tell you
> off hand. If you send me the original file, I can try to patch it manually.
>

I've pasted it below my signature, since something seems to be stripping my
attachments.

>
> Have you looked at it ? Does it look difficult for you to do ?
I don't have a lot of experience with setuptools, so I'd be doing a lot of
poor guesswork.  If you'd be willing to take a crack at it, I'd be very
grateful.

Thanks,

Scott

# =============== build_ext.py =======from distutils.command.build_ext import build_ext as _du_build_ext
try:
    # Attempt to use Pyrex for building extensions, if available
    from Pyrex.Distutils.build_ext import build_ext as _build_ext
except ImportError:
    _build_ext = _du_build_ext

import os, sys
from distutils.file_util import copy_file
from setuptools.extension import Library
from distutils.ccompiler import new_compiler
from distutils.sysconfig import customize_compiler, get_config_var
get_config_var("LDSHARED")  # make sure _config_vars is initialized
from distutils.sysconfig import _config_vars
from distutils import log
from distutils.errors import *

have_rtld = False
use_stubs = False
libtype = 'shared'

if sys.platform == "darwin":
    use_stubs = True
elif os.name != 'nt':
    try:
        from dl import RTLD_NOW
        have_rtld = True
        use_stubs = True
    except ImportError:
        pass

def if_dl(s):
    if have_rtld:
        return s
    return ''
class build_ext(_build_ext):
    def run(self):
        """Build extensions in build directory, then copy if --inplace"""
        old_inplace, self.inplace = self.inplace, 0
        _build_ext.run(self)
        self.inplace = old_inplace
        if old_inplace:
            self.copy_extensions_to_source()

    def copy_extensions_to_source(self):
        build_py = self.get_finalized_command('build_py')
        for ext in self.extensions:
            fullname = self.get_ext_fullname(ext.name)
            filename = self.get_ext_filename(fullname)
            modpath = fullname.split('.')
            package = '.'.join(modpath[:-1])
            package_dir = build_py.get_package_dir(package)
            dest_filename os.path.join(package_dir,os.path.basename(filename))
            src_filename = os.path.join(self.build_lib,filename)

            # Always copy, even if source is older than destination, to
ensure
            # that the right extensions for the current Python/platform are
            # used.
            copy_file(
                src_filename, dest_filename, verbose=self.verbose,
                dry_run=self.dry_run
            )
            if ext._needs_stub:
                self.write_stub(package_dir or os.curdir, ext, True)
    if _build_ext is not _du_build_ext and not
hasattr(_build_ext,'pyrex_sources'):
        # Workaround for problems using some Pyrex versions w/SWIG and/or
2.4
        def swig_sources(self, sources, *otherargs):
            # first do any Pyrex processing
            sources = _build_ext.swig_sources(self, sources) or sources
            # Then do any actual SWIG stuff on the remainder
            return _du_build_ext.swig_sources(self, sources, *otherargs)

    def get_ext_filename(self, fullname):
        filename = _build_ext.get_ext_filename(self,fullname)
        if fullname not in self.ext_map:
            return filename
        ext = self.ext_map[fullname]
        if isinstance(ext,Library):
            fn, ext = os.path.splitext(filename)
            return self.shlib_compiler.library_filename(fn,libtype)
        elif use_stubs and ext._links_to_dynamic:
            d,fn = os.path.split(filename)
            return os.path.join(d,'dl-'+fn)
        else:
            return filename

    def initialize_options(self):
        _build_ext.initialize_options(self)
        self.shlib_compiler = None
        self.shlibs = []
        self.ext_map = {}

    def finalize_options(self):
        _build_ext.finalize_options(self)
        self.extensions = self.extensions or []
        self.check_extensions_list(self.extensions)
        self.shlibs = [ext for ext in self.extensions
                        if isinstance(ext,Library)]
        if self.shlibs:
            self.setup_shlib_compiler()
        for ext in self.extensions:
            ext._full_name = self.get_ext_fullname(ext.name)
        for ext in self.extensions:
            fullname = ext._full_name
            self.ext_map[fullname] = ext

            # distutils 3.1 will also ask for module names
            # XXX what to do with conflicts?
            self.ext_map[fullname.split('.')[-1]] = ext

            ltd = ext._links_to_dynamic = \
                self.shlibs and self.links_to_dynamic(ext) or False
            ext._needs_stub = ltd and use_stubs and not
isinstance(ext,Library)
            filename = ext._file_name = self.get_ext_filename(fullname)
            libdir = os.path.dirname(os.path.join(self.build_lib,filename))
            if ltd and libdir not in ext.library_dirs:
                ext.library_dirs.append(libdir)
            if ltd and use_stubs and os.curdir not in
ext.runtime_library_dirs:
                ext.runtime_library_dirs.append(os.curdir)

    def setup_shlib_compiler(self):
        compiler = self.shlib

  Hi Scott,

On Thu, 26 Jan 2012, Scott D Anderson wrote:

> Ah, my apologies. I didn't realize that the patch had half-succeeded.   I
> have reinstalled the python-setuptools package from the Fedora repo and
> I've verified that the package is unmodified.  I've attached the
> build_ext.py.orig file, which is identical to the build_ext.py file.
>
> Sorry for the double-attachment last time.  I was trying to get it to
> attach as plain text instead of octet stream, and then I couldn't delete
> the extra attachment.

This is what I was afraid of. The original setuptools sources on PyPI,
http://pypi.python.org/packages/source/s/setuptools/setuptools-0.6c11.tar.gz#md5=7df2a529a074f613b509fb44feefe74e
(which list the author, pje, as the file owner even), has a different
version of build_ext.py from what you sent me.

The diff is attached and included below.

Unless you can explain where your modified version is coming from and why, I
really don't see what I can do to 'properly' fix this.

The comment about distutils 3.1 in your version that doesn't exist in the
original seems to indicate that your version is newer than what it's
purported to be, 0.6c11.

If there is a newer version of setuptools out there that needs a new patch,
I can certainly write it but what is it and where is it coming from ?

Andi..

--- /Users/vajda/build_ext.py.orig 2012-01-27 15:13:36.000000000 -0800
+++ /Users/vajda/tmp/s.b/setuptools-0.6c11/setuptools/command/build_ext.py 2009-10-19 10:35:44.000000000 -0700
@@ -77,27 +77,25 @@
              sources = _build_ext.swig_sources(self, sources) or sources
              # Then do any actual SWIG stuff on the remainder
              return _du_build_ext.swig_sources(self, sources, *otherargs)

      def get_ext_filename(self, fullname):
          filename = _build_ext.get_ext_filename(self,fullname)
-        if fullname not in self.ext_map:
-            return filename
-        ext = self.ext_map[fullname]
-        if isinstance(ext,Library):
-            fn, ext = os.path.splitext(filename)
-            return self.shlib_compiler.library_filename(fn,libtype)
-        elif use_stubs and ext._links_to_dynamic:
-            d,fn = os.path.split(filename)
-            return os.path.join(d,'dl-'+fn)
-        else:
-            return filename
+        if fullname in self.ext_map:
+            ext = self.ext_map[fullname]
+            if isinstance(ext,Library):
+                fn, ext = os.path.splitext(filename)
+                return self.shlib_compiler.library_filename(fn,libtype)
+            elif use_stubs and ext._links_to_dynamic:
+                d,fn = os.path.split(filename)
+                return os.path.join(d,'dl-'+fn)
+        return filename

      def initialize_options(self):
          _build_ext.initialize_options(self)
          self.shlib_compiler = None
          self.shlibs = []
          self.ext_map = {}

      def finalize_options(self):
@@ -108,21 +106,16 @@
                          if isinstance(ext,Library)]
          if self.shlibs:
              self.setup_shlib_compiler()
          for ext in self.extensions:
              ext._full_name = self.get_ext_fullname(ext.name)
          for ext in self.extensions:
              fullname = ext._full_name
              self.ext_map[fullname] = ext
-
-            # distutils 3.1 will also ask for module names
-            # XXX what to do with conflicts?
-            self.ext_map[fullname.split('.')[-1]] = ext
-
              ltd = ext._links_to_dynamic = \
                  self.shlibs and self.links_to_dynamic(ext) or False
              ext._needs_stub = ltd and use_stubs and not isinstance(ext,Library)
              filename = ext._file_name = self.get_ext_filename(fullname)
              libdir = os.path.dirname(os.path.join(self.build_lib,filename))
              if ltd and libdir not in ext.library_dirs:
                  ext.library_dirs.append(libdir)
              if ltd and use_stubs and os.curdir not in ext.runtime_library_dirs:

On Fri, 27 Jan 2012, Scott D Anderson wrote:

> Well, my code would be coming from the RPMs in the Fedora 14 repository.
> Maybe tweaked by someone in the Fedora community?
>
> Hmm.  Here's a thought.  Would it make sense to use pip to re-install
> setuptools?  Presumably pip would look at the pypi.python.org place that
> you referenced, rather than the Fedora repository, and if it replaced the
> Fedora stuff, your original code might just work.  I don't know a lot about
> pip and setuptools, so I may be missing something.

To reinstall setuptools, just get and run the setuptools egg as outlined
here:
   http://pypi.python.org/pypi/setuptools/0.6c11#cygwin-mac-os-x-linux-other

Andi..

>
> Thanks,
>
> Scott
>
> On Fri, Jan 27, 2012 at 6:23 PM, Andi Vajda <[EMAIL PROTECTED]> wrote:
>
>>
>>  Hi Scott,
>>
>>
>> On Thu, 26 Jan 2012, Scott D Anderson wrote:
>>
>>  Ah, my apologies. I didn't realize that the patch had half-succeeded.   I
>>> have reinstalled the python-setuptools package from the Fedora repo and
>>> I've verified that the package is unmodified.  I've attached the
>>> build_ext.py.orig file, which is identical to the build_ext.py file.
>>>
>>> Sorry for the double-attachment last time.  I was trying to get it to
>>> attach as plain text instead of octet stream, and then I couldn't delete
>>> the extra attachment.
>>>
>>
>> This is what I was afraid of. The original setuptools sources on PyPI,
>> http://pypi.python.org/**packages/source/s/setuptools/**
>> setuptools-0.6c11.tar.gz#md5=**7df2a529a074f613b509fb44feefe7**4e<http://pypi.python.org/packages/source/s/setuptools/setuptools-0.6c11.tar.gz#md5=7df2a529a074f613b509fb44feefe74e>
>> (which list the author, pje, as the file owner even), has a different
>> version of build_ext.py from what you sent me.
>>
>> The diff is attached and included below.
>>
>> Unless you can explain where your modified version is coming from and why,
>> I really don't see what I can do to 'properly' fix this.
>>
>> The comment about distutils 3.1 in your version that doesn't exist in the
>> original seems to indicate that your version is newer than what it's
>> purported to be, 0.6c11.
>>
>> If there is a newer version of setuptools out there that needs a new
>> patch, I can certainly write it but what is it and where is it coming from ?
>>
>> Andi..
>>
>> --- /Users/vajda/build_ext.py.orig      2012-01-27 15:13:36.000000000
>> -0800
>> +++ /Users/vajda/tmp/s.b/**setuptools-0.6c11/setuptools/**command/build_ext.py
>>      2009-10-19 10:35:44.000000000 -0700
>> @@ -77,27 +77,25 @@
>>
>>             sources = _build_ext.swig_sources(self, sources) or sources
>>             # Then do any actual SWIG stuff on the remainder
>>             return _du_build_ext.swig_sources(**self, sources, *otherargs)
>>
>>
>>
>>     def get_ext_filename(self, fullname):
>>         filename = _build_ext.get_ext_filename(**self,fullname)
>> -        if fullname not in self.ext_map:
>> -            return filename
>> -        ext = self.ext_map[fullname]
>> -        if isinstance(ext,Library):
>> -            fn, ext = os.path.splitext(filename)
>> -            return self.shlib_compiler.library_**filename(fn,libtype)
>> -        elif use_stubs and ext._links_to_dynamic:
>> -            d,fn = os.path.split(filename)
>> -            return os.path.join(d,'dl-'+fn)
>> -        else:
>> -            return filename
>> +        if fullname in self.ext_map:
>> +            ext = self.ext_map[fullname]
>> +            if isinstance(ext,Library):
>> +                fn, ext = os.path.splitext(filename)
>> +                return self.shlib_compiler.library_**filename(fn,libtype)
>> +            elif use_stubs and ext._links_to_dynamic:
>> +                d,fn = os.path.split(filename)
>> +                return os.path.join(d,'dl-'+fn)
>> +        return filename
>>
>>
>>     def initialize_options(self):
>>         _build_ext.initialize_options(**self)
>>         self.shlib_compiler = None
>>         self.shlibs = []