It'd be better if it did.

doesn't it ? it works here

Well anyway, some sample code:

import zipfile

z = zipfile.ZipFile('aaa.zip', mode='w')
z.writestr('aa.py', 'def x(): print "hi there"\n\ndef y(): print "hi"')
z.close()


and then:


import sys
import inspect

sys.path.append('aaa.zip')
import aa

inspect.getsource(aa.x)
inspect.getsource(aa.y)


Doesn't that work for you ?

It seems to depend on working directory:

exarkun@charm:/tmp$ python
Python 2.5.2 (r252:60911, Jul 31 2008, 17:28:52) 
[GCC 4.2.3 (Ubuntu 4.2.3-2ubuntu7)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.path.append('/home/exarkun/foobar.zip')
>>> import foobar, inspect
>>> inspect.getsource(foobar.foo)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python2.5/inspect.py", line 629, in getsource
    lines, lnum = getsourcelines(object)
  File "/usr/lib/python2.5/inspect.py", line 618, in getsourcelines
    lines, lnum = findsource(object)
  File "/usr/lib/python2.5/inspect.py", line 468, in findsource
    raise IOError('could not get source code')
IOError: could not get source code
>>> 

versus:

exarkun@charm:~$ python
Python 2.5.2 (r252:60911, Jul 31 2008, 17:28:52) 
[GCC 4.2.3 (Ubuntu 4.2.3-2ubuntu7)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.path.append('/home/exarkun/foobar.zip')
>>> import foobar, inspect
>>> inspect.getsource(foobar.foo)
'def foo():\n    pass\n'
>>>

This must be very similar to #4197 and #4201 that I reported a few days 
ago.  I'll see if a similar patch would work in this case.

Hmm, apparently inspect was made to work with zipped modules back in r45248 .

I cannot reproduce the problem either.

Jean-Paul, can you attach your foobar.zip?  What else do you have in your 
/tmp directory?

In any case, it's clearly not a candidate for 2.5.3, due to the lack of
a clear problem description, and a working patch. That essentially means
that associating the version 2.5 is also fairly pointless, but I'll
leave that in, anyway.

Here is a transcript for a complete, self-contained, minimal reproduction:

exarkun@charm:~$ cd /tmp
exarkun@charm:/tmp$ mkdir zipimportbugexample
exarkun@charm:/tmp$ cd zipimportbugexample/
exarkun@charm:/tmp/zipimportbugexample$ mkdir foobar
exarkun@charm:/tmp/zipimportbugexample$ echo "def foo(): pass" >
foobar/__init__.py
exarkun@charm:/tmp/zipimportbugexample$ zip foobar.zip foobar
  adding: foobar/ (stored 0%)
exarkun@charm:/tmp/zipimportbugexample$ zip foobar.zip foobar/__init__.py 
  adding: foobar/__init__.py (stored 0%)
exarkun@charm:/tmp/zipimportbugexample$ rm -r foobar
exarkun@charm:/tmp/zipimportbugexample$ mkdir workingdirectory
exarkun@charm:/tmp/zipimportbugexample$ cd workingdirectory/
exarkun@charm:/tmp/zipimportbugexample/workingdirectory$ python
Python 2.5.2 (r252:60911, Jul 31 2008, 17:28:52) 
[GCC 4.2.3 (Ubuntu 4.2.3-2ubuntu7)] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import sys
>>> sys.path.append('/tmp/zipimportbugexample/foobar.zip')
>>> import foobar, inspect
>>> inspect.getsource(foobar.foo)
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/usr/lib/python2.5/inspect.py", line 629, in getsource
    lines, lnum = getsourcelines(object)
  File "/usr/lib/python2.5/inspect.py", line 618, in getsourcelines
    lines, lnum = findsource(object)
  File "/usr/lib/python2.5/inspect.py", line 468, in findsource
    raise IOError('could not get source code')
IOError: could not get source code
>>>

Thanks, Jean-Paul, I can now reproduce your problem and I think I found 
the culprit.

I believe the file name check in linecache.updatecache is unnecessary 
and is responsible for this problem.  With the following patch:

--- Lib/linecache.py	(revision 67040)
+++ Lib/linecache.py	(working copy)
@@ -88,7 +88,7 @@
             get_source = getattr(loader, 'get_source', None)
 
             if name and get_source:
-                if basename.startswith(name.split('.')[-1]+'.'):
+#                if basename.startswith(name.split('.')[-1]+'.'):
                     try:
                         data = get_source(name)
                     except (ImportError, IOError):


and Jean-Paul's foobar:

$ PYTHONPATH=/tmp/foobar.zip ./python.exe  -c "import inspect,foobar; 
print inspect.getsource(foobar)"
def foo(): pass


Martin, is there a chance for this issue and  #4197, #4201  to make it 
to 2.5.3?  If so, I'll prepare a combined patch with tests for your 
consideration shortly.

After looking into this, I think Alexander is correct. There is no
standard mapping between __file__ and __name__ and linecache is mistaken
in assuming such a mapping exists for all importers (and is the same as
the standard filesystem to name mapping to boot).

In this particular case, it was the differences between the way the two
relate for a package vs a normal module that was confusing linecache,
but for more exotic cases the filesystem based rules that linecache was
attempting to enforce may be completely irrelevant.

Fixed for 2.7 in r67750.

Will be ported to 2.6, 3.0 and 3.1.

Final revisions for fix:
2.7 = r67751 (there was a new test file missing from the initial checkin)
2.6 = r67752
3.1 = r67753
3.0 = r67754