subversion-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Bert Huijben" <b...@qqmail.nl>
Subject RE: copy_test, update crash and dirent_uri failure on Windows
Date Wed, 06 Jan 2010 08:54:38 GMT


> -----Original Message-----
> From: Ed [mailto:ed@kdtc.net]
> Sent: woensdag 6 januari 2010 7:21
> To: Subversion Dev
> Subject: copy_test, update crash and dirent_uri failure on Windows
> 
> Hi,
> 
> It's been some time since I last compiled svn on my Windows box; but,
> as far as I can recall, I've had this problem before.  I'm running
> VS2008 on a Vista machine and compiling the trunk (checked out today
> as of r896326.
> 
> During the copy_tests (specifically #7), I would get a dialog which
> states:
> 
>     DEBUG ERROR
> 
>     Program:
>     d:\other_projs\svn\trunk\Debug\subversion\svn\svn.exe

^^^^^
> 
>     The application has requested the Runtime to terminate it
>     in an unusual way.
> 
> Ditto with update_tests.py but it gave me an AppCrash dialog:
> 
>   Problem Event Name:    APPCRASH
>    Application Name:     svn.exe
>    Application Version:  1.7.0.0
>    Application Timestamp:        4b441a79
>    Fault Module Name:    libsvn_subr-1.dll
>    Fault Module Version: 1.7.0.0
>    Fault Module Timestamp:       4b441a11
>    Exception Code:       c0000005
>    Exception Offset:     0008cdd7
>    OS Version:   6.0.6001.2.1.0.256.6
>    Locale ID:    1033
>    Additional Information 1:     22eb
>    Additional Information 2:     40dd979078fb82d9393b5a88c8b4a483
>    Additional Information 3:     cb58
>    Additional Information 4:     0e665ec469d0f64cf5e3772ec0a4460f
> 
> Copy_tests.py crash log = svn_crash1.txt
> Update_tests.py crash log = svn_crash2.txt
> 
> with both crashes, I can close the program and yet the tests
> continue with an eventual 'success'.
> 
> Also, dirent_uri_test # 37 is failing on Windows (at least,
> on my box) with the tests.log dump as follows:
> 
> .\..\..\subversion\tests\libsvn_subr\dirent_uri-test.c:2295:
> (apr_err=125001)
> ..\..\..\subversion\libsvn_subr\dirent_uri.c:1553: (apr_err=125001)
> svn_tests: Couldn't determine absolute path of 'C:abc'
> ..\..\..\subversion\libsvn_subr\dirent_uri.c:1553: (apr_err=20022)
> svn_tests: The given path is incomplete
> XFAIL: dirent_uri-test.exe 36: test svn_dirent_get_absolute with lc drive
> ..\..\..\subversion\tests\libsvn_subr\dirent_uri-test.c:2378:
> (apr_err=125001)
> ..\..\..\subversion\libsvn_subr\dirent_uri.c:1812: (apr_err=125001)
> ..\..\..\subversion\libsvn_subr\dirent_uri.c:1553: (apr_err=125001)
> svn_tests: Couldn't determine absolute path of 'C:A'
> ..\..\..\subversion\libsvn_subr\dirent_uri.c:1553: (apr_err=20022)
> svn_tests: The given path is incomplete
> FAIL:  dirent_uri-test.exe 37: test svn_dirent_condense_targets

This is a known issue in APR. Apr doesn't like when any of the active drives
has a lower case drive letter in its path.
E.g. see  http://svn.haxx.se/dev/archive-2009-10/0682.shtml
and http://svn.haxx.se/dev/archive-2009-11/0075.shtml

Most Windows applications never create lower case drive letters in the
current directories. E.g

You have to do special operations to make it happen. Not just switching
drives.
E.g.
C:\> d:
D:\> cd /d c:\windows
c:\windows>

Here you have a lower case drive letter in the current directory. And if you
would run the tests from this situation you will see the test failure.

In most cases I heard of this happens because the user uses an alternative
shell instead of cmd.exe like Cygwin.

The fix should be implemented in APR and before that you can work around
this issue by fixing the drive letter casing via the same trick.

> 
> I included test # 36 because it's an XFAIL and as far as I can
> see the results of the errors are nearly the same as #37 with
> the minor exception of the first error being in dirent_uri-test.c:2295
> for #36 and dirent_uri-test.c:2378 for #37.  Since I
> don't see anyone talking about it here, I'm assuming it's
> a fault with my box.

That XFail test you quoted tests for this exact issue by replicating the
behavior from the C code.

	Bert
> 
> Can someone clarify whether this is the case of faulty
> compiling on my part?
> 
> Thanks.
> 
> Edmund


Mime
View raw message