mingw test failures

[Martin Hauner <ha...@web.de>]

Hi,

I'm currently trying to build subversion with mingw. After starting with a lot 
of test failures (I think it were 249 test failures or something like that) I'm 
down to 5 test failures after a few minimal changes and acouple of rebuilds:

At least one test FAILED, checking c:\Development\mingw\svn-1.6.x\tests.log
FAIL:  dirent_uri-test.exe 18: test svn_dirent_get_absolute
FAIL:  commit_tests.py 22: commit from a dir with a longer name than the wc
FAIL:  update_tests.py 27: conflict markers should match the file's eol style
FAIL:  prop_tests.py 13: some svn: properties should be converted
FAIL:  merge_tests.py 40: conflict markers should match the file's eol style
Summary of test results:
   43 tests SKIPPED
   25 tests XFAILED
   5 tests FAILED


The dirent_uri-test.exe fails in test_dirent_get_absolute { "C:abc", "@/abc" }.
But this is a binary where I can simply run it with gdb.


My question is how do I debug svn inside the python tests?

I tried to replace the svn binary (svn_binary in svntest/main.py) with a script 
that calls gdb. But that doesn't seem to work. It just stops and I don't get a 
gdb prompt although gdb is running (I can see it in the task mananger).



-- 
Martin

Subcommander 2.0.0 Beta 5 - http://subcommander.tigris.org
a Win32/Unix/MacOSX subversion GUI client & diff/merge tool.

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2406798

Re: mingw test failures

[Julian Foad <ju...@btopenworld.com>]

On Mon, 2009-10-12, Martin Hauner wrote:
> My question is how do I debug svn inside the python tests?
> 
> I tried to replace the svn binary (svn_binary in svntest/main.py) with a script 
> that calls gdb. But that doesn't seem to work. It just stops and I don't get a 
> gdb prompt although gdb is running (I can see it in the task mananger).

Martin,

What I usually do is edit the Python test, insert a "raise" statement
just before the line that runs the "svn" command that fails, then re-run
the test with "--no-cleanup" mode (so it does all the preparation and
then stops), and then manually run the required next command with GDB
(copying and pasting the parameters from the output of the test run that
failed).

It's not really neat, I know, but it works.

(In your attempt, I guess GDB's stdin/stdout were not connected to the
console.)

- Julian

------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=2407175