Ruediger Pluem wrote:
On 11/19/2007 12:38 AM, Bojan Smojver wrote:
Both apr-0.9.17 and apr-1.2.12 fail SHM tests:
0.9.17:
-------------------------
starting consumer.....
starting producer.....
Name-based shared memory test FAILED: [2] No such file or directory
Name-based shared memory test FAILED: [2] No such file or directory
That's a bug in the test. Just ensure that the test subdirectory is in your PATH
such that testshmconsumer and testshmcontributor can be executed.
Yes; see the recent changes in Makefile.win to see where we set up PATH= to the
corresponding binary subdirectories. Of course if you cross compile, these
simply can't work.
Given that Ruediger already committed patches for both of these, it
would probably be best to roll again.
apr-util 1.2.12 is on it's way. I have one build failure as I had jumped to an
older nmake+older cmd.exe, my path-hack to work out the users' APR_PREFIX etc on
win32 doesn't behave as it did on server 2003+modern nmake, failing to then run
the test/ Makefile.win build properly. Grumpf. Should be an hour or few to get
that right (started it back in Atlanta), thanks for the loan of the wireless usb
dongle, Issac).
I haven't commited for testdate so far. I just wanted to hear some other
opinions
especially on the patches to apr_date_parse_rfc as I couldn't believe that these
bugs have not been found in a function that has not been touched for such a
long time.
So please comment and I commit or adjust.
Amazing how long a bug can live unnoticed, eh?
