Alexandre Goujon ale.goujon@gmail.com writes:
Should fix bug #17037 I followed your advice and now, I only ask the mountmgr when needed.
dlls/kernel32/volume.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-)
This is failing here when the drive is empty:
../../../tools/runtest -q -P wine -M kernel32.dll -T ../../.. -p kernel32_test.exe.so volume.c && touch volume.ok volume.c:132: Tests skipped: can't test removing fake drive volume.c:785: Test failed: IOCTL_DVD_READ_STRUCTURE (DvdPhysicalDescriptor) failed, last error = 87 make: *** [volume.ok] Error 1
If I put a DVD in I get:
volume.c:813: Test failed: IOCTL_DVD_READ_STRUCTURE should have failed
Probably the tests need some fixing.
On 08/29/2012 09:35 PM, Alexandre Julliard wrote:
Alexandre Goujon ale.goujon@gmail.com writes:
Should fix bug #17037 I followed your advice and now, I only ask the mountmgr when needed.
dlls/kernel32/volume.c | 10 +++++++++- 1 file changed, 9 insertions(+), 1 deletion(-)
This is failing here when the drive is empty:
../../../tools/runtest -q -P wine -M kernel32.dll -T ../../.. -p kernel32_test.exe.so volume.c && touch volume.ok volume.c:132: Tests skipped: can't test removing fake drive volume.c:785: Test failed: IOCTL_DVD_READ_STRUCTURE (DvdPhysicalDescriptor) failed, last error = 87 make: *** [volume.ok] Error 1
If I put a DVD in I get:
volume.c:813: Test failed: IOCTL_DVD_READ_STRUCTURE should have failed
Probably the tests need some fixing.
I worked on these issues and found that : + "make volume.ok" is different than "wine kernel32_test.exe.so volume" (WINETEST_PLATFORM=wine is set in the first command) + the error
"volume.c:813: Test failed: IOCTL_DVD_READ_STRUCTURE should have failed"
comes from a missing check. I fixed it with http://source.winehq.org/patches/data/89457. + the error
"volume.c:785: Test failed: IOCTL_DVD_READ_STRUCTURE (DvdPhysicalDescriptor) failed, last error = 87"
comes from ioctl() failing (errno = ENOMEDIUM), returning STATUS_INVALID_PARAMETER hence last error = 87 (ERROR_INVALID_PARAMETER). I fixed that with http://source.winehq.org/patches/data/89458.
Except the fact that I wasn't testing my patch the right way (I'm not very proud of it), this also proved the TestBot is not very efficient regarding volume tests. So, I am proposing to add several virtual drives to some virtual machines. Is it possible ? I can work on it if needed. With genisomage and mkudffs, it's very easy to generate ISO files of different kinds. It won't be the same as with real discs and real drives but at least, it would prevent some regressions.. Any comment on this ?
On Fri, 31 Aug 2012, GOUJON Alexandre wrote: [...]
Except the fact that I wasn't testing my patch the right way (I'm not very proud of it), this also proved the TestBot is not very efficient regarding volume tests. So, I am proposing to add several virtual drives to some virtual machines. Is it possible ? I can work on it if needed. With genisomage and mkudffs, it's very easy to generate ISO files of different kinds. It won't be the same as with real discs and real drives but at least, it would prevent some regressions.. Any comment on this ?
The existing WineTestBot infrastructure is going to be replaced soonish... hopefully. So it's unlikely that we'll change the VMs now.
However we can certainly add some drives or even ISOs as we build the VMs of the new WineTestBot. You'd need to detail which configurations you're interested in.
One thing to keep in mind is we should strive to test as much as possible in as few configurations as possible, especially for the base VMs (those that run every single test patch). But at the same time it should be possible to isolate the effect of one configuration change.
I'm pretty sure others have things they'd like to see in the new WineTestBot. Where can we keep track of those?
Should we add a WineTestBot product to Bugzilla? Should the WineHQ.org product be coopted?
On Fri, Aug 31, 2012 at 9:59 AM, Francois Gouget fgouget@codeweavers.com wrote:
... I'm pretty sure others have things they'd like to see in the new WineTestBot. Where can we keep track of those?
If you're collecting a wish list then multiple network interfaces would be awesome, though I recognize that there is only a small number of tests that would be affected by it.
Erich Hoover ehoover@mines.edu
-
Alexandre Julliard -
Erich E. Hoover -
Francois Gouget -
GOUJON Alexandre