[parted-devel] tests/t3310-flags failure

Brian C. Lane bcl at redhat.com
Wed Jun 26 20:17:34 BST 2019


On Wed, May 22, 2019 at 09:17:26AM -0700, Brian C. Lane wrote:
> On Wed, May 22, 2019 at 09:03:15AM -0400, Phillip Susi wrote:
> > Brian C. Lane writes:
> > 
> > > On Mon, May 20, 2019 at 03:30:28PM -0400, Phillip Susi wrote:
> > >> 
> > >> Brian C. Lane writes:
> > >> > I haven't seen this fail recently, the last fedora build passed just
> > >> > fine. But I don't have the patch swapping those, and it doesn't look
> > >> > like it made it to master yet.
> > >> 
> > >> Oops... I hadn't pushed it yet.  Here it is.
> > >
> > > Ah, right I'd meant to try that one out. Just applied it and sure enough
> > > t3310 fails for me as well.
> > 
> > Right... any idea why?  When I run those commands by hand they work just
> > fine.
> 
> Weird. I threw a print into the test and it says it is a msdos label
> with no partitions:
> 
> parted -s dev-file u s p
> Model:  (file)
> Disk /builddir/build/BUILD/parted-3.2/tests/gt-t3310-flags.sh.D1cY/dev-file: 8192s
> Sector size (logical/physical): 1024B/1024B
> Partition Table: msdos
> Disk Flags: 
> Number  Start  End  Size  Type  File system  Flags
> 
> So it isn't just a failure of the print-flags helper. Currently I
> don't have much time to dig any deeper, maybe by the end of the week.

I took another look at this again today, and managed to get even more
confused. If I reproduce the test commands from the cmdline, after
setting LD_LIBRARY_PATH to point to the new build, and setting
PARTED_SECTOR_SIZE=1024 it works :/ But it still fails in the test
script, and the output from a parted u s p on the failing file is still
empty.

So I don't have any new ideas, and really have no idea what's going on
with this seemingly simple change.

-- 
Brian C. Lane (PST8PDT) - weldr.io - lorax - parted



More information about the parted-devel mailing list