On Sun, Dec 13, 2009 at 5:46 PM, Roger Leigh <[email protected]> wrote: > On Sun, Dec 13, 2009 at 09:36:18PM +0000, Roger Leigh wrote: >> On Sun, Dec 13, 2009 at 04:12:37PM -0500, Evan Broder wrote: >> > On Sun, Dec 13, 2009 at 2:33 PM, Roger Leigh <[email protected]> wrote: >> > > [Please could you not top-post. Thanks!] >> > > >> > > On Sun, Dec 13, 2009 at 01:57:15PM -0500, Evan Broder wrote: >> > >> I checked out the repository, ran ./bootstrap, then debuild -S, then >> > >> sbuild, then installed the resulting packages. I've rebooted since >> > >> installing the packages just to make sure there's no state sticking >> > >> around. >> > > >> > > I've put a copy of the sources and debian packages here: >> > > http://www.codelibre.net/~rleigh/schroot/ >> > > >> > > Could you try them, or rebuild them as appropriate? >> > > >> > > I've verified they work for me. However, it looked from your >> > > description that you might additionally be using filesystem >> > > unions; is this the case? If so, please could you test without >> > > a union to see if this works? >> > > >> > > If it's broken only when unions are enabled, there's some >> > > issue writing out the session file when unions are enabled. >> > >> > I've tried with both your packages and mine - I get the same behavior, >> > but you're right that it's only for union chroots. Without unions, >> > schroot -b works on type=block-device chroots. >> > >> > However, union chroots of type=directory /do/ work - it's just a >> > problem for chroots that are both union chroots and type=block-device. >> >> Thanks for checking. I'll have to have a further look to see what's >> different between the two types. > > It's actually a simple bug in the union code for block-device chroots. > If enabled, it immediately returned, skipping the session file > creation. The attached patch should amend this. Could you please > give it a try and let me know if this fixes things up for you? >
Yep - that works for me on block-device union chroots. Thanks, - Evan -- To UNSUBSCRIBE, email to [email protected] with a subject of "unsubscribe". Trouble? Contact [email protected]
