[oe] mtn: fatal: St11logic_error: database.cc:845: invariant 'I(cancel_size < pending_writes_size)' violated

Koen Kooi koen at dominion.kabel.utwente.nl
Sun Sep 10 12:39:31 UTC 2006


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

[11:02] koen: mtn: fatal: St11logic_error: database.cc:845: invariant 'I(cancel_size <
pending_writes_size)' violated
[11:02] • koen wonders
[11:05] njs: what version are you using?
[11:05] koen: monotone 0.29 (base revision: 796bc8596d6c5c68828fbfb6a620a09a3af15ded)
[11:06] koen: I'm suddenly getting 131 revs in
[11:06] njs: oh, dude, 0.29 has that bug?
[11:06] koen: and the invariant trips on 35/131
[11:06] njs: huh, I guess it does, just events conspire to make it really hard to hit
[11:06] njs: it's a very stupid bug, that "<" should be "<="
[11:07] njs: but I didn't realize it was possible to trigger it in 0.29 (I ran into it
while doing the roster cache stuff, because it changed the data flow patterns so the bug
got hit every time)

So, these are your options:

* downgrade to 0.28
* upgrade to head
* rebuild with 'I(cancel_size <= pending_writes_size)' instead of 'I(cancel_size <
pending_writes_size)'
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (Darwin)

iD8DBQFFBAeDMkyGM64RGpERAlBfAJ92FLBJyyq8WzuHdo6MqXc1Qxy1QgCfWCsB
VOl2KmZt3Za2meZzJUEeo8A=
=XFbh
-----END PGP SIGNATURE-----




More information about the Openembedded-devel mailing list