The bridge failed last night on cindi's wad.  Internally what happened 
was that she did a putback on 11/3 that the bridge failed on.  Danek 
backed it out due to merge turds, and it was re-putback on 11/4.

The original putback was what got pushed to the Hg onnv-gate, as 
revision 3100.  I've had to rollback && revert this, and I've re-bridged 
her follow-on putback on 11/4 as revision 3100.

So: anyone who has taken a child/cloned/updated their repositories are 
now detached.  Please do a 'hg rollback && revert' or when you do your 
next update, you will probably grow a second head.

Unfortunately, while I was debugging this; I noticed a discrepency in 
usr/src/lib/fm/topo/libtopo/common/mod.c, specifically:

[stevel at donuthole:onnv_external] 499$ diff 
usr/src/lib/fm/topo/libtopo/common/mod.c 
/ws/onnv-clone/usr/src/lib/fm/topo/libtopo/common/mod.c
27c27
< #pragma ident "%Z%%M% %I%     %E% SMI"
---
 > #pragma ident "@(#)mod.c      1.2     06/02/11 SMI"
144a145
 >       (void) close(fd);
149a151,152
 >       if (fd >= 0)
 >               (void) close(fd);


It looks like the bridge failed somewhere in the past... most likely on 
revision 1417 (the FMA amd64 putback).  I think at this point I will 
probably need to regenerate the onnv-gate, rolling back to at least 
revision 1417 (if not altogether).

I realise this causes an inconvenience - but hey... it's beta, right? 
;-)  (Sorry, that's a crappy excuse, but an excuse nonetheless).  I 
don't know if I'll have time to do this this week since I'm OOTO; but if 
not, then I will have it done the following week.

And on a side note, I think a good sanity check to put in the bridge is 
to have it diff the files against /ws/onnv-gate after each putback to 
make sure there are no discrepencies.

cheers,
steve
-- 
stephen lau // stevel at sun.com | 650.786.0845 | http://whacked.net
opensolaris // solaris kernel development

Reply via email to