[Monotone-devel] Re: New strange monotone errors

[graydon hoare]

Christof Petig wrote:
> Christof Petig schrieb:
>
> > Well, the netsync performed just fine this morning (though I did get 
> > an error in the end, see failure.gz):

your failure, for what it's worth, helped me track down a broken format 
string in netsync.cc. some other people have already run into it (and 
I've put in machinery to make it easier to isolate such things in the 
future).

the upshot was that during netsync, you wrote some manifest certs to 
disk but not the accompanying manifests (because monotone crashed during 
formatting of an internal log message). as a result, you had a cert 
without a manifest, and it complained when you tried to update to the 
manifest (located by examining certs).

I should make the update algorithm a bit more tolerant about this. in 
general, it is always possible to have a cert referring to a 
not-yet-existant version. perhaps it should issue a warning if it 
selects one as an update target, but then it should back off and 
continue updating, not crash.

> monotone fetch did the trick - all works well again.

yeah, you got the manifest from the fetch. this is another thing which 
should be fixed: when your db contains a cert which refers to a 
nonexistent manifest, netsync should always request it in pull/sync 
modes, to try to "fill in the gaps". currently the manifest request 
cycle is only triggered by the arrival of *new* certs, and since you 
already had the cert in question, further netsyncs would not have fixed it.

> PS: What's wrong with the mailing list? I do not get my posts back 
> (neither do I see any other).

it appears to have a 3-4 day delay on postings, it has been like this 
all last week too. perhaps something is wrong with their mailing list 
software. I have received no reply posting to the savannah hackers list 
about it.. but then, they probably have such a delay too :(

-graydon