Mark,

I found the immediate preceding file and removed it but still got the error. So I added some verbosity to mixrbld and reran it with this output:

/esnet/setup/island.setup/sysfiles/mixrbld .
./.mix470175e2
UID: 2
UID: 3
UID: 4
UID: 5
UID: 6
UID: 7
UID: 8
UID: 9
UID: 10
UID: 11
UID: 12
UID: 13
UID: 14
UID: 15
UID: 16
UID: 17
UID: 18
UID: 19
UID: 20
UID: 21
UID: 22
UID: 23
UID: 24
UID: 25
UID: 26
UID: 27
UID: 28
UID: 29
UID: 30
UID: 31
UID: 32
UID: 33
UID: 34
./.mix4702870a
UID: 140
./.mix47028bdd
UID: 1
Data file ./.mix47028bdd duplicate UID @ 0: 1

So since there were only two files before the complaining file I moved them around and determined that the first file is somehow in conflict with the third file. I processed that one in a separate directory.

Also, about a dozen lines down from line 177 the original printf format problem exists again. There may be others as well. Seems to be where cur->uid is printed.

Thoughts?
thanks,
David

ps I think I'm going to need a couple of Long Island Iced Teas personally ;-)

Mark Crispin wrote:
I think that I understand what is going on. You have two data files that start with UID 1, so the duplication is inter-file instead of intra-file. mixdfix only remedies the intra-file case.

Look in that directory. The other file is almost certainly the file that immediately preceeds .mix47028bdd (and .mix47028bdd is probably the second file in the overall list).

Decide which of these two files that you want, and delete the other. Then re-run mixrbld.

If for some reason you need both, let me know and we'll proceed from there.

How you happened to have multiple files with duplicate UIDs is probably best told over a pitcher of beer... ;-)

On Fri, 9 Nov 2007, David Severance wrote:

Mark,

I applied the patch and as you indicated got this error:

/usr/local/sbin/mixrbld SAVE
Data file SAVE/.mix47028bdd duplicate UID @ 0: 1

So I ran mixdfix and got this message:

/usr/local/sbin/mixdfix SAVE/.mix47028bdd
SAVE/.mix47028bdd: no repair needed

This was after I copied over .mixstatus and .mixmeta files from a newly created folder.

Ideas?
thanks,
David

Mark Crispin wrote:
The following patch will fix the core dump. At line 177 of mixrbld.c, change: printf ("Data file %s duplicate UID @ %ld: %ls\n",file,pos,cur->uid);
to:
printf ("Data file %s duplicate UID @ %ld: %ld\n",file,pos,cur->uid);
                                                 ^^^ note change here
In other words, that %ls should be %ld.

However, that is not the end of your problem, since mixrbld will now complain about the data file. You'll need mixdfix to repair that. I'm surprised that you say that mixdfix didn't complain about that data file, since it's supposed to check for that condition.

For the .mixmeta and .mixstatus files, just create an new mailbox and copy the files from that.

On Fri, 9 Nov 2007, David Severance wrote:

I have some .mix data files and no .mixindex, .mixmeta or .mixstatus files through a file system manipulation error. I figured I could just run mixrbld to take the data files and rebuild the other files. To my horror mixrbld core dumps. I have a core file which compresses to 12K which I can send along separately. It's been 10 years since I've had to use gdb and I don't remember how to be honest. I am using the 2006k final release with all the latest copies of the mix tools recompiled against the final 2006k release. Besides the core dump problem is there a document on how to deal with rebuilding or recovering mix folders. Also, mixdfix indicated no problems with the data files.

thanks,
David

--
David Severance
Network and Academic Computing Services
(949) 824-7552
[EMAIL PROTECTED]

_______________________________________________
Imap-uw mailing list
Imap-uw@u.washington.edu
https://mailman1.u.washington.edu/mailman/listinfo/imap-uw


-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.

--
David Severance
Network and Academic Computing Services
(949) 824-7552
[EMAIL PROTECTED]



-- Mark --

http://staff.washington.edu/mrc
Science does not emerge from voting, party politics, or public debate.
Si vis pacem, para bellum.

--
David Severance
Network and Academic Computing Services
(949) 824-7552
[EMAIL PROTECTED]

_______________________________________________
Imap-uw mailing list
Imap-uw@u.washington.edu
https://mailman1.u.washington.edu/mailman/listinfo/imap-uw

Reply via email to