[Imap-uw] mixcvt giving me an incinsistency message size error
nlin
nlin at newton.berkeley.edu
Fri May 4 15:34:08 PDT 2007
Hi Mark
I also just had a thought. If the problem is in the imap toolkit, does
this mean that mailutil would not have worked properly as well? So that if
I converted a user's mailbox from either unix format or mbx format to the
mixfmt using pre 1/18/07 version, it would have corrected the mailbox?
thanks
nancy
-------------------------------------
Nancy Lin DECF
1109A Etcheverry Hall 510-642-7291
Office Hours: 2PM-4PM Mon-Thu
-------------------------------------
Mark Crispin wrote:
> Hi Nancy -
>
> You misunderstood. The bug fixed on January 18, 2007 was in the mix
> driver in the IMAP toolkit, not in mixcvt.
>
> mixcvt is simply reporting that it encountered a mailbox that was
> impacted by that old, now fixed, bug.
>
> On Tue, 1 May 2007, nlin wrote:
>
>> Hi Mark
>>
>> That's kind of strange. I only download mixcvt on 4/17 and the
>> mixcvt.c file has the following line:
>>
>> Last Edited: 14 April 2007
>>
>> I also just downloaded and compiled the Apr 23rd version and that one
>> gave me the same error.
>>
>> These mailboxes are big. That's why I wanted to convert them. One is
>> about 500MB and the other 625MB. I'm currently running the 2006f
>> version of imapd. Do I need to upgrade 2006g?
>>
>> nancy
>>
>> -------------------------------------
>> Nancy Lin DECF
>> 1109A Etcheverry Hall 510-642-7291
>> Office Hours: 2PM-4PM Mon-Thu
>> -------------------------------------
>>
>>
>> Mark Crispin wrote:
>>> On Tue, 1 May 2007, nlin wrote:
>>>> warning: Inconsistency in mix message size, uid=0000150d (20987904
>>>> != 2098899
>>>
>>> This message indicates a repair of a damage caused by a known bug in
>>> earlier versions of the mix driver. I think that you omitted a digit
>>> from the end of the second value. Assuming that to be the case, the
>>> message is stored in the index as having 20,987,904 bytes (exactly
>>> 20496K) whereas the message was supposed to have 20,988,99x bytes.
>>> Thus, between 1086 and 1095 bytes from the end of the message were lost.
>>>
>>> The bug was one in which a final buffer's worth of bytes would not be
>>> written to the mailbox on append if the number of iterations through
>>> the message copy loop (64K at a time) exceeded the number of bytes
>>> remaining to be written to the message. The larger the message, the
>>> more likely this bug was to bite. It was fixed on January 18, 2007.
>>>
>>>> Is this similar to the data length error that was posted on the
>>>> mailing list the other day?
>>>
>>> The "data length error" problem was a bug in the traditional UNIX
>>> format driver. That too is fixed.
>
> -- Mark --
>
> http://panda.com/mrc
> Democracy is two wolves and a sheep deciding what to eat for lunch.
> Liberty is a well-armed sheep contesting the vote.
>
More information about the Imap-uw
mailing list