Re: Q: Did you do something to message number recently?

[ebiederm@xmission.com (Eric W. Biederman)]

Eric Wong <e@80x24.org> writes:

> "Eric W. Biederman" <ebiederm@xmission.com> wrote:
>> The add method computes the number using num_for, which uses
>> Msgmpa::mid_insert.
>> 
>> Short of the sequence number for msgmap getting scrambled I don't
>> see how that can go wrong.  Sigh.
>
> That seems to be what went wrong with "INSERT OR IGNORE"
>
> cf. https://public-inbox.org/meta/20190624233809.1721-1-e@80x24.org/

Is it INSERT OR IGNORE or is it the autoincrement?  As I read
the sqlite docs autoincrement is allowed to skip message numbers.
Especially in cases where the insert fails.

This fix should come with a test case shouldn't it?  Such a test
case should be pretty straight forward to write.  Insert a message
multiple times and then insert a new one.

I don't recall the transaction constraints on sqlite but I am wondering
if instead of INSERT OR IGNORE we should read and the increment
num_highwater in the public-inbox code and then we would not need to
depend upon autoincrement only incrementing once.  Even when the sqlite
docs don't promise it will.

>> I will have to instrument my client, and see if that will give
>> me a clue what is going on.
>
> I guess your client ends up seeing the same message
> over-and-over again like -watch does?

It usually doesn't but for dealing with a different imap server I
tweaked the code and one of the servers feeding me the last message in
the mail box every time I checked the mailbox for new messages.

I have also narrowed it down to duplication insertions.  I was just
about to start looking at what the fix would be.

>> I think most of my previous work was all in the index rebuild path.
>> Which may be why we don't have a regression test catching this.
>
> Yeah, this isn't a new bug at all.  I didn't notice since I
> mainly use w3m or (neo)mutt for NNTP and didn't look too closely
> at NNTP article numbers.  lynx makes it apparent messages are
> missing: lynx news.public-inbox.org/inbox.comp.lib.glibc.alpha

It is a little obscure in gnus but I start getting suspcious when
I tell it to give me the last 10 messages and I get only one.

Eric