Hi Mark,
At 3:06 PM -0400 7/29/09, Mark wrote:
Even when I corrected the bad from lines, the popper would still
give the from lines error until I removed both X-UID lines. Then it
popped no issue.
The error was that Qpopper was unable to parse the spool because it
couldn't find the "from " separator lines?
As far as the X-UID lines, I am certain popper puts them in there
when people set their client to leave mail on server.
If people leave mail on server and writing status/uidl hasn't been
disabled, then Qpopper writes a "Status:" header line. This is done
in popper/pop_updt.c lines 557-564 (an existing "Status:" header is
updated to indicate a message has been downloaded at lines 586-588).
So the difference I saw was the different type of X-UID written.
4.015 writes X-UIDL and 4.1 writes X-UID
Qpopper writes an "X-UIDL:" header line if writing status/uidl hasn't
been disabled. It does this in popper/pop_updt.c lines 539-545. It
doesn't write "X-UID" lines, at least not intentionally.
Writing status/x-uidl headers is enabled by default, but can be
disabled by adding "set update-status-headers = false" in a
configuration file.
I do not want to run traces until we agree on popper writing X-UID
lines :-) because I feel that is where the problem lies
A trace can help show what is going on. Qpopper doesn't have any
code to generate or write a line such as "X-UID:3" so either there is
a weird bug in Qpopper or some other process is writing it.
Are you using server mode? Do you set any of the options which
affect cache files?
One difference between 4.0 and 4.1 is that 4.1 uses the cache file
even if new mail has arrived. This speeds up session start-up when
users leave mail on the server, since only the new mail is read from
the spool. However, if another process has modified the spool,
corruption can occur. Try setting the "conservative-cache" option,
which tells Qpopper not to use the cache if the spool doesn't exactly
match the cache. To do this, add "set conservative-cache = true" to
the configuration file. If you have a situation where some users
might use IMAP as well as POP, while others use only POP, you can set
the "conservative-cache" option for only those users who might use
IMAP. This will speed things up for the POP-only users if they leave
mail on the server. (Or, you can set the option in a global
configuration file and then set it to false for POP-only users.)
At 01:29 PM 7/27/2009, Randall Gellens wrote:
At 8:16 AM -0400 7/24/09, Mark wrote:
I installed the 4.1b17 and the problems between IMAP and qpopper
were magnified by a factor of 5. Using the 4.0xx series of
qpopper I would have an issue once in awhile and easily fix the
From lines being broken. After getting a rash of support calls
and my support staff having a mini melt down, I restored the
4.014 qpopper I was running and everything went back to normal.
Editing mail spools to fix the broken From lines I noticed two
X-UID lines in the header which I did not see before.
For example:
normal everyday X-UIDL line
X-UIDL: c...@+"!-$`"!^gF"!^3f"!
and
X-UID:3
I assumed 4.b17 put the X-UID:3 in there and it was conflicting
with the old type of X-UIDL line because even fixing the From
lines people still could not pop those spools until I removed
BOTH X-UID lines
Any ideas?
Thanks
Mark
Hi Mark,
Qpopper doesn't write "X-UID" lines. Maybe the IMAP server is
doing this? Even so, I'm not sure why there would be a difference
between 4.0 and 4.1.
When you say users couldn't pop their mail until you deleted the
X--UID and X-UIDL lines, what was happening?
Would it be possible for you to reproduce the problem with Qpopper tracing?
To enable tracing in Qpopper in standalone mode:
1. Do a 'make clean'
2. Re-run ./configure, adding '--enable-debugging'.
4. Add '-t <tracefile-path>' to the command line for Qpopper.
This causes detailed tracing to be written to the file specified
as 'tracefile-path'.
At 07:00 PM 7/23/2009, Randall Gellens wrote:
At 3:57 PM -0400 7/22/09, Mark wrote:
Does 4.1x write X-UID differently then 4.0x?
No, but there are several options which control if and how
X-UID is handled.
Are you seeing a problem with X-UIDL?
--
Randall Gellens
Opinions are personal; facts are suspect; I speak for myself only
-------------- Randomly selected tag: ---------------
You have the right to remain silent. Anything you say will be
misquoted, then used against you.
--
Randall Gellens
Opinions are personal; facts are suspect; I speak for myself only
-------------- Randomly selected tag: ---------------
Flappity, Floppity, Flip!
The Mouse on the Möbius Strip.
The Strip revolved,
The Mouse dissolved
In a chronodimensional skip.
--Frederick Winsor, "The Space Child's Mother Goose"
[Simon & Schuster 1966], paperback
