Re: [Evolution] Problems in CVS - Known?

[Lonnie Borntreger <email borntreger com>]

Some followup on the issues....

On Fri, 2002-03-22 at 18:58, Lonnie Borntreger wrote:
    1) Evolution has become MUCH less stable.  It crashes/freezes quite
    regularly this last week or two (usually within 30 minutes).

This is related to issue 3, it seems.  Whenever Evolution seems slow or
acting like it's going to crash, the executive-summary is eating a CPU.

    2) The INBOX message count in the "Shortcuts" section always stays at
    the initial number of new messages from when evolution was started, and
    the vfolders count in shortcuts never shows.  This started two or three
    days ago.

As reported by Jeff, this is now fixed.

    3) The executive summary process is a CPU HOG.  This has been the case
    for the last several weeks, although it's getting common (used to only
    happen occasionally, now it happens all the time).  This is what I see
    when only having run evolution for around 17-18 minutes:
    >From TOP:
    24476 67goat    17   0 10592  10M  7132 R    99.8  2.0  16:06
    evolution-execu
    
    notice that it is using 99.8% of CPU.  Luckily I have two of them, so I
    can continue to work, and the other threads can continue (other than the
    aformentioned freezes).  Notice the CPU time allocated to the process
    after 17-18 minutes.
    >From PS:
    24476 ?        R     16:39 evolution-executive-summary 
    
It looks like it is in a hard loop.  Here is a snapshot of what was
shown when I attached to the pid using strace.  This series of output
repeats ad infinitum, with only the second number in the braces of
"gettimeofday" changing.
NOTE: everything after the "poll" is one line
NOTE1: I can recompile with -g if it helps... I normally don't.

gettimeofday({1016933112, 457019}, NULL) = 0
read(33, "", 8192)                      = 0
read(32, "", 8192)                      = 0
read(28, "", 8192)                      = 0
gettimeofday({1016933112, 457128}, NULL) = 0
ioctl(6, 0x541b, [0])                   = 0
poll([{fd=6, events=POLLIN}, {fd=14, events=POLLIN}, {fd=28,
events=0}, {fd=32, events=0}, {fd=33, events=0}, {fd=33,
events=POLLIN, revents=POLLIN}, {fd=33, events=0}, {fd=32,
events=POLLIN, revents=POLLIN}, {fd=32, events=0}, {fd=28,
events=POLLIN, revents=POLLIN}, {fd=28, events=0}, {fd=10,
events=POLLIN}, {fd=11, events=POLLIN}, {fd=12,
events=POLLIN}, {fd=13, events=POLLIN}, {fd=16,
events=POLLIN}, {fd=17, events=POLLIN}, {fd=18,
events=POLLIN}, {fd=19, events=POLLIN}, {fd=20,
events=POLLIN, revents=POLLIN}, {fd=21, events=POLLIN},
{fd=29, events=POLLIN}], 22, 7364) = 4


    4) If I happen to type into the To: field a little too fast, the To:/Cc:
    pane goes blank, and I have to killev to recover. (this "may" have been
    fixed by the last CVS update, I'll watch it)

This is not as straight forward as I thought.  It doesn't always happen,
and it's hard to see how to reproduce it with regularity.  It just
happens when it feels like it I guess.....


-- 
TTFN,
Lonnie Borntreger