Re: [gamin] greedy gam_server

On Sunday 05 November 2006 02:50, Daniel Veillard wrote:
> > Well, there don't appear to be any developers around....
>   I posted a new release 0.1.8 no later than 5 days ago

Yep, I saw that.  And none of the fixes appeared to address the following
bug reports:

Fri, 3 Nov 2006 23:42:34 +0000 (GMT)   julie guneratne ntlworld com
Wed, 11 Oct 2006 17:31:56 -0700 "Patrick Lin" <Patrick Lin hyperpia com>
Tue, 3 Oct 2006 13:34:46 -0700 Dan <dmisc charter net>
Wed, 27 Sep 2006 09:14:51 -0400 Barry Rountree <rountree cs uga edu>
Thu, 21 Sep 2006 11:44:31 -0300 "Walter Cruz" <walter php gmail com>

I thought you were using this list in write-only mode.  I'm happy to see I 
was mistaken.

> I can't comment on problems I have no idea how to reproduce 

In my email on 27Sep2006 I provided my distro version, gamin version, 
symptoms of the problem, an analysis of the debugging messages, 
an .xsession-errors dump, and what I did to cause and stop the problem.  
Please let me know what additional information you need. 

> I could explain why gamin sucks, it's based on what it was asked to do
> and what is available to implement it, I don't think anybody care, 

Well, I care, actually, and a number of other people seem to as well.  Could
you provide a writeup and put it in the faq on the gamin website?

> it's 
> just simpler to complain, right ?

The folks who emailed went pretty far out of their way to let you know 
that there was as issue.  None of them were complaining -- they were
asking for help.  

And with that out of the way....

I turned on debugging and let it run for a good long time.  Obviously,
that generates millions of messages.  Far and away the greatest number
of them were:

   6534 Event to kded [kdeinit] : 127, 1, .xsession-errors Changed

That's interesting....  Let's have a look at .xsession-errors

   rountree malaise:~$ ls -l .xsession-errors
   -rw------- 1 rountree rountree 262071980 2006-09-27 08:48 .xsession-errors

Ok, that's huge.  What's in there?

   end from FAM server connection
   invalid length 24902
   invalid length 24902
   invalid length 24902
   invalid length 24902

Over and over and over again.

This starts to happen sometime after I start up Konqueror (although based on
what I've googled, it appears that many kde applications can trigger it) and
stops when I kill konqueror.  

It's possible that konqueror is generating .xsession-errors input at such a 
high rate that gamin ends up using a lot of cpu just trying to keep up.  
However, since I'm not seeing a huge number of xsession errors when gamin 
isn't running, this probably isn't the case.

I'm more inclined to think that some event causes gamin to error out, the 
error gets logged in .xsession-errors, thus gamin detects that the file has 
changed, but this detection (or handling it) generates the same error, and 
that error is logged to .xsession-errors, thus the file is changed, noted, 
another error generated -- gamin ends up chasing its own tail and chewing
up a lot of cpu in the process.

The really simple and stupid workaround could be something as easy as 
making the default to be ignoring any .xsession-errors file.  

Shall I send you a patch?



> Daniel

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]