[Xprint] xprint 009 and firefox 0.8

[Matthew Hooper]

Sorry about the delay in responding Roland, but I've been following a 
few related threads to do with the Mozilla 1.7rc1 browser and chasing up 
some Solaris patching.

To answer your questions:

    * No we don't have the required patches for the latest firefox
      release - though we did when 0.8 was installed.
    * The load issues I'm referring to seem to be related to a
      long-standing Mozilla/firefox bug regarding browser inactivity and
      browser responsiveness, which from following the thread for the
      Bugzilla entry has only recently been addressed and should be
      included in the 0.9 release of firefox.
    * We're running SunRay software, so I have added a line exporting
      XSUNTRANSPORT="shmem" to see what effect that will have once the
      students are back from holiday, and the load is back at "normal"
      levels.

We're holding off on putting any contrib builds into production until 
after we've done some testing on the various solaris tarballs, but I'm 
hopeful that once we've configured the new build of Xprint + firefox 0.9 
when it's released, that things will become more stable.

Thanks,

    Matt.


Roland Mainz wrote:

> Matthew Hooper wrote:
>
>> With your build of Firefox for Solaris, is it built with support for
>> multiple processors?
>
>
> I didn't alter anything special like threading support - everying which
> was used to build FireFox is described in the README and
> about:buildconfig.
> Mozilla/FireFox/ThunderBird are running the GUI, rendering and event
> queue processing in the same thread and other threads are only used to
> offload tasks like networking support etc. (basically a
> "single-threaded"-like design which uses other threads for
> "helper"-purposes) ...
>
> ... but before I ask more let's check the most common source for
> problems with Mozilla/FireFox/ThunderBird - missing OS patches:
> Do you have all Solaris/SPARC patches listed in
> http://lxr.mozilla.org/seamonkey/source/xpfe/bootstrap/init.d/moz_patch_checker.dtksh#53 
>
> installed in your system ?
>
>> We've been having some load issues
>
>
> Can you be little bit more specific what happens ? Does it help when you
> use pbind(1m) to force one FireFox process to use exactly only ONE CPU ?
>
>> on our server which
>> has 4 processors and 4 Gig of memory with around 60+ firefox sessions,
>
>
> Are you running the SunRay software ?
> If "yes" - does it help when you start FireFox like this
> % (export XSUNTRANSPORT="blablabla" ; ./firefox)
>
>> and
>> I was wondering if the -02 option
>
>
> "-O2" ? I guess you mean "-xO2", right ? Mozilla.org default is -"xO4"
> but I have lowered that to "-xO2" to avoid bugs in the Sun
> Workshop/Forte compilers... I am trading 5%-25% performance reduction
> for stability here since I prefer stabilty little bit over performance
> (crashing binaries aren't good - nor are random bugs caused by compiler
> bugs) ... :)
>
>> might make some difference in regard to
>> this as well - processors aside?
>
>
> Mozilla and FireFox are generally resource hogs (both CPU and memory)
> and there cannot be done much about that except hoping that the code
> becomes better over the time... ;-(
> BTW: There are nightly contrib builds for FireFox available under
> ftp://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/contrib/latest-trunk/ 
>
> ... maybe they work better for you...
>
> ----
>
> Bye,
> Roland
>

-- 

Matthew Hooper
Library Systems Officer
Flinders University Library
G.P.O. Box 2100
ADELAIDE, South Australia 5001
t  +618 8201 2068
f  +618 8201 2508
http://www.lib.flinders.edu.au/

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mozdev.org/pipermail/xprint/attachments/20040427/5d785bde/attachment.htm