mozilla launching script proposed xremote fix
by Warren Togami
Below describes two broken xremote functions that our current
/usr/bin/mozilla script relies upon, and a method to fix the script for
equivalent functionality so that it works for mozilla-1.6+ and
MozillaFirebird-0.8+ while Thunderbird is running. If you are running
an earlier mozilla or Firebird, then you must run
/usr/lib/thunderbird/mozilla-xremote-client rather than the browser's
own client or the browser will fail.
Please help to verify these findings and suggested patch to mozilla-1.6.
Verify that there are no regressions in behavior, and it behaves
equivalently to before the patch both with and without
MozillaThunderbird-0.4 running.
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=114524
Tracking this proposal within this report
Christopher Blizzard wrote:
> Warren Togami wrote:
>> Blizzard,
>>
>> Can we improve the /usr/bin/mozilla script to make it possible to
>> launch while Thunderbird is running? It is currently impossible to do
>> so with the current /usr/bin/mozilla script. This may severely hinder
>> adoption of Thunderbird for Linux in the future.
>>
>> Do you already have a recipe to make this possible for
>> /usr/bin/mozilla?
>
> If we're talking about the problem that I think we're talking about,
> Mozilla 1.6 should fix it. The client in 1.6 (along with thunderbird
> 0.4, I believe) include the code changes required to allow the client
> to keep looking for a server that will support its request.
>
> --Chris
>
I did testing with mozilla-1.6-0 and fedora.us MozillaThunderbird-0.4,
and it appears that xremote behaviors are broken in several ways that
still make the /usr/bin/mozilla script fail while Thunderbird is running
with this thunderbird pop-up dialog:
Alert
Error launching browser window:TypeError:
Components.classes['@mozilla.org/appshell/component/browser/instance;1']
has no properties
There is a slight improvement over mozilla-1.4.1 and MozillaFirebird-0.7
though, in that you no longer need to run thunderbird's xremote client
rather than the browser's xremote
in order to send the browser a signal. This improvement is
unfortunately unrelated to the remaining xremote brokenness, which
currently make it impossible for the /usr/bin/mozilla script of 1.6 to
launch while Thunderbird-0.4 is running.
http://www.mozilla.org/unix/remote.html
Of the xremote commands listed on this page, these two commands in
mozilla-1.6 seem to be broken while Thunderbird is running:
ping()
xfeDoCommand (openBrowser)
function check_running() {
$MOZ_CLIENT_PROGRAM 'ping()' 2>/dev/null >/dev/null
RETURN_VAL=$?
if [ "$RETURN_VAL" -eq "2" ]; then
echo 0
return 0
else
echo 1
return 1
fi
}
xremote's ping() functionality never works properly while Thunderbird is
running. It normally returns "2" if Mozilla is not running, or "0" if
it is running. While Thunderbird is running, ping() always returns "0".
This leads the /usr/bin/mozilla script to think it needs to use
xremote rather than launch mozilla.
# check to see if there's an already running instance or not
ALREADY_RUNNING=`check_running`
# If there is no command line argument at all then try to open a new
# window in an already running instance.
if [ "${ALREADY_RUNNING}" -eq "1" ] && [ -z "$1" ]; then
exec $MOZ_CLIENT_PROGRAM "xfeDoCommand(openBrowser)" 2>/dev/null
>/dev/null
fi
The xremote xfeDoCommand(openBrowser) option is what causes the pop-up
error dialog from Thunderbird mentioned near the top of this report. It
seems that xfeDoCommand(openBrowser) is completely inoperable with
either Thunderbird's or mozilla-1.6's xremote client binary.
Since ping() is broken, we no longer have a functional way in which to
test for a running browser. What other distributions have done to
workaround this problem are shell snippets like this:
$MOZ_CLIENT_PROGRAM "openURL($1,new-window)" && exit 0
# If the script reaches this line, then assume xremote failed
# launch mozilla directly
Unfortunately I have found that some programs like xchat call the
/usr/bin/mozilla script directly with parameters like "-remote
openURL(%s,new-window)", so the above snippet fails in that case. Below
is a minimal patch to /usr/bin/mozilla that should work.
--- mozilla.orig 2004-01-21 06:50:01.633665318 -1000
+++ mozilla 2004-01-22 00:08:39.492016912 -1000
@@ -164,14 +164,18 @@
# set our JVM vars
set_jvm_vars
-# check to see if there's an already running instance or not
-ALREADY_RUNNING=`check_running`
-
-# If there is no command line argument at all then try to open a new
-# window in an already running instance.
-if [ "${ALREADY_RUNNING}" -eq "1" ] && [ -z "$1" ]; then
- exec $MOZ_CLIENT_PROGRAM "xfeDoCommand(openBrowser)" 2>/dev/null
>/dev/null
+if [ "$1" == "-remote" ]; then
+ $MOZ_CLIENT_PROGRAM "$2" && exit 0
+else
+ URL="$1"
+ if [ -z $1 ]; then
+ URL="about:blank"
+ fi
+ $MOZ_CLIENT_PROGRAM "openURL($1,new-window)" && exit 0
fi
+ALREADY_RUNNING=0
+# xremote failed, indicating that a browser is not already running
+# attempt to launch new browser
# check system locale
MOZARGS=
In my testing this patch behaves identically to the previous
/usr/bin/mozilla script, with the exception of fixing this
launching/xremote behavior of course.
Your thoughts?
Warren Togami
wtogami(a)redhat.com
20 years, 4 months
Re: I have a problem with the package mount but mount is nowhere to be found on bugzilla
by lupus
k thx I will use rpm -qfi foo instead to see from which component it is
part. BTW is there a way to see all bugreports in rawhide from all
packages. Or the last 50 or so? :) Just trying to make my bug reports
better.
On do, 2004-01-29 at 00:36 +0000, Pedro Morais wrote:
> Em Quinta 29 Janeiro 2004 00:36, Kristof vansant escreveu:
> > Mount is not on the component list on bugzilla, is there no support for
> > this component or???
> >
> > BTW can it be that because of the .sel behind the version dependencies
> > of other packages can't correctly detect that I have a higher mount
> > version then required?
> >
> > [root@d5e03fba lupus]# rpm -q mount
> > mount-2.11y-35.sel
>
> rpm -qif $(which mount)
> Look for the Source RPM: value.
> (in this case, util-linux)
>
> --
> Pedro Morais - morais(a)kde.org - http://www.rnl.ist.utl.pt/~pmmm/
--
lupus (Kristof Vansant Belgium)
20 years, 4 months
weird fc devel upgrade dependencie problem mount and libgtop seem to be missing, but they are installed
by lupus
Does anyone know what can be the cause?
[root@d5e03fba lupus]# yum upgrade
Gathering header information file(s) from server(s)
Server: Fedora Core 1 - i386 - Base
Server: Fedora Core 1 - i386 - Unreleased Updates
Server: Fedora.us Extras (Stable)
Server: Fedora.us Extras (Testing)
Server: Fedora.us Extras (Unstable)
Server: Livna.org - Fedora Compatible Packages (stable)
Server: Livna.org - Fedora Compatible Packages (testing)
Server: Livna.org - Fedora Compatible Packages (unstable)
Server: macromedia.mplug.org - Flash Plugin
Server: Fedora Core 1 - i386 - Released Updates
Finding updated packages
Downloading needed headers
Finding obsoleted packages
Resolving dependencies
.Package gnome-system-monitor needs libgtop-2.0.so.0, this is not
available.
Package gnome-applets needs libgtop-2.0.so.0, this is not available.
Package gnome-system-monitor needs libgtop_common-2.0.so.0, this is not
available.
Package gnome-applets needs libgtop_common-2.0.so.0, this is not
available.
Package gnome-system-monitor needs libgtop_sysdeps-2.0.so.0, this is not
available.
Package gnome-applets needs libgtop_sysdeps-2.0.so.0, this is not
available.
Package initscripts needs mount >= 2.11l, this is not available.
Package mkinitrd needs mount, this is not available.
[root@d5e03fba lupus]# rpm -q libgtop2
libgtop2-2.0.3-1
[root@d5e03fba lupus]# rpm -q mount
mount-2.11y-35.sel
20 years, 4 months
Updating RPMs using binary deltas (demo)
by M A Young
Here is a couple of demo short scripts for generating and applying deltas
between two rpms. The key step is to use the minigzip program from the
zlib source (see the docs directory of the zlib-devel package), as
minigzip uses compression compatible with the compression rpm uses. If I
was writing this properly, I would consider using python, which seems to
have a zlib modules, and could also hook into some rpm libraries.
Michael Young
20 years, 4 months
Sql
by Chadley Wilson
Hi everyone,
I cant seem to find anyone who can help
I am trying to get our in house app which connects to an SQL2000-ent
sever to run in linux, I have the app running just fine but I cant get
connection to the database,
When you install it in windoze all you have to install is client
connectivity, I have browsed the win dirs on my XP system but there are
no files installed. but it works.
Cant I use the linux sql to some how get the link.
The other Problem I think I am faced with is the fact that linux doesn't
log into the domain. Is it possible to logon to the windoze domain with
Fedora?
The I also need to map the directory to the U drive in my wine folder.
Please if you have any Ideas I will be throwing a party on your behalf.
A BIG BIG thanks for any help in advance,
You folks are great.
Chad
--
Chadley Wilson <chadley(a)pinteq.co.za>
20 years, 4 months
Misc other rawhide issues
by Warren Togami
Please let me know if the below are already known bugs so I wont report
duplicates.
openoffice.org
BTW, GREAT JOB on the gtk2 native widgets. The entire suite looks GORGEOUS!
1) Spreadsheet user interface bug
Enter a value within a cell and hit ENTER. It does not go to the next
cell below like openoffice and staroffice before the gtk2-native widgets
conversion.
2) Opening a spreadsheet document and closing always segfaults. Let me
know if you cannot trigger this behavior and I will send the reproducer.
3) Japanese translations in the UI seem missing. Are all tranlations
missing?
Kernel
4) cd /proc; find; causes kernel oops. According to notting this is a
ACPI bug that is triggerable on Thinkpad hardware. Anyone know a
Bugzilla #?
tora
5) Lacking plugins to connect to any database? I might just totally
misunderstand how it works...
6) Segfault when closing
mozilla
7) Missing japanese translation. missing all translations?
8) ALL KDE APPS missing Japanese translation
<warren> than: Is it a known bug that all KDE apps in rawhide appear to
be missing japanese (all?) translations?
<than> warren: yes, KDE developer has removed kde-i18-japanese in KDE 3.2:(
<warren> than: yikes, any explanation?
<than> i just added the old one from Beta2 in our rpm
<warren> than: I can get the fedora-ja team to work on the translations
<than> warren: i think the Japanese translation is incompletely at the
moment
<mharris> warren: be careful.... I remember redhat-config-network's
"Python Network Configuration Tool" being translated into German which
more or less read "Tool for Configuration of Python Networks"
<mharris> ;o)
<than> warren: it's the reason why they have removed it
<than> warren: it's great if you could get fedora-ja team to work on the
translation!
Apparently both Japanese and Korean translations are judged to be so
poor and incomplete that KDE upstream has decided to remove them from
KDE 3.2. For now than has agreed to keep the old set of translations in
the distribution until the translation freeze on March 5th, after which
they will need to be removed from FC2 unless the community has fixed the
translations to a sufficient degree.
http://i18n.kde.org/contact.html
KDE i18n contact
http://fedora.jp/
I am alerting the already organized and enthused Fedora-JA community,
but I don't know any Korean Fedora users or groups, so please help if
you care about keeping the Korean i18n for KDE.
Misc processes need kill -9
9) Has anyone noticed that many miscellaneous processes when they become
either deadlocked or livelocked are unkillable with regular "kill", and
you need "kill -9"? This happened to me with:
gdm-binary, kinput2, soffice, top, and several other processes...
Only managed to strace one of them here:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=104296
gdm
10) gdm sometimes does not startup X-and-login-screen again
automatically after killed with CTRL-ALT-Backspace. Difficult to
reproduce but potentially bad.
11)
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=110315
This bug still exists in rawhide where gdm's maximum simultaneous
connection counter became a maximum connection counter between RH9->
FC1. It is particularly crippling for K12LTSP and my LUG's
super-Linux-advocacy presentation at a huge educator's conference coming
on Feburary 3rd. Help!!! =(
Warren Togami
wtogami(a)redhat.com
20 years, 4 months
Kernel Panic with 2.6.1-1.61
by Lorenzo Luconi Trombacchi
Hi,
I have just installed Kernel 2.6.1-1.61 x86_64 (CPU Athlon64 Mainboard
Asus K8V), but
immediately after the kernel starts the load a Kernel Panic message is
displayed.
The Kernel 2.6.1-1.53 works fine.
Lorenzo
20 years, 4 months
Re: sylpheed (and inclusions)
by Pete Zaitcev
On Wed, 28 Jan 2004 03:32:09 -0500
Bill Nottingham <notting(a)redhat.com> wrote:
> IMO, things like sylpheed, balsa, etc. all eventually should
> probably be in Extras. But I may be in the minority. :)
Right. Might even give in impetus to have Extras functioning properly.
As a sylpheed user, I do not mind it going to Extras.
-- Pete
20 years, 4 months