note to self....
Monday, April 24, 2006
memory issues
Why does kamix take more memory than Emacs? I mean, for that matter, why does Emacs take more memory than Firefox? Still, wtF? Fully graphical web browser - Mozilla Firefox 1.5.0.2 is listed in top as taking 0.5% of memory (we're talking 512M here, I believe) - and that's running maximized and with the Performancing blogging extension open.
GNU Emacs, on the other hand - a text editor with a text inerface, but running under X (so I guess there are a few widgets allocated there) - Emacs is taking up (again: according to top) 3.7% of RAM, with no files open, and no child processes running (?).
Now, this is nothing, really (3.7%) as we look up the list from bottom to top - in the same vicinity as Emacs on this listing (sorted by %MEM) we have such notable applications and services as kalarmd (3.8%), gkrellm (3.3%), and (less than 3%) xterm, artsd, etc. No, this is nothing - we're in the lightweight end of the listing - let's look closer to the top.
Let's see ... kdeinit (4.1%), kdeinit (4.2%), kdeinit (4.5%) ... the list goes on - altogether there are something like 16 instances of kdeinit running, each of them taking up more RAM individually than either Emacs or Firefox - more, in fact than any other of the 15-odd applications running except for e.g. X (7.9%), kgpg (6.3%), kamix (5.4%) ... whoops - wait, I've overlooked something: firefox-bin is running up at the top, consuming some 12.7% - followed by the #1 instance of kdeinit (10.5%) second, and X itself (7.9%) third....
So please ignore what I said about Firefox at the beginning - that was 'firefox' the script that starts 'firefox-bin' - everything else was [i think] pretty accurate.
Labels: kde, linux, memory
catastrophic Firefox 1.5.0.2 (Linux) failure - near miss
Apparently the
blogger.com | "Publish Status" | "Publishing in progress..." throbber page is causing memory leakage and ultimately crashage in Mozilla Firefox 1.5.0.2 (Linux) - I caught it when the CPU load was up to 37%, and the available RAM was down to around 20M (these numbers according to Gkrellm) - I killed it (successfully) before it locked up X, but I am becoming more convinced that it is page content like this that is crashing the bitch...
Labels: 1.5, firefox, linux, mozilla
And furthermore, why is the sound Notification system in KDE such a
piece of crap?
First it was Konsole -
SCREAMING at me each time I touched the Backspace key while the cursor was at the beginning of the line -WtF? Now it's Konqueror
BLASTING the same damned noise at me each time I push the down arrow key to scroll down the page!?
And on
my system, each of those
staticy,
noisy, piece of shit system beeps is about
100Db - that is, I have this PoS notebook plugged into a rather high-powered speaker system (not that the crappy system beep isn't
too loud even on crappy little laptop speakers), and no matter what I do, it seems that KDE just won't agree to either
shut up with the stupid system beeps, or even just turn them down to something a lot lower in amplitude than what I want from e.g. XMMS or JuK or whatever media player I'm using to blast rock'n'roll to the rest of the neighborhood...
Put it on the to do list to fix this crap...
Labels: kde, keyboard, konsole, linux, sound
Intrusive Firefox Extensions - and other stuff
Fukking Klipper! Goddess save us all, the Advertisers have learned what a browser extension is. Can the spyware spooks and the virus kiddies be far behind?
A little while ago, as Firefox 1.5.0.2 (Linux) was crashing spectacularly on one workstation (vertigo) - and this time w/o the help of any extensions or themes, i might add - I was browsing the
prefs.js file for another Firefox install on another workstation (spirit).
The first thing I noticed about the install on spirit (and this would have been true for the vertigo install until I blew away the user profile directory in an effort to get Firefox to quit locking up the X server) is that there are an entire assload of settings that are added by extensions, and (in one case) by something that is not even an extension (I think - still looking at this one).
The next thing I noticed was that certain of the settings contained information about me - or rather, about some things that were on my disc drive.
Some examples - not all of these are "intrusive", I suppose, but most of them are remarkable (in some way), to me:
- user_pref("extensions.lastAppVersion", "1.0");
- This one looks to me like an artifact of a prgramming error, probably in an extension. I mean, 'lastAppVersion' of *what*?
- browser.download.dir
- Well, this one seems to be part of the browser, but how much could an untrusted extension - or even an un-noticed Javascript - learn about the directory structure of the local drive from watching this and, say, browser.download.lastDir ?
- browser.startup.homepage_override.mstone, rv:1.7.12
- What is this? I didn't see it listed - maybe I need to look again...
- extensions.mediaplayerconnectivity.*
- Well, these seem to be heirarchaly correct, but - there are a few like '.playerpls, "/usr/bin/xmms" that could conceivably leak system information to other processes that have access to this file, or even to those who released mediaplayerconnectivity
- fgupdater.*
- The fgupdater creators didn't bother with the 'extensions' prefix ... neither did the flashgot guys, or the gmnotifier guys, or the menux guys - perhaps this is a point of disagreement amongst developers? Or maybe it's already been solved, and something here is obsolete?. More of these found are sessionsaver.*, stumble.*
- gm-notifier.users.default
- ... has my username for the service - and menux.editor.path has '/usr/bin/gedit' - need to change that, since I'm not sure this system even has gedit - isn't that some Gnomish editor thing?
- print.tmp.printerfeatures.printer.*
- A bunch of these - odd, considering this workstation has never known a printer.
- security.*
- Seems like a poor place for things like .warn_viewing_mixed setting - couldn't another app [extension] change the value from, say, "true" to "false"
- stumble.784508.interests
- user_pref("stumble.784508.last_incat", "0");
- user_pref("stumble.784508.last_stumble", "1133658391177");
- user_pref("stumble.784508.last_uploaded", "1133658738236");
- user_pref("stumble.784508.newmessage", false);
- user_pref("stumble.784508.nick", "zerohex");
- user_pref("stumble.784508.password", "eshoog");
- user_pref("stumble.784508.prefetch", true);
- user_pref("stumble.784508.referral_count", "0");
- user_pref("stumble.current_user", "784508");
- Bunch of stuff there I'm not entirely comfortable with .... username, password, secret numeric username [usernumber] ... Hmmm, I might want to rethink my use of the StumbleUpon extension...
- user_pref("update_notifications.provider.0.last_checked", 1129893293);
- Huh?
- user_pref("yahoo.photos.yphLastBrowseDir", "/home/zerohex/Pictures");
- Hmmm. Severe discomfort realizing that I am a Yahoo user, and Yahoo has thin scruples... I don't recall installing any Yahoo firefox extensions, though - I will have to look into that.
Interesting - Firefox Edit menu options don't seem to interoperate as expected with the Klipboard tool under KDE - more of Gnomish foolishness.
Labels: add-ons, extensions, firefox, howto, linux
Sunday, April 23, 2006
Fixing the telnet:// protocol handler in Firefox 1.5.0.2 (Linux)
I had some problems getting Firefox to accept/provide arguments to a protocol handler command entered in the about:config page [see
previous post for about:config settings], but I have found a workaround, and posted a HOWTO document about it. See:
HOWTO Enable Telnet URI Handling in Firefox.
The gripe here is that telnet://host.name:portnum/ URLs didn't work by default in Firefox. The HOWTO gives a fix.
Labels: bash, config, firefox, hack, howto, linux, protocol, script, telnet, x11, xterm
Firefox about:config settings for telnet: URLs
Thursday, April 20, 2006
One night they took the car - the next night they took the phone - it was terrible, I tell you, this captivity of the mind - the singular fixation on being
About a Day
Well, it seems like it's been about a day now, and wurx is still up and running - have not started Mozilla Firefox this session. Posting this under Opera 8.52 -
Went so far before now - okay here it is:
a) Firefox 1.0.7 with various extensions and themes loading was crashing a not-quite-stock install of whatever X server shipped with SuSE 9.2
2) Remove 1.0.7 and install Newest Whizbang 2.5.0.2 or something
[NOTE: Couldn't figure out how to do a system-wide, multi-user install for the newest Firefox - may want to look into that later - running local this time]
3) Only a few extensions plugged in this time.
BANG) it didn't work anyway.
So what do YOU think. This is not life it is a SOAP-OPERA
For the record, just a few moments ago,
$uptime
10:01pm up 10:29, 3 users, load average: 0.05, 0.31, 0.31
Labels: 10.07, 2.5, add-ons, crash, extensions, firefox, linux, mozilla, opera, problem, suse
Sunday, April 16, 2006
FVWM: enable backing store and save unders for XFree86 (X11)
Got this from the Google cache:
[note that under SuSE 9.2 the XF86Config file says "do not edit", but there does not appear any way to add these options using SaX2 -ed.]
"Slow rendering of graphics: On Linux with Xfree4, if I move a window on top of the graphics window the graphics windows is continuously redrawn. It seems that the graphics windows needs to be re-rendered every time I do something with any other window.
"If your X server is run with backing store you won't have this problem. If backing store is present scilab does nothing, If there is no backing store then Scilab redraw graphics to emulate backing store (we will certainly change this in the future). By default, it seems that Xfree4 does not use backing store. Here is a remedy proposed by Lydia van Dijk: Those folks using XFree86 4.x can put the backing-store and the save-under setting into their XF86Config files and in the Screen sections:
Section "Screen"
# more stuff here
Option "backing_store"
Option "save_unders"
# probably even more stuff here
EndSection
"On Suse the configuration file is /etc/X11/XF86Config, on
Mandrake 8.1, XFree86-4.1.0-17mdk maybe like Redhat 7.1 (7.2):
it is in /etc/X11/XF86Config-4."
Wednesday, April 12, 2006
Weeelllll - I b'lieve now, ohiowa, that loaded firearms and alchohol can be a lethal combination. Futhermore, you should call me Tammy.
KBlogger Sux (in a nice way, of course)
Fuck yeah - that would be okay [not having a "blog this" extension in Konqueror] - but i'd rather have it as a separate app(let) *anyway*, but KBlogger Sux. How do I configure it? How do I even know it _did_ anything?
UPDATE: Actually did get KBlogger to work, sort of, but it still sux. The
Performancing extension for Firefox is better, but it apparently doesn't play nice with blogger.com if you have Blogger.com set up to publish your blog to another [i.e. not
blogger.com] site using e.g.
sftp.
Sunday, April 09, 2006
possible solution to the classic definition list problem in HTML...
Visual formatting model:
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0//EN">
<HTML>
<HEAD>
<TITLE>A compact box example</TITLE>
<STYLE type="text/css">
DT { display: compact }
DD { margin-left: 4em }
</STYLE>
</HEAD>
<BODY>
<DL>
<DT>Short
<DD><P>Description goes here.
<DT>too long for the margin
<DD><P>Description goes here.
</DL>
</BODY>
</HTML>
This example might be formatted as:
short Description goes here
too long for the margin
Description goes here
possible solution to the classic definition list problem in HTML...
Visual formatting model:
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0//EN">
<HTML>
<HEAD>
<TITLE>A compact box example</TITLE>
<STYLE type="text/css">
DT { display: compact }
DD { margin-left: 4em }
</STYLE>
</HEAD>
<BODY>
<DL>
<DT>Short
<DD><P>Description goes here.
<DT>too long for the margin
<DD><P>Description goes here.
</DL>
</BODY>
</HTML>
This example might be formatted as:
short Description goes here
too long for the margin
Description goes here
Archives
2004/09
2005/03
2005/04
2005/05
2005/06
2005/07
2005/08
2005/09
2005/10
2005/11
2006/01
2006/02
2006/04
2006/05
2006/06
2008/01
Subscribe to Posts [Atom]