LinuxQuestions.org - Blogs - Lumak's Guide to Random Things by lumak

Gnome 3 Tip - Quick Application Search

lumak — Mon, 19 Dec 2011 18:07:11 GMT

Gnome 3 will search the "comments" value of .desktop files when you start typing in the search box of the overlay screen. This search will include symbols and punctuation as well. This means that there are various ways to exploit the system and start grouping applications by symbols.

1. Copy your favorite applications' .desktop files from /usr/share/applications to ~/.local/share/applications
2. Edit the comments section to include a symbol for a group. E.G. ` ~ ! @ # $ < > . or anything you want... even a word.
3. Reset your session with Alt+F2 (type r, then enter) or log out and back in.
4. Press Alt+F1
5. Type one of your symbols
6. Observe all applications with that symbol in the comments appears.

Note, this is also the reason you can type something like 'music' and end up with several options relating to play music. Nothing magical, no special database, just edit .desktop files.

It would be nice if they added some customize options to the application launchers from within the shell instead of having to manually edit .desktop files.

More playing with BtrFS on Slackware

lumak — Tue, 03 May 2011 12:58:50 GMT

Ok well I'm done playing with all that for now.

Some issues I have with BtrFS for now are:
- Quotas can not be implemented on subvolumes.
- EDIT: A segfault when trying to use an encrypted volume with either a seed or its rewritable counterpart..
- Slackware does not ship with btrfstune to toggle the seed flag anyway.
- Slackware isn't equipped to multiboot multiple roots from the same btrfs.

However, in general, everything that is supported worked without issues. Even running bad commands that should break the file system did nothing than caused some reboot issues and a delay in the next mount (e.g. thinking that quotas were implemented and trying to use the filesystem resize options on subvolumes then specifying a size much smaller than the data stored on the volume)

So right now I reverted to a 15G btrfs lvm partition with later plans to see if seeding will eventually work with my setup. I don't really like running without Quotas placed on var, but I've only ever had a problem with MythTV reporting way to much to the log file by default. I don't plan on messing with that for a while. Maybe I can play with user quotas in that case.

Anyway, to get around a multiboot in Slackware, it seems very possible to edit the mount line in the 'init' file in the initrd-tree and re run mkinitrd without clearing the tree. by appending "subvolid="

Line 274 of init in initrd-tree

Code:

  # Switch to real root partition:
  echo 0x0100 > /proc/sys/kernel/real-root-dev
  mount -o ro,subvolid=257 -t $ROOTFS $ROOTDEV /mnt

The id can be found by running:

Code:

btrfs subvolume list /path/to/mounted/filesystem

If you don't want to multiboot from the same btrfs, then you can set the default subvolume to mount and you won't have to change anything:

Code:

btrfs subvolume set-default  /path/to/mounted/filesystem

However, none of that really matters if you don't tweak anything before/after the slackware install.
If you create the btrfs in the Slackware setup menus, then it will install Slackware to the default subvolume of the file system. This is not recommended as it makes snapshots appear in the normal filesystem. Since I didn't understand what I was doing when I first installed using btrfs, I had to modify and move files around after installing Slackware. At the moment, it seems the best thing to do is setup everything before running 'setup' to install Slackware.

That is...
- Create a partition
- run mkfs.btrfs on the partition
- create a subvolume for slackware
- set that subvolume to the default mount.
- create any other subvolumes you want (var, usr, etc.) However, this is only important, at the moment, if you want to create individual snapshots or mount a particular subvolume from another OS. Later it will be important for imposing quotas

Then after installing Slackware the normal way, you can:
- change the default subvolume mount (if dual booting and this isn't the primary)
- edit the fstab entry to include "subvolid="
- create an initrd, edit the init file, recreate the initrd

Not to mention, this all gets more complicated if you are using luks and lvm.

If you can avoid it, do not use btrfs on an lvm partition. It works, but it's needless and it may be the cause of my Segfault with attempting a seed partition... maybe I did something wrong. I don't know.

Overall the experience was interesting and the commands for managing btrfs are easy to use. The 'help' output from them explain enough to use the tools without knowing anything prior.

Playing with BtrFs and thinking out loud.

lumak — Sat, 30 Apr 2011 10:53:19 GMT

Warning: I'm mainly jotting things down so it may be a bit out of order.

During the install of Slackware 13.37, I figured I would give BtrFs a try and see how it goes. I quickly found out that BtrFs is NOT designed for small partitions. Where it used to be common practice to have something like the following, the same is not true for BtrFs:

Code:

Primary Partition
/boot  - 100M - ext3

Luks encrypted LVM
/      - 1G   - btrfs
/var   - 500M - btrfs
/usr   - 37G  - btrfs
/home  - 7.5G - ext4
swap   - 3G   - This was an afterthought.  Even with 5gigs of ram, you still need swap.  I had space left outside of the Luks/LVM but I wanted encrypted swap without having to manage another LUKS partition.

Large Share of insensitive data on Logical Extended Partition
/home/share - 625G - reiserfs - I haven't got around to playing the data movement game to change.

Tmpfs
/tmp - tmpfs - I've never cared to save this data between rebooting.

As I found out, the structure of btrfs does not allow you to have the full space available of the partition so 1G is more like 500M (600M-ish if you re-balance) and 500M is 250M. The goal of BtrFs seems to be a lot like LVM where you can expand over multiple drives and expand as your needs expand.

That's fine, I now have the chance to play around with LVM sizing and BtrFs subvolumes.

BTW I KNOW (now) - BtrFs in it self doesn't match well with LVM. Please spare me the response about how I'm not supposed to put a BtrFs on an LVM logical volume.

Anyway, it seems my solution now (for experimentation) is to consolidate the LVM partitions root, var, and usr -- increase the BtrFs over that partition -- create sub volumes -- move some file names around to the subvolumes -- edit fstab -- rebalance and defrag for good measure.

At least... I think I can do all that. To my understanding LVM is flexible on how it manages its data as it does not have to be contiguous. And because of that, it should be nothing to increase the BtrFS size after consolidating the three partitions. I'm not playing with important data here so I can afford to screw up. I don't think I'll mess up that badly and destroy the data on /home and /home/share

----------

Anyway... All this got me thinking, what are even the use cases of subvolumes? if you mount the root subvolume, you automatically have access to all the data in the subvolumes. And it's not like remounting a subvolume is anything significant, you can already do this with bind.

However, while typing the above I thought of a very interesting scenario that could be useful to software developers. For example:

We are still transitioning from 32bit to 64bit. So it's often the case that we may need to dual boot, develop in one or the other without reboot, use multilib, etc.

BtrFs actually lets you easily do the above with subvolumes while reducing the need to duplicate data for the multilib.

BtrFs
root (default subvolume)
slackware32-13.37
.. /
.. usr
.. usr/bin
.. usr/lib
.. lib
.. anything else you want to share
slackware64-13.37
.. /
.. Anything you think you might need to make separate snapshots of.

With this setup, you can now share the full partition between both OSes and boot into either one. Data stays separate and there is no need to specify how large each subvolume is. They all report the same size (a minor inconvenience to bar-graphs of diskusage... just use one subvolume). This alone is a good reason for using btrfs.

Now you can mount lib and usr/bin of the 32bit os into your 64bit os! Both oses are independent and you get a multilib compatibility layer (other than the compilers). Since both slackware64 and slackware32 are identical, there is no need to duplicate non architecture sensitive data like configurations or documentation.

However. There is a 'Gotcha' with using /lib from the 32bit OS in the 64bit OS. Slackware uses the FSF recommendation of /lib and /lib64 under multilib. But others use /lib and /lib32. This causes problems with a direct share of the 32bit /lib on the 64bit OS because there are still files installed in /lib from the 64bit os. The dhcpcd, firmware, modules, and udev directories for example. This would involve doing some playing around with the subvolumes on both OSes to avoid collisions. Also, there are files in the main alienbob multilib packages that get installed to /lib... glibc-solibs, gcc, glibc. However, these are probably binary compatible with the same files that would get installed under a normal 32bit install. I haven't quite worked out how to get around this yet. Maybe there is a way to mount the 32bit /lib in the 64bit os while keeping changed data on the 64bit copy of /lib?

I suppose you could be really messy have 32bit subvolumes as:
lib-bin
lib-dirs (with all directories, not just dhcpcd, firmware, modules, and udev)

The 32bit OS would symlink all the files in lib-bin to /lib and all directories in lib-dirs to /lib.
This would obviously take some setup after installing the 32bit os.

The 64bit OS would have to do something strange like mounting lib-bin as /lib. Maybe removing the colliding binaries from the glibc-solibs, gcc, and glibc multilib packages. And symlink the dhcpcd, firmware, modules, and udev folders.

I say Simlink (rather than a subvolume mount), because the root filesystem has to be complete with at least /bin /sbin and /lib directories. The subvolumes are automatically available when any one of the parent volumes are mounted. e.g. slackware32-13.37.

You 'should' be able to get away with a blind mount of the 32bit bin, usr/bin sbin/, usr/bin to "32" subfolders on the 64bit OS. (bin/32, usr/bin/32, etc.) This is sort of what Alien Bob's multilib "massconvert" script does with the 32bit packages.

Ok that theory of use needs some work. But the above mentioned solution (while messy) would work. If I can do some more reading and figuring out, it's something that you can't exactly do with LVM and ext234 formats. You would have to somehow mount the root filesystem for both OSes and do some bind mounts or other strange things in the initrd image. That would require some custom scripts in the initrd to handover control to the correct OS folder and further script manipulation in the OS to ensure it stays with it.

I know BtrFs shares a lot of the same concepts as ZFS, and you can probably do this same thing with ZFS or maybe there is a more elegant way of doing it. I'm not claiming to know what I'm doing. Just thinking.

Looks like I need to do more reading before I try mucking with stuff.

EDIT: there is 'btrfstune' which can be used for enabling filesystem seeding, however it does not appear to function on sub volumes. Only block devices. I suppose you could have two partitions. One for the 32bit seed volumes and another volume that contains the readwrite 32bit filesystem(s) and 64bit filesystems(s). Arguably, this is even more useful for software development. It allows you to always have a static snapshot of the original OS as intended by the developers and multiple rw subvolumes that can be independently booted. There may need to be some additional setup, but I think this could work out well. If only Slackware came with btrfstune... It's not compiled by a base run of the 'make' command in the source tree.

HOWTO: LAMP server on Slackware

lumak — Thu, 20 Jan 2011 17:01:16 GMT

This is a short HOWTO in getting a basic LAMP server running on Slackware. You will have to do more reading if you really want to know the details about configuring everything. I call this 'ready to confirm you want the defaults.' Please send me a message if you think I missed something.

Apache
---------------------------
set the execute bit on /etc/rc.d/rc.httpd

Code:

chmod +x /etc/rc.d/rc.httpd

As far as getting a basic web server up and running, that's all you need.
Slackware's web server directory is /var/www. You would place your webpage files in htdocs subdirectory. You will also find /srv has two symbolic links in it that both point to /var/www. Being that the /etc/httpd/httpd.conf file uses:

Code:

DocumentRoot "/srv/httpd/htdocs"

You can change the location of your webpage folders by copying /var/www somewhere and changing the symbolic links in /srv... Or you can customize /etc/httpd/httpd.conf anyway you want. It's up to you.

MySQL
---------------------------
MySQL is not set up with an initial database in Slackware. This is why so many people can't figure out how to start it. However, the start up script it self contains the instructions on configuring MySQL with the initial database. Read the comments on the script /etc/rc.d/rc.mysqld as well as the man pages for the following two commands.

Code:

mysql_install_db --user=mysql
mysql_secure_installation

Set the execute bit on /etc/rc.d/rc.mysqld

Code:

chmod +x /etc/rc.d/rc.mysqld

Additionally, if you need other computers on the network to access the database, comment out the following line in the start up script /etc/rc.d/rc.mysqld:

Code:

# This is a security thing.  If the database is just for local access,
# there is no reason to make it visible on the network.
SKIP="--skip-networking"

While not specifically needed, you can choose a configuration file from those listed in /etc/my-*.cnf to better control the resources that mySQL uses. This is really only needed if you have less than 2 gigs of ram on your system and access the database at lot.

Note, the mod_dbd module is already included in the httpd.conf file. SQL access should work for any module that uses it.

PHP
---------------------------
Uncomment the following line at the end of the /etc/httpd/httpd.conf file

Code:

Include /etc/httpd/mod_php.conf

On around line 231 in the same script there is the following block

Code:


  DirectoryIndex index.html

You may want to append "index.php" to allow you to have the default page of a directory with a php extension.

Now copy the example config files in place for php. Read them for more detail.

Code:

cp /etc/httpd/mod_php.conf.example /etc/httpd/mod_php.conf
cp /etc/httpd/php.ini-recommended /etc/httpd/php.ini

Perl/Python
---------------------------
Install mod_perl/mod_python from slackbuilds.org. Enable it in the /etc/httpd/httpd.conf file by adding to the end:

Code:

Include /etc/httpd/mod_perl.conf
Include /etc/httpd/mod_python.conf

You may want to look at their configuration files and make sure they are actually there. They may have a .new extension on them and or some other suffix.

Starting it up
---------------------------
You can choose to reboot or start Apache and MySQL manually.

Code:

sh /etc/rc.d/rc.httpd start
sh /etc/rc.d/rc.mysqld start

Note! if you already had the execute bits set when the computer was booted, then Apache is already running! MySQL probably was not if it was not configured. You can always restart the servers by using "restart" instead. For that matter, you can also stop them by using "stop".

UGH! I hate broken configure scripts/pkg-config setups (cairo-compmgr/vala)

lumak — Thu, 30 Dec 2010 21:15:42 GMT

Seriously, one of the most annoying things about linux and open source is the constant changing of library names and versions and the inconsistent use of pkg-config or the lack their of!

OK so the configure script for cairo-compmgr does a check for vala which goes fine. It knows where the compiler is and what to do. But later in the script it uses pkg-config to double check the library dependencies and it can't find vala... why? well because the script checks for the library 'vala-1.0' BUT?! the Slackbuild script doesn't install vala-1.0.pc it installs libvala-0.12.pc. Not only is the 'lib' part different but the version reference is also wrong! Most other .pc files use the standard .pc or -.pc. I suppose vala may still be too new to adhere to this and now I need to send a note to the Slackbuilds maintainer to consider adding a symbolic link so stupid configure scripts don't make this mistake again.

At least I've fiddled with make/configure and broken scripts enough to diagnose problems like this.

Fluxbox - Display Message on Desktop

lumak — Thu, 30 Dec 2010 07:05:46 GMT

Well I couldn't find a good solution to displaying text on the desktop similar to what conky does without actually using conky with only text display, So I decided to go with controlling window settings based on the custom window class of an xterm. I think it came out nicely.

desktop.sh

Code:

#!/bin/bash
# Not -R is a vim flag for Readonly to prevent accidental edits.
xterm -rv -fn 10x20 -class DMsg -e vim /path/to/textfile.txt -R

.fluxbox/apps

Code:

[app] (name=xterm) (class=DMsg)
  [Dimensions] {500 500}
  [Position] (BOTTOMRIGHT) {0 0}
  [Deco]   {NONE}
  [IconHidden]  {yes}
  [Sticky] {yes}
  [Layer] {Desktop}
  [Alpha]  {90}
[end]

The upside is I get a nice transparent terminal window on the desktop that's opened to a read only file in vim. It's still editable and sizeable (if you can alt+mouse to control sizes). It is also on the Desktop layer so it doesn't get in the way of other windows.

However, on the down side, it's still focusable and you can not create static text displays.

I suppose to get exactly what I want, I will have to install conky and set it up to display only a text file. Theoretically, since there is nothing to update, this should take up less memory and resources than xterm/vim... right?

Window Manager Ponderings

lumak — Mon, 13 Dec 2010 06:06:46 GMT

Why is it that two features of fluxbox aren't implemented in any other window managers?

1. fluxbox manages the transparency settings of any application you want. It's in the right click menu and it's extreemly easy to set there or even in the manual configuration. This seems like the easiest feature for any eye candy capable window manager to do.

2. fluxbox can group any kind of windows you want together and style them differently than other windows. So if you are working with gimp, you can group all the tool box windows together and they have a tab at the top. Definitely one of the more obscure features but extreemly usefull under the right situations. Not only that, but you can have very fine grained control over how to alt-tab through your apps and do other things like toggle key controls

I would probably use openbox if these features were implemented. So far, fluxbox is the only window manager that meets all my expectations. The only thing I could hopefor is maybe more development activity and regular releases... but then... for the most part it doesn't need it.

LibreOffice back at RC1

lumak — Thu, 09 Dec 2010 15:55:50 GMT

Haha... well that was exciting to see them create the document foundation and start a new port then immediately create a release candidate 1, then soon after create a RC2 that was some how worse than RC1, then drop the RC completely and go back to beta.

I assume this was because of the 'Go' open office portions they were adding among other things. Ya know... RCs are for bug fixes only, not feature additions.

Honestly, LibreOffice 3.3.0 should have been a direct copy of OpenOffice 3.3.0 with new branding. This would have allowed for the same bug fixes that openoffice was going through as a maintenance issue, all the while LibreOffice 3.3.1 would be worked on with the added features. We would have seen a more stable, more complete, and a better organize LibreOffice.

I'm starting to really doubt the quality of LibreOffice and its ability to replace OpenOffice.

CSS small caps

lumak — Mon, 06 Dec 2010 15:58:58 GMT

As many know, the proper css style to produce small caps does not function as expected:

Code:

font-variant: small-caps;

This style will often cause all the letters to be normal sized uppercase letter essentially producing the same style as:

Code:

text-transform: uppercase;

This is aside from the w3 recommendations stating to do this as a LAST resort. The browser is supposed to attempt to force the small-caps first by turning all the lower case letters to uppercase and resizing them smaller.

Anyway, as a css user, you have a few options. The quickest and easiest being:

Code:

font-size: smaller;
text-transform: uppercase;

However, you loose out on full size capital letters as 'small-caps' is only supposed to change the lower case letters.

If however, you are using small-caps to get a nice header and won't be using it on more than 3 words at a time, you can do the following:

EDIT: it appears my version of firefox ( 3.6.10 ) has a bug with adding extra space after a block that changes the font size of a block element using the virtual class :first-letter. I'm diagnosing a workaround...
EDIT EDIT: OK it appears this is a really old bug that has never been fixed and I can't think of an elegant workaround. Webkit browsers will display this properly so it's still some what good code but being that it's completely worthless to all of firefox, it's probably not worth using. Just use the messy example posted in the comment below.

Code:



  small caps test
  



  
    T
    
      This 
      Is 
      Small 
      Caps

The down side? you must have each word as a separate span tag, it will only work on Properly Capitalized Words not RaNdoMly CapItalIzEd, and you must turn the words in to floating blocks so that the :first-letter virtual class works. This makes positioning troublesome if you have other normal words on the same line. If positioning around other block elements, you may want to wrap it in another block element that's positioned the way you want in relation to the others.

The values of "1.35" and "0.74" were judged on what I thought looked good when testing under midori and firefox. This proportion may not look right to you and you will have to change them to readjust the first font size in proportion to the rest of the page. Test it by placing a normal sized capital letter just before the first word and playing with the values.

Oh and notice I gave it the class '.title'

Do not be tempted to miss use CSS where your html contains formatting information in the classes like class="smallcaps bold large" CSS is designed to make formatting transparent to the source. This is why I say to put the words in 'span' classes. This way you can change the style by only altering the root title class. The sub styles of '.title .title' will automatically inherit the main '.title' style. And if you fail to style it... say a print style... the words will appear all on one line without having to think about styling it. Additionally, if you place it all in a 'h1' tag then it will be auto formatted for other processing applications as well as have a default 3em size in bold.

I hate when I see css examples that tell people to tag html will really long multi class strings. At that point, you might as well code pure html with the old style tag properties because you obviously don't care about portability and replacing the style later.

People that complain about KDE4

lumak — Wed, 24 Nov 2010 19:10:08 GMT

Curious... I wonder, what is it about KDE4 that people don't like? New tools? New lay out? Visually different? Resource Hog? QT4?

This is all akin to wanting Windows 7 (Aero) to function just like XP or its predecessors... sure, you may be able to make the gui and theme look and feel like the previous, but all the system tools are going to be updated.

As with all history of computing, when something significantly changes, it's going to take time for programs to adapt and be updated. This is still a transition period where programs are still going to support only QT3, and some are switching to only QT4, and some still may try and support both for a period.

If you were a windows user, what would you complain about? Would you expect Microsoft to stay stagnate and unchanging? How far back do you expect backwards compatibility? If your programs didn't run on Windows 7, you would probably still use XP, that's fine. That's what you need. But you wouldn't need to complain about it. So your tool of choice here is It works as expected and it's what you need. Don't block progress, or the illusion that you think others see as progress.

As others have said. If you want to have KDE3/Trinity then you support it. Hell, maybe if Trinity becomes its own and plays extremely nice with QT4 and the rest of , it might even be supported by . That being, it would probably have to use QT4 or make a derivative QT3 library that doesn't conflict with QT4 in anyway.

As for Slackware choosing to drop KDE3 because it was determined to be 'unmaintained' is BS. Blackbox isn't maintained (though arguably extremely stable and functions perfectly as-is). However, it is still installed with Slackware. Assuming the rules of 'maintained' status, BB should be removed and just keep Fluxbox. No. I think the real reason is QT. The ability to easily compile QT software between QT3 and QT4 on the same system take a bit of extra knowledge. For example, using PKG_CONFIG_PATH and LD_LIBRARY_PATH. If you could install KDE3 along side KDE4 without /opt and have no conflicts, then there wouldn't be an issue.

//-------------

But I use fluxbox... even that hasn't had a new version in over a year. But even if Slackware dropped BB and FB support, I would probably just use Openbox. Additionally, they would probably provide binaries for Openbox if that was the case. I wouldn't be sad or complain. I would post a farewell in the designated thread and move on.

Tar progress bar for scripts

lumak — Wed, 20 Oct 2010 06:55:21 GMT

I use SlackBuild scripts a lot and the typical standard is to use the verbose option to print all the file names as they are extracted. This is annoying. However there really isn't a progress bar for tar... or is there?!

Well if we knew the number of files in the tar file we could probably calculate the percentage progress of the file list, but that would take too much time... Thankfully there is --checkpoint

So, outside the script, you find the general number of files that the archive is going to have.

Code:

# Don't do this in your script.  It takes too long on really large archives.
tar --list -f  | wc -l

Then we divide this number by our standard console width and add in the script as follows:

Code:

# 76 dashes and 2 pipes for our standard console width
echo "|----------------------------------------------------------------------------|"

# please note the dot in the checkpoint.  This is so it outputs only a dot every
# specified number of files.
tar xf  --checkpoint=.

# start a new line for niceness of display
echo ""

for webkit, this was roughly 11120 / 78 = 142 and the output is:

Code:

|----------------------------------------------------------------------------|
........................................................................

Now doesn't that look nicer than the annoying file list that nobody cares about?... Yes I know the line of dots doesn't go to the end. It's not exact and you will have to tweak the number a bit.

OK OK I hear your argument. "--checkpoint only a gnu/tar option." To which I reply, "It's a gnu/linux OS, not BSD or other." I'm not trying to be compliant with anything other than a recent gnu/linux OS.

Slacklet - a new OS Mod Proposal

lumak — Sat, 09 Jan 2010 03:49:14 GMT

If you have read my other post, you'll know I want to install GNU/Linux onto a Tablet PC. I've thought about it long and hard and think that Tablets really deserve an OS design for them.

I propose a Slackware Tablet - Slacklet.

Why?

Slackware is simple to manage and it's the OS I prefer.

Real-world Justifications?

Less work on creating a Tool Chain.
The user can use their main machine running Slackware to compile the extra packages they will need for their tablet.

Slackware is stable.
I trust the brand to provide stable vanilla software.

Ease of custom install.
Anybody can easily customize an install of Slackware using tagfiles

The concept of a Demi OS.
I don't plan on distribution of an entire OS. I want to educate people on customizing an install to their exact needs. Basically offer tagfiles for a bare-minimum Slackware Install with only things useful to a tablet, then offer the ability for a new "s" source tree containing all the extra packages needed to get a useful tablet going. E.G. Matchbox and Journal programs. Additionally have guides written by users (screened and cleaned up of course) for each tablet they have gotten it to work on as well as the extras they had to do (manufacture buttons, stylus, and peripherals.)

Negatives?
- Requires a desktop machine with a full Slackware install to compile software
- People shy away from Slackware
- Download to install time would take weeks to sort through and work out
- Low user base to offer support
- Lots more.

I don't know... suggestions, questions, concerns, insults?
ideas for a host website?

BTW enjoy the logo :)

Attached Thumbnails

having fun with bash varaibles: indirect variable references

lumak — Sat, 28 Nov 2009 19:38:28 GMT

Well I recently started messing around with bash scripts as cgi scripts and was wondering how to pass variables in and what not.

Well it turns out, Apache will define the variable $QUERY_STRING which will contain everything after the first question mark. You can then pars this string to define variables. The general format is "name1=val1&name2=val2". It turns out you can actually declare variables in bash using 'declare' and parameter expansion.

Code:

VAR="hello"
VAL="how are you"
declare $VAR="$VAL"

echo $hello
# >> how are you

unfortunately, I couldn't find a very intuitive way to do the opposite. That is, get the value of a variable name if the name is given as text... But this seemd to do the job well enough.

Code:

eval echo '${'$VAR'}'
# >> how are you

In that example, $VAR is expanded to "hello" then eval will execute echo as if you had actually typed "echo ${hello}" on the console! if you remove the 'echo' it will be as if you typed "`$hello`" and in this case say "bash: how: command not found" which may or may not be your intention.

Anyways... I'm cleaning up a sample cgi script that utilizes this and will upload it when it's done.

Attached Files

cgi_vars.txt (2.4 KB, 529 views)

Working Scsi Robot DVD Library! Yeah!

lumak — Sat, 21 Nov 2009 07:24:02 GMT

So I finally got off my ass and got a Plasmon D-Series dvd scsi library robot working! Honestly, for basic functionality, it doesn't take anything special.

Steps:
1. Full install of slackware 13.0
2. Make startup script for autofs (or grab it from Alien Bob's page)
3. Configure it to use autojuke (one of the tools from the scsi-changer project):
auto.master should have "/juke /usr/sbin/autojuke" and configure it with /etc/autojuke.conf
4. Configure it to NOT reinitialize the robot every time it starts:
create a modprobe configuration for the 'ch' module with "options ch init=0"

5. Set up the /juke share with http, smb, and sshfs if needed. Unfortunatly, I don't believe it's possible to set it up for nfs or ftp!
6. Access any slot by number or volume name!

Wow that's simple.

OK now for the hard part:
* Extend autojuke to use double sided medium
* Allow for remote management of burning CDs/DVDs on either a timed, data fill, or manual from both windows and linux
* Attempt remote streaming of encrypted dvds (more than likely with VLC)
* Attempt UPnP sharing. This could be a long tedious task of mounting each slot, refreshing the UPnP server to fill it's database and testing by allowing the automounter to unmount and store the slot then attempting to access from a UPnP client
* Making a database and auto filler for all the content on the CDs
* Add needless feature support to the autojuke to allow for audio CDs to be accessed as if they were wave files.

WOOT!

Linux on a Tablet PC - which way to go.

lumak — Fri, 18 Sep 2009 16:21:27 GMT

Which way to go from here? I recently received a Fujitsu Stylistic ST5010 tablet PC with windows xp tablet edition installed. I instantly thought of how great it would be to get linux installed on it but haven't taken the plunge yet... Mainly because there are no distros specifically for tablets! Not to mention, I'm a slackware user.

This leaves me pondering:

1. A lot of the software I looked at seems unmaintained.
2. Matchbox-project looks promising but appears to have no updates for 2 and a half years. http://matchbox-project.org/
3. The only decent journal software is java based (no offense, I just don't like java)
4. The only way I can think to compile things is on the tablet it self (a 1 gig Pentium M) or to compile in a virtual machine with an AMD Athlon(tm) 64 X2 Dual Core Processor 5600+
5. Do I force a default user login into a graphical environment? or do I provide a gui log in... how do I handle if X crashes? Do I forcefully turn off the system? Do I go the way of Knoppics and very annoyingly reboot the x server?
6. WHAT distribution do I choose?!

OK granted, old software is still relevant and works as is (e.g. blackbox). AND Jarnal may be java but it supports text and a pdf export. AND I don't mind compiling on a virtual machine.

All options of dealing with the x server are equally good options.

That only leaves the distribution...
I know slackware, I can deal with others. Choosing a distro with a repository based package manager will make things 'easier' But it won't be slackware... I will have an opportunity to learn other distros. I've been spoiled with the very direct and to the point '/etc/rc.d/rc.my_daemon start|stop|restart'

But slackware... is it really up to the task? I suppose it is what I make of it... I have the docking station for keyboard, mouse, and cd incase something goes wrong...

Hrmmmms.... Maybe I'll give it another month to decide.

slackware64 multilib QT issues

lumak — Thu, 17 Sep 2009 23:13:23 GMT

UGH! ok I was trying to compile q4wine for god knows what reason and half the variables in the build/CMakeCache.txt were pointing to /usr/lib64/qt and the other half were pointint to /usr/lib/qt.

I solved my issue by adding

Code:

CMAKE_PREFIX_PATH=/usr/lib$LIBDIRSUFFIX/qt \

on the line right before
cmake \

Hope this helps somebody else with cmake and qt problems.

Slackware64 nvidia-driver-compat32

lumak — Thu, 17 Sep 2009 20:45:21 GMT

EDIT: I added a new script for 195.36.24 as well as a script to compile libvdpau-compat32 on a 64 bit system. I never hear any feedback on this stuff... is it working for other people? Do I have errors anywhere?

ORIGINAL POST:
I'm one of those people that personally don't like installing things on my system without having proper packages made for tracking the files. As such I edited the scripts on SlackBuilds.org for an additional package to support this.

Attached is the build script and accompanying nvidia-switch32. You will need the binary package from nvidia that has pkg2.run on the end as it contains the required usr/lib32 directory.

Also, being that 32bit compatibility for the nvidia driver isn't actually a driver... perhaps calling it nvidia-driver-compat32 isn't exactly right. It is a few OpenGL libraries that get installed to /usr/lib. Maybe nvidia-libs-compat32. Either way, the name is not important. This package does not belong on SlackBuilds.org because they don't really have compat32 packages. Wine and Virtual box don't count.

I doubt this package/build script really has any other place that to be distributed here as an example for the stubborn.

Use and distribute as you wish.

Please provide feedback if you find errors.

Attached Files

	nvidia-switch32.txt (9.1 KB, 846 views)
	nvidia-driver-compat32-195.32.24.SlackBuild.txt (4.0 KB, 12 views)
	libvdpau-compat32.SlackBuild.txt (2.7 KB, 9 views)

man to pdf converter

lumak — Tue, 08 Sep 2009 21:21:38 GMT

EDIT:
while it was fun to do, this was actually completely useless. Just hotkey your terminal and type "man'! Original post as follows.

I recently got a tablet PC and instantly wanted to fill it with a wealth of knowledge. This lead me to want PDFs of the man pages. Luckily, man can output in post script and there is a ps2pdf converter on most Linux distributions... But you may not want to convert them manually and you may want a nice folder structure for them.

It's not perfect but go ahead and check it out.

EDIT:
Known Bugs:
- Some manual pages may contain errors when converted to ps then to pdf. These errors are not logged yet and the script continues happily.

- some file names are name.

x[.gz] and others are name.

p[.gz]. The ones with the p are actually POSIX manual pages and the matching files with no p are Linux manual pages.

# package bug
muttbug may be an actual man page but redirects to flea with a man page link. It may link to a non existent file.

# package bug
compress-dummy.1.gz appears to be an empty file and is not converted.

version numbers in file name affect sed substitutions.
gimp-2.6.1.gz
gimptool-2.0.1.gz
gimp-console-2.6.1.gz
gimp-remote-2.6.1.gz
gimprc-2.6.5.gz

# package bug
pthread_sigmask.3p.gz redirects to man3p/sigprocmask.3p which does not exist
additionally pthread_sigmask.3.gz reported the same issue. Not shure why.

mount.nfs.8 and umount.nfs.8 were not converted. But for some reason [u]mount.cifs.8 were!

WINDOWS ISSUES WITH FILE NAMES

- the man3 section has files named ExtUtils::name::suffix.3
mann on slackware has pkg::create and platform::shell as well.
These file names are not allowed by windows because of the colons ':' and will get copied over with names like SIH273~1.PDF. This is a windows issue and will not be fixed. Consider doing a rename on the file BEFORE copying to windows.
find -name '*::*' -exec rename '::' '..' {} \;
It will need to be run 4 times. CPANPLUSS::Shell::Default::Plugins::CustomSource.3.pdf

- if copying files to a windows machine, the following files will be in naming conflict due to windows not being case sensitive. This is a windows problem and I don't plan on fixing things that arn't issues. Additionally, windows may copy them over in a different order than what is expected and you will have to make sure of which ones to copy.

DB.3.pdf - db.3.pdf
Standards.7.pdf - standards.7.pdf
CORE.3.pdf - Core.3.pdf
Errno.3.pdf - errno.3.pdf
NAN.3.pdf - nan.3.pdf
_Exit.2.pdf - _exit.2.pdf

END EDIT

Code:

#!/bin/bash
# man2pdf
# Convert the man page(s) specified to pdf files
# 
# This script is in the public domain and comes with no warranty
#
# Options:
# --all - convert all manpages in $MANDIR.  Currently only grabs
#         those in man[1-9,n]. Optionally pass a section number
#         as a second argument to convert only those in man
#
# Without the above option, the user can convert man pages by
# man2pdf [section] man_page [[section] man_page] [...]

MAN_PAGES="$@"             # If --all is not the first parameter then all
                           # the parameters are assumed to be man pages

SECTION=${2:-0}            # used with --all as the section number.

OUTPUT=${OUTPUT:-$(pwd)}   # Your output directory


MANDIR=${MANDIR:-/usr/man} # You may have your own copy some place else

LOGFILE=${OUTPUT}/output.log # currently not supported

if [ $1 != "--all" ]; then
    # This may match several and will store them as a space separated
    # list of exact paths to file names
    page_list=$(man -M $MANDIR -aW "$MAN_PAGES")
else
    # Store all the file names found in the MANDIR and convert them.
    # NOTE! this will only produce a list of actual files and symbolic links
    # will be ignored.
    page_list=""
    if [[ $SECTION == [1-9,n] ]]; then
        # Find all the files in the given section
        page_list=$(find $MANDIR/man$SECTION -type f)
    else
        # Recursively add all the files listed in each section
        for current_section in 1 2 3 4 5 6 7 8 9 n; do
            page_list="$page_list $(find $MANDIR/man$current_section -type f)"
        done
    fi
fi

#  Loop through all the files found 
for current in $page_list; do

    # These sed substitutions make the assumption that the section
    # name is found after the first dot in the name.  Man pages may be compressed
    # or uncompressed, as long as they are name.n

    # Find just the name of the man page and strip the section
    current_page=$(basename $current | sed -e 's@\.[1-9,n].*@@')
    # Find the section number from the file name
    current_section=$(basename $current | sed -e 's@.*\.\([1-9,n]\).*@\1@')

    # Quick check to ensure we will get a man page
    # We let it print here so we know our progress for the --all command
    # TODO: print filenames and errors to a log file when we come across them.
    man -M "$MANDIR" -aW "$current_section" "$current_page"

    # Just in case my regular expression magic fails.
    if [ $PIPESTATUS -eq 0 ]; then
        # Create the storage directory.  There doesn't appear to be any harm in letting
        # this happen for every loop
        mkdir -p "${OUTPUT}/man${current_section}"

        # Our file name
        pdf_file="${OUTPUT}/man${current_section}/${current_page}.${current_section}.pdf"

        # Pipe the man page into ps2pdf
        # This line may complain about the format of some characters or specifics of some
        # man pages.  I didn't make a list and I didn't log them and I didn't check
        # the produced man pages in detail.
        man -M "$MANDIR" -t "${current_section}" "${current_page}" | ps2pdf - "${pdf_file}"
    fi
done

Fluxbox/Openbox - calendar in menu

lumak — Sun, 23 Aug 2009 21:05:17 GMT

MOVED: added this to the fluxbox wiki as "Calendar Menu" updates may be posted here or there.

Here is a nifty thing to include a calendar printout in your fluxbox menu. It currently coded in bash and is very crude. The print out only looks best when your menu is left justified and the text is mono spaced. An optional fix for this may be to convert spaces to underscore (not yet implemented).

create a directory ~/.fluxbox/cal and place the attached script in it.

initialize it once by running it as

Code:

~/.fluxbox/cal/updatecal.sh reset

Then add the newly created menu to your main menu

Code:

[submenu] (Calendar)
   # the name is different between version .1 and .2
   [include] (~/.fluxbox/cal/calendar-menu)
[end]

NOTE
this script was designed to handle fluxbox's inability for dynamic menus and stores the menu as ~/.fluxbox/cal/cal in .1 and ~/.fluxbox/cal/calendar-menu in .2

I plan on expanding this to use a crude appointment system for adding annual events like birthdays. Unfortunately, because it is essentially a menu, you would see important dates as a menu off to the right for the current week.

Known Issues:
I can't find any, openbox may experience some. The speed issues of version 0.1 seem to be gone now that I added a separate file for storing the main calendar month and year.

Version 0.2
- Optimized Some functionality
- separated menu format from the rest of the code
- added theoretical support for openbox
- added dynamic menu support

Version 0.1
- Basic Calendar
- Tools menu option to: reset, change to previous month, change to next month
- Menu option to view previous month
- Menu option to view next month

Attached Files

	updatecal.sh.txt (3.5 KB, 758 views)
	updatecal-0.2.sh.txt (10.4 KB, 803 views)

Fluxbox - Terminal Window PC Game Style

lumak — Sat, 22 Aug 2009 07:26:13 GMT

MOVED: I added this to the fluxbox wiki under "Category:Configuration howtos" Any changes may go on either. Enjoy.

Warning. I tried this on my much older 32 bit laptop and it didn't respond that quickly to toggling the consoles. I'm guessing that long bit of code for the key command actually takes more processing than I thought.

OK I realized I could make xterm (any terminal?) act like the command console in a FPS game. So here are all the settings to do that.

edit ~/.fluxbox/apps

Code:

# Group all xterms into tabs
[group]
 [app] (name=xterm) (class=XTerm)
  # edit to your screen size
  [Dimensions]  {1280 491}

  # I like it here.
  [Position] (TOPLEFT) {0 0}

  # Why not?
  [Alpha]   {190}

  # available on all work spaces regardless of Iconify
  [Sticky]  {yes}

  # optional - do not use [Hidden] or [FocusHidden]
  [IconHidden]  {yes}

  # set to be on top of everything
  [Layer] {2}

  # display only a tab and a border regardless of
  # window manager settings.  This looks nicer than
  # with all the window decorations.
  [Deco] {516}
[end]

edit ~/.fluxbox/keys

Code:

# (name!=xterm) prevents xterm from being included
# in the alt-tab cycle
# alt-tab
Mod1 Tab :NextWindow {groups} (name!=xterm)
Mod1 Shift Tab :PrevWindow {groups} (name!=xterm)


# Toggle xterms
# Control (tilde key):
# If there are xterms open then switch between showing
# them and hiding them.  If there are not any open, then
# open one up.

# all one line
Control 49 : If {Some Matches (xterm)}   {ToggleCmd {NextGroup (name=xterm)} {MacroCmd {NextGroup (name=xterm)} {Delay {Iconify} 1}}}    {Exec xterm}

edit ~/.fluxbox/init
# add/replace as needed or use menu options to select tab placement.

Code:

session.screen0.tab.placement: BottomLeft

If you want to have the label on the tabs of the xterms numbered (purely aesthetics) then you can use the attached script. Instead of executing xterm with all your hot keys and menu options, you would want to execute this script. It's in the public domain and comes with no warranty.

Attached Files

open_term.txt (1.6 KB, 680 views)

hplip hpaio sane setup

lumak — Tue, 18 Aug 2009 02:39:46 GMT

I have a PSC 1610 All-in-one printer that always gets auto recognized when I scan for it as root but I always forget how to configure it for regular users.

1. add "usb 0x03f0 0x4811" to /etc/sane.d/hpaio.conf
2. ensure 'hpaio' is uncommented in /etc/sane.d/dll.conf
3. ensure the user is added to the 'lp' group! or whatever group owns the device.

My device is /dev/lp0.
you can type "ls -l /dev/* | grep lp" to see if your printer/scanner was auto detected and configured by Hal.

There are various ways to find the values for step 1. the first number is the vendor ID and is more than likely going to be 0x03f0 for this driver.
dmesg | grep -i 0x03F0
dmesg | grep -i printer
dmesg | grep -i lp0

as a last resort try going through /sys/bus/usb/devices and trying to find the correct vendor and product id's in idVendor idProduct and serial if you know it (scanimage -L, as root)

Hope that helps somebody aside from me when I need to set it up again!

blender x86_64 compile guide for slackware64-pre-13.0-current

lumak — Sun, 24 May 2009 05:01:25 GMT

Just use the latest blender 2.49. They now support python 2.6.2 AND ffmpeg. Only follow the below if you require verse-server (I don't know if their pre-compiled binaries support verse client functions though). Also, you have to export PYTHONPATH=/usr/lib64/python2.6 to run the pre-compiled binaries.

--------------------------------------------------
Original Post:
--------------------------------------------------

Anyway... you don't have to modify too much. OpenAL and freealut are optional for the build. I suggest modifying the SlackBuilds.org scripts to get them to install into /usr/lib64

You can do this by the following modifications as well as passing ARCH=x86_64 to the script:

OpenAL

Code:

cmake \
  -DCMAKE_C_FLAGS="$SLKCFLAGS" \
  -DCMAKE_INSTALL_PREFIX=/usr \
  -DLIB_INSTALL_DIR=/usr/lib64

freealut

Code:

CFLAGS="$SLKCFLAGS" \
./configure \
  --prefix=/usr \
  --sysconfdir=/etc \
  --localstatedir=/var/lib \
  --libdir=/usr/lib64 \
  || exit 1

Blender has several compile choices: make, scons, and cmake

Nan - make
The make files littered about are just regular make compatible files that get configured through environment variables or through a user-def.mk. To compile on slackware64, you have to patch the make files to use /usr/lib64 for a few system libraries. Once you got that, you can do the following:

Code:

REL_CFLAGS="$SLKCFLAGS" \
REL_CCFLAGS="$SLKCFLAGS" \
NAN_PYTHON_VERSION=$PYTHON_VERSION \
NAN_PYTHON_LIB="/usr/lib$LIBSUFFIX/python$PYTHON_VERSION/config/libpython$PYTHON_VERSION.a" \
NAN_NO_OPENAL=0 \
make

Blender - scons

Code:

scons WITH_BF_VERSE="1" \
      BF_PYTHON_VERSION="2.6" \
      WITH_BF_OPENAL="1" \
      BF_OPENAL_LIB="openal alut" \
      REL_CFLAGS="$SLKCFLAGS" \
      REL_CCFLAGS="$SLKCFLAGS" \
      blender blenderplayer

Blender - cmake

Code:

  mkdir cmake-build-dir
  cd cmake-build-dir
  cmake -G "Unix Makefiles" path/to/source/blender-2.48a \
    -DWITH_PLAYER=1 \
    -DCMAKE_C_FLAGS="-O2 -fPIC" \
    -DCMAKE_CXX_FLAGS="-O2 -fPIC"

  make

Other features you can enable are WITH_VERSE to edit shared assets online with others and WITH_OPENPM which I think makes blender run more efficiently.

After a few compiles, it appears the following features require extra dependencies.

WITH_QUICKTIME=0
Quick time in blender is not supported for *nix.
Everywhere in the source code when WITH_QUICKTIME is enabled, it includes for win32 and for apple. There is no such compatible header in libquicktime for linux. It's not a big loss anyway, just convert the files before or after import.

WITH_FFMPEG
If you install the ffmpeg package 0.5 from slackbuilds, the headers go into include folders for each seperate lib compiled for ffmpeg. Blender expects the headers to all be in /usr/include/ffmpeg/. The source needs to be patched for a few files. I thought I had a working patch but it errors out when attempting to use ffmpeg.
You are better off on trying to use the ffmpeg distributed with blender. You may be able to statically link it in. I'll be working on it a bit more.

WITH_OPENAL
2 dependencies, might as well use them.

WITH_WEBPLUGIN
Requires a build time dependency http://releases.mozilla.org/pub/mozi....8.0.4.tar.bz2. It builds on x86_64 even though the sdk is i686. objdump says it is an x86_64 binary. The README for the plug-in is hidden under source/gameengine/GamePlayer/netscape/src/README.

I'll probably edit this blog everytime I find a mistake...

OK the uploaded files should be good to stay as is for now. the nan_make does not allow for building the web plug-in. It seems like the Makefiles just don't point to working source for it. If you really must build it, use the cmake method. If you need ffmpeg support, use the nan_make as this will compile a static version of ffmpeg into blender. I did get a buffer overflow with some random settings but other settings worked and rendered a video.

Attached Files

	blender.nan_make.SlackBuild.txt (17.5 KB, 336 views)
	blender-2.48a-nan_make_x86_64.patch.txt (1.4 KB, 299 views)
	blender.cmake.SlackBuild.txt (15.4 KB, 422 views)

Slackware Undocumented 'Well Duh' Quark #2917 - console vim and X11 clipboard

lumak — Sat, 21 Mar 2009 04:41:57 GMT

The special registers * and + do not function in console vim in Slackware. Vim has been compiled both with -clipboard and -xterm_clipboard. If you enable GPM, you can shift highlight in console vim to middle click past in any x application. You can also Shift+middle click in console vim to past text from any x application. If you don't want to use the Shift modifier, you can sacrifice mouse functionality and type ":set mouse=r" in normal mode. This issue does not affect gvim.

------------------ Rant ----------------------

This was bugging me for a while because I much prefer the xterm when the gui version has the same functionality. I never took the time to find out why the hell the console version could not paste text from the clipboard. After a short bit of searching I found out the secret and how to find out if I have the support. "vim --version | grep clipboard"
5 seconds after this I realized the importance of a conscious choice by the Slackware maintainers to use -clipboard and -xterm_clipboard. Why the hell should a console program include links to libraries that may not even exist on a bare bones server?! What if you have to fix 3 characters in that HTML file to fix that problem and vim is you favorite/only text editor? I know I wouldn't want to have X11 installed on this server and I sure as hell wouldn't want to move the file some where else just to edit it then move it back. Thank you Slackware for asking "what if there is no X11?" What about copy/paste between two vim instances running in the same xterm? I don't know, the normal registers don't appear to be shared between them. I'm thinking you are better off using the 'tab' or window feature the buffers will work then. But why the hell does ":set mouse=r" work if the clipboard is not compiled in? Because the options +mouse +mouse_gpm and +mouse_xterm are all enabled. I'm guessing these are enabled because one is more than likely to not have X11 but leave smaller minor packages in? OK well I thought I did have a rant as a thank you for providing an x11 free vim but now I'm confused about the mouse_gpm and mouse_xterm options...

Compiling with new static libs

lumak — Sun, 01 Mar 2009 17:27:43 GMT

OK, had you read my other post on gnuCash, you know that I hate gnome dependencies. Some of the programs I use have up to 5 gnome dependencies and I don't feel it's worth it to install them. But I managed to compile them into the gtkpod binary using static libs. This goes against everything slackware has taught, but I feel that it was worth it in the end.

Basically, I have a build script that builds gtkpod. Before it does that, it will build its gnome dependencies as static libs only into a tmp folder and the user can freely delete them when they are done.

Before I begin, I will give the standard warning on such things. If the library is updated due to a security fix or other such things, your main program now has that security flaw. You have to recompile the main program with the static library again so that you can have the security fix.

This is why having 10 programs all with the same dependency should use shared library files so that they are updated all at once. But I'm sure you already knew this.

OK now here are the basics. gtkPod is missing the libgnomecanvas which is missing gail. We want to run gtkPod without infecting our /usr/{lib,include} with these files. We just have to compile using the proper options for the program.

Code:

GTKPOD=/tmp/location/of/gtkpod-src
# compile gale
# cd /to/gail-src
./configure \
  --prefix=$GTKPOD/staticlibs\
  --sysconfdir=/etc \
  --localstatedir=/var \
  --enable-static=yes \
  --enable-shared=no \
  --build=$ARCH-slackware-linux

make
# This installs the files to the --prefix which only contains the
# static libs.
make install-strip

Code:

# compile libgnomecanvas
# cd /to/libgnomecanvas-src
PKG_CONFIG_PATH="$PKG_CONFIG_PATH:$GTKPOD/staticlibs/lib/pkgconfig" \
./configure \
  --prefix=$GTKPOD/staticlibs \
  --sysconfdir=/etc \
  --localstatedir=/var \
  --enable-static=yes \
  --enable-shared=no \
  --build=$ARCH-slackware-linux

make
make install-strip

Now you just compile gtkpod as normal except you add "$PKG_CONFIG_PATH:$GTKPOD/staticlibs/lib/pkgconfig \" before typing ./configure in gtkpod's src directory.

This works nice for simple dependencies. I personally wouldn't use this method on things that used gconf as that installs stuff into /etc and does other retarded things.

What I belive can also be done is installing with the prefix /opt/gnucash for all programs and still keep shared libraries. However, I don't know the commands/settings to do it properly. The only way I could think to do it would be to "make install DESTDIR=/opt/gnucash" for everything. package opt/gnucash. then delete the directory and install the package. That seems messy.

Additionally, if you have static and shared libraries on your main system, you have to do something else to target specifically that static libraries. I'm not sure how to do that either. Enjoy.

THANK "BOB"!

lumak — Tue, 20 Jan 2009 05:12:12 GMT

OK well maybe not "Bob" personally but 'god' didn't seem the right thing to thank at the moment. OK Off topic.

The other day I realized that the only way to fix webpage style conflicts with dark themed conflicts was to use a firefox add on. A little bit of searching later revealed "Stylish" Tnx-to-da-god-be this solves so many problems. It basically applies your own personal css styles to any page individually or all at once! After some tweaking, here is my all-in-one-fix for a handful of pages with various different -isms.

input,input[type="text"] {
background-color: #666666 !important;
color: #CCCCCC !important;
}

input[type="submit"] {
background-color: #CCCCCC !important;
color: #000000 !important;
}

select {
background-color: #CCCCCC !important;
color: #000000 !important;
}

Note the use of "!important" on everything. Some webpages, despite this add-on's greatness, still override somethings (certain online banking webpages for example) Now their theme is just as generic as anybody else! damn inconsiderate @)(@#*$ Not everybody has lightly colored themes!

OK anyway the reason 'submit' and 'select' are reversed from most dark themes is because I've noticed that, at least without a separate firefox theme, buttons and drop down lists typically have a white/light color base but still use your default chosen colors for text. The result is a barely readable button. So the simplest thing is to just force the text back to black. I suppose you may run into problems that use graphic buttons with real text but a graphic background. But, I'll fix that when the time comes.

Definitely check out "Stylish" add-on if you use a dark desktop theme with firefox and are frustrated with things not appearing correctly.

Now if somebody could fix the problem with certain applications that have a multitude of black icons. At least "the gimp" realized this flaw and has blue icons and icons that include both light and dark colors where you can either see the outline or the color of the tool.

The pointlessness of gnuCash

lumak — Fri, 19 Dec 2008 05:12:58 GMT

Generally this points to gnuCash on slackware BUT I feel it goes towards everything.

gnuCash doesn't work.

I don't mean that if you actually get it compiled and running it doesn't work. I mean it doesn't work to begin with. A program should either be easy to compile and install with minimal fuss or it should be prepackaged binaries to install. gnuCash is neither!

From slackware's end here is a list of required dependencies.

Code:

- libgsf-gnome
  o libbonobo
  o libgnomevfs2
     + gnome-mime-data
     + gconf2
       o ORbit2
- libgnomeui
  o libbonoboui
     + libgnomecanvas
        - libgnome
        - gail
  o gnome-keyring
     + libtasn1
- libgtkhtml3
- goffice
- slib
- swig

WHAT THE HELL!? is this any type of reasonable dependency list for a standard linux system after a full install? NO
some would say then, "Why don't you install gnome and then install it" I don't like gnome for this reason!

This is the reason that applications should NOT use gnome libraries. Gnome is badly designed to begin with and has no business causing so many nonstandard dependencies. Gnome libraries seem in many cases to take existing standard libraries and make pointless interfaces to 'smooth' out the code built with them.

Where was I... gnuCash... too many dependencies. I actually installed all them too (taking great care to choose versions and build scripts compatible with slackware 12.2)! the ./configure goes through without a hitch. Towards the end of the compile it complains

Code:

/usr/lib/libgoffice-0.8.so: undefined reference to `gsf_input_gnomevfs_new'
/usr/lib/libgoffice-0.8.so: undefined reference to `gsf_output_gnomevfs_new'

What the hell! everything is there!

OK OK one could point out the --disable-gui configure flag
That just fails the config step with

Code:

configure: error: conditional "GTKHTML_USES_GTKPRINT" was never defined.
Usually this means the macro was only invoked conditionally.

I'm done with this. I had gotten gnucash working on slackware 11 and 12.0 before but this is too much. I know they originally intended to focus on features instead of fighting with appearance coding so they went with gnome... But I bet it doesn't even save them any time at all! All that stuff is back end and once it's done it just requires tweaking. Look at how much gnome libraries change, they can't even keep the library names consistent and things are constantly being depreciated... More importantly... WHY THE HELL ARE YOU USING GCONF?! this is such a pointless piece of software! there is a reason people don't like the windows registry. We don't need something like it on linux to take over our machines.

I'm sure I could come up with more things to complain about but gnucash angers me too much to care anymore.

Fixing the MBR without loosing data.

lumak — Sun, 14 Dec 2008 20:21:28 GMT

First let me explain my situation. I was playing with Slackware 12.1 and making partitions on a new system, I got lost somewhere withmy partitions and had the crazy idea of dualbooting slackware to have a safe build environment from my system. (I ended up just using the build environment which lead me to wanting a chroot jail instead... ANYWAY)

I screwed things up and had lilo installed on boot records in several different places. How do you fix this then?!

You restore your boot record and make sure to install lilo ON YOUR PARTITION BOOT RECORD.

First thing is you got to remove only the executable code from the MBR.

Code:

# /dev/hda is your harddrive device and NOT a partition
# a partition boot record is 512
# executable code is in the first 446
dd if=/dev/zero of=/dev/hda bs=446 count=1

BUT there is a catch!
This blanks the executable code!
Your bios now doesn't know how to send control from the MBR to the actual partition you want to boot!

a few man minutes later

Code:

# use mbr if the boot partition is on a primary partition
# use ext if the boot partition is on a logical partition
lilo -M /dev/hda mbr
# This is the same as using (c)fdisk to activate a partition
# for booting.  Also note the space.
lilo -A /dev/hda 1

Without this step, even though I had lilo installed onto a partition boot record, nothing was booting. I'm guessing that at the very least there needs to be generic code in the MBR for anything to happen at all. The real question would be, what about windows? I have no clue. I don't dual boot windows anymore. IF I do install windows, it's going to be with a virtual machine. I would rather control windows rather than windows control my machine.

Now you can go about installing lilo properly. Me personally, I have a boot partition /boot just in case I want to boot other linux OSes for testing, I can build a kernel and place it on the boot partition and run lilo again... or chroot my main os and run lilo again.

I also install lilo into the partition boot record of the /boot partition... believe this is done by having

Code:

boot = /dev/hda1

in the lilo.conf

I'm sure I'm misunderstanding something here... BUT this is how I fixed things before installing slackware 12.2 Everything seems to be working now.

Incase you were wondering how badly my system was partitioned before... I had space for 2 or 3 oses on /dev/hda as well as wasted space for an os on /dev/sda. I fixed everything now by blanking /dev/hda, tossing all my saved files there, blanking /dev/sda, fixing partitions for only 1 os (any 'dual' booting will be going on the slower older /dev/hda) then I ran into my booting problem with old lilo reminance, figured out all this crap to fix it, and restored my system with a fresh 12.2! No more random lilo code laying around. (at least I hope it was cleaned off all the other partition boot records. I destroyed all the partition tables and it was 'different' enough that it's probably all gone.) Never figured out how to clean partition boot records. Don't even know if such a thing is possible. Read a better manual. I fixed my problem, that's all that was needed.

Mmmm Fluxbox

lumak — Sun, 14 Dec 2008 07:25:41 GMT

Why didn't anybody tell me fluxbox is everything that blackbox could be with active development. Every feature I tried to add to blackbox my self (basically through scripts) is already in fluxbox and then some!

I tried openbox briefly because the reviews and descriptions made it sound more of what I wanted... BUT I'm sold on fluxbox now. what was the last feature that I found in it that spurred this message? A customizable tool bar + system tray to-boot! It also doesn't freak out when changing the font sizes and shrinking to smaller than it really should go.

Among other great features that I implemented my self for blackbox through scripts was a menu generator and a backbground cache/setup type thing. Both of these are in fluxbox and easy to use!

I also felt like playing with the very pointless transparency features. Also to my amazement, well supported by fluxbox.

Awesome Job Indeed.

What could one complain about? the config files are not xml... BUT they are intuitive and human readable. Additionally, the config files go into .fluxbox instead of .config/fluxbox... not a big deal I guess... I'm sure I can also edit an .xinitrc file to set that up for me... There are a few other things... but I'm still much happier with fluxbox.

Good bye blackbox .70... you will be missed... wait no... all your same features are in fluxbox and then some... You will be remembered.

I was going to post in general... but I figure I'll get just as many comments or people interested in my opinion by posting here.

One Liner Dependency Check for existing binaries

lumak — Thu, 04 Dec 2008 05:15:47 GMT

I ran across this a while ago and was reminded of it recently. I'm placing it here for safe keeping in case I ever need to do a full install of gnome to compile a program just to find the 26 specific dependencies it needs so that everything else can be uninstalled (did I say gnuCash? you lie! it works on all modern OS's... well except for slackware ;)

Code:

#!/bin/sh
#
# Find out all the needed dependencies of a binary

objdump -x $1 | grep NEEDED

USB drive auto mounter script!

lumak — Mon, 01 Dec 2008 05:35:26 GMT

Yes that's right! I kid you not! A marvelous bash script to automatically mount a usb flash drive (or at least set it up to mount it by users) it is commented and should be 'easy' to follow. There are probably better ways to do this, but at the time I didn't want to install anything (e.g. pmount) and I didn't feel like using autofs.

what does the script do if it doesn't automatically mount the drive?
- It determines an appropriate mount name from the file system label or defaults to 'flash'.
- It creates a symbolic link in /dev that points to the mountable partition
- It makes the mount point (e.g. /media/flash)
- It auto increments by 1 if the mountpoint is not empty OR the link is already there (e.g. /media/flash1 (it skips 0))

- IT EDITS FSTAB TO MAKE THE PARTITION MOUNTABLE BY ALL USERS. (bold because this is a security risk and I really didn't want to edit fstab either... oh well)
- It attempts to respect fstab rules by not creating any for a partition if they already exist.

Exciting huh?

It cleans it self up after you eject the device. I did not place any unmounts in the removal so if you don't use 'eject /dev/linkname' or at the very least use 'unmount' manually, you may have some issues. I'm tired of the script for now and just wanted to share it.

SO... here is what to do.
create /etc/udev/rules.d/90-local.rules (or whatever is appropriate for your OS) and place the following.

Code:

# This is messy but I'm no expert on the matter and it works
SUBSYSTEMS=="scsi", KERNEL=="sd*", ACTION=="add", GROUP="plugdev", RUN+="flashhandler.sh $env{DEVNAME} start"

ACTION=="remove", SUBSYSTEMS=="scsi", KERNEL=="sd*", RUN+="flashhandler.sh $env{DEVNAME} stop"

create a file and make it executable. from above, I named mine flashhandler.sh and I placed mine in the 'working/script' directory for udev. On my system (slackware) this is /lib/udev. If you place it some other place, the udev rules file should contain a direct path name to flashhandler.sh

Code:

#!/bin/bash
#
# This is in the public domain!
#
# A 'simple' script to create the a set up to allow users to
# mount their usb drives.

DEVICE=$1
COMMAND=$2

# How odd... most of the commands are in /bin and sym linked
# in /usr/bin. execpt for 'find'.  'find' is only in /usr/bin
PATH=/bin:/usr/bin
DEVDIR=/dev
MNTDIR=/media
CFGDIR=/etc

# just to be clear, put this in fstab format not mount format,
# the '0 0' is part of the line later in the script. You could
# move it here if you wish
FSTABOPTIONS="$ID_FS_TYPE noauto,user,rw,sync,dmask=027,fmask=137"

function fh_start () {
  # If the device is not mounted
  if ! grep $DEVICE $CFGDIR/mtab &> $DEVDIR/null; then
    # Find the file system label to name the link and mount point after
    NAME=`echo $ID_FS_LABEL | tr [:upper:] [:lower:]`
    if [ ! $NAME ]; then
      # Fall back to a default value
      NAME="flash"
    fi
    # By default, we want no suffix
    SUFFIX=
    # Loop until we have a safe link/mount point name.  Hopefully
    # this is done on the first try
    MADELINK=0
    while [ $MADELINK -eq 0 ]; do
      # is the mount point empty? is there a dev sym link?
      if [ ! "$(ls -A $MNTDIR/${NAME}${SUFFIX} 2> $DEVDIR/null)" ] && \
         [ ! "$(find $DEVDIR -name ${NAME}${SUFFIX} 2> $DEVDIR/null)" ]; then
        # Make the sym link
        ( cd $DEVDIR; ln -s `basename $DEVICE` ${NAME}${SUFFIX} )
        # Create the mount point
        mkdir -p $MNTDIR/${NAME}${SUFFIX}
        # Write an entry into /etc/fstab so that users can mount
        # the device as long as the device is not already listed.
        if ! grep ^$DEVICE $CFGDIR/fstab &> $DEVDIR/null && \
           ! grep ^$DEVDIR/${NAME}${SUFFIX} $CFGDIR/fstab &> $DEVDIR/null; then
           FSTABENTRY="$DEVDIR/${NAME}${SUFFIX} $MNTDIR/${NAME}${SUFFIX} $FSTABOPTIONS 0 0"
           echo $FSTABENTRY >> $CFGDIR/fstab
        fi
        
	# Do not mount the device here.  It will be owned by
        # root:root. Only mount it here if you specify
        # uid=root,uid=somegroup" along with permissions for the
        # group to do what you want them to.
	# if grep $MNTDIR/${NAME}${SUFFIX} $CFGDIR/fstab &> $DEVDIR/null; then
        #   mount $MNTDIR/${NAME}${SUFFIX}
        # fi

        # Exit the loop
        MADELINK=1
      else
        # To make the sym link and mount point match, we need to
        # increment it and try again.  This is a dirty solution and
        # really should have a better exit stratigy.  Additionally,
        # we do no error checking for people that name their
        # partitions drives after stupid things that already
        # appear in the /dev directory.  e.g. sd, fd, cdrom, etc.
        let SUFFIX++
      fi
    done
  fi
}

function fh_stop () {
  # The device unmounting is handeled by 'eject'.  A safe check
  # should probably be placed here just to make sure it really
  # is unmounted.  Have the script check /etc/mtab.
  # Since there is a possibility of multiple links to the device,
  # we should loop through all the links and undo everything.
  # You should probably do some error checking to make sure the
  # file is there for you to remove it or something of the such
  BASEDEV=`basename $DEVICE`
  for LINKFILE in `find $DEVDIR -lname $BASEDEV 2> $DEVDIR/null`; do
    NAME=`basename $LINKFILE`

    # Remove the sym link and mount point
    rm -f $LINKFILE
    rm -f $MNTDIR/$NAME

    # Remove our exact fstab entry that we put in before.  This is
    # sloppy. We don't actually know if we definitely added a line
    # in or not.  Maybe some sort of log file is needed to check
    # this first.
    FSTABENTRY="$LINKFILE $MNTDIR/$NAME $FSTABOPTIONS 0 0"
    sed -i -e "\@$FSTABENTRY@d" $CFGDIR/fstab
  done
}

# All scsi devices get passed into this script by udev. we are only
# interested in those connected through usb and that are
# filesystems... Hopefully this works when the filesystem is on the
# main device itself and not on a partiton.  e.g. /dev/sda instead of
# /dev/sda1.

# This should actually be placed into the udev rule to avoid
# confusion at the source and avoid running the script in the first
# place.
if [ $ID_BUS == "usb" ] && [ $ID_FS_USAGE == "filesystem" ]; then
  case $COMMAND in
    start)
      fh_start
    ;;
  
    stop)
      fh_stop
    ;;
  
    *)
      echo "This script should be run by udev rules only"
    ;;
  esac
fi

I tested this with my ipod nano (hey it was free) and my flash drive. The ipod was properly set up for sd?2 and the flash drive was properly set up for sd?1. I did not test different methods. bugs, tips, experience with this script would be appreciated; just leave a comment.

Some things I have not tested:
- make a hole flash drive the file system with no partitions (e.g. /dev/sda)
- make a flash drive with multiple partitions of various file system types.
- Test external hard drives (or make the script support it)

Check back later for when I post a blog entry on getting autofs to do something similar.

A script to extract RPMs

lumak — Sun, 30 Nov 2008 05:44:10 GMT

Since my blog is a guide to random things... here is a nice bash script to extract RPMs to a directory named after them self where the RPM can be located anywhere... that is, relative to the current directory or exact path

Code:

#!/bin/sh
#
# This script is in the public domain.
#
# A simple script to extract RPM files automagically
# Honestly... why isn't this feature built into
# the rpm tools?! unless I missed something...

RPM="$1"
CWD=`pwd`

DESTDIR="$CWD/"`basename $RPM .rpm`
mkdir "$DESTDIR"
cd "$DESTDIR"
if [ -f "$RPM" ]; then
  FILE="$RPM"
elif [ -f "$CWD/$RPM" ]; then
  FILE="$CWD/$RPM"
fi

rpm2cpio $FILE | cpio -idmv

NAT, SCSI-changer, Plasmon D Series, and Slackware

lumak — Thu, 20 Nov 2008 06:19:34 GMT

UPDATE: new blog note on how I somewhat got it working!

To prove the power of open source and linux to my roommate, I'm going to get his Plasmon D Series dvd robot working under linux as a NAT box that will not only be accessible from windows but hopefully the PS3 and capable of trans coding incompatible files. I have mostly been researching the topic and I believe I have all the software... Then why isn't it done yet? configuring... and figuring out the damn system. Though this is a completely new thing to me (well not the linux half) I have faith I'll get it all put together. At which point I will back up the hardrive/setup, make donations to the major components, and make a YouTube video of it all working on marvelous harmony.

Considering the software from plasmon to do all this on windows is about $5000 I good comparable donation would have to be in order. Anyway... I'll probably post progress and a guide here and make available configuration files and startup scripts where possible.

Currently I've been playing with getting autofs (con)figured out and setting up a packaging environment for and necessary start up scripts and any extra support scripts I may write. The next step is getting all the other support software that is required by dvd-vault (osvault/optivault/openfiler derivative/whatever the hell it wants to be called) and configuring it on slackware.

Why slackware you ask?
* It's a clean base system.
* I'll learn more of what goes into it.
* slackware startup scripts are extremely easy to follow
* why not?
* wouldn't you?
* I'd rather make a donation to slackware than most other distros
* etc.

OK that's enough for now.

LinuxQuestions.org - Blogs - Lumak's Guide to Random Things by lumak

Gnome 3 Tip - Quick Application Search

More playing with BtrFS on Slackware

Playing with BtrFs and thinking out loud.

HOWTO: LAMP server on Slackware

UGH! I hate broken configure scripts/pkg-config setups (cairo-compmgr/vala)

Fluxbox - Display Message on Desktop

Window Manager Ponderings

LibreOffice back at RC1

CSS small caps

T This Is Small Caps

People that complain about KDE4

Tar progress bar for scripts

Slacklet - a new OS Mod Proposal

having fun with bash varaibles: indirect variable references

Working Scsi Robot DVD Library! Yeah!

Linux on a Tablet PC - which way to go.

slackware64 multilib QT issues

Slackware64 nvidia-driver-compat32

man to pdf converter

Fluxbox/Openbox - calendar in menu

Fluxbox - Terminal Window PC Game Style

hplip hpaio sane setup

blender x86_64 compile guide for slackware64-pre-13.0-current

Slackware Undocumented 'Well Duh' Quark #2917 - console vim and X11 clipboard

Compiling with new static libs

THANK "BOB"!

The pointlessness of gnuCash

Fixing the MBR without loosing data.

Mmmm Fluxbox

One Liner Dependency Check for existing binaries

USB drive auto mounter script!

A script to extract RPMs

NAT, SCSI-changer, Plasmon D Series, and Slackware