Category Archives: Script

Get down with getopt

What is getopt?

getopt is a command used in scripts to parse their options and add a basic error checking ability. getopt is not getopts the bash built-in that has similiar functionality. Urban Vagabond explains:

getopt and getopts are different beasts, and people seem to have a bit of misunderstanding of what they do. getopts is a built-in command to bash [that] processes command-line options in a loop and assigns each found option and value in turn to built-in variables [(so that they can be further processed)]. getopt, however, is an external utility program, and it doesn’t actually process your options for you the way that [(e.g.)] bash getopts, the Perl Getopt module, or the Python optparse/argparse modules do. All that getopt does is canonicalize the options that are passed in — (i.e. convert them to a more standard form) so that it’s easier for a shell script to process them.

For example, a use of getopt converts the following:

tmpscript -abd -ooutfile.txt

into this:

tmpscript -a -b -d -o outfile.txt

getopt can also process the long format option of --output=/tmp.... Also basic error check abilities, it has:

tmpscript -c
tmpscript: invalid option -- 'c'

or

tmpscript -o
tmpscript: option requires and argument -- 'o'

getopt gets defined by telling it the available options in short and long form (if you like). An option with a colon (:) following denotes that an argument is required for that option. If followed by two colons the argument is optional. Here is an example:

Options are generally set as variables so that they can defined how to be used after the while loop (usually, so that mainly they are processed in the right order). getopt parses in the same order as specified on the input, so a -a -d -b input would not work real well if the -d option required the -b option to run correctly.

One caveat of getopt is that each option allows either one or no arguments to follow. So in input like tmpscript -v --files file1.txt file2.txt file3.txt just isn’t possible in any type of predictable-fashion. One can however use quotes on the input:

tmpscript -v --files "file1.txt file2.txt"

or use mutiple options (-f file1.txt -f file2.txt) and append them to an array, as seen, here.

Numbered file backups – Easy as can beasy

file-twoWhen I’m about to edit a configuration file or a bash script, sometimes I don’t know if the edit will work. In these cases I create a backup of it in numbered order (to keep track of them if I need to revert the changes). Because this is something I regularly do, I decided to build a script to make this easy. The usage is basic: define the file and destination-directory:

# bckfile definedfile.bash .
‘/home/johndoe/.local/bin/definedfile.bash’ -> ‘/home/johndoe/.local/bin/definedfile_00.bash’

Optionally a tag can be added too:

bckfile youdle .vault/ new
‘/home/johndoe/.local/bin/youdle’ -> ‘/home/johndoe/.local/bin/youdle_02-new’

A limitation of the script is that filename cannot have ‘periods’ in their filename. A heuristics of discovering multiple extensions and with . in the filename itself would be tough.

gt – script to create a temp file in gedit

posticon

I am really lazy with my editors. I have aliases in my shell configuration for gedit and vim that are very basic:

alias v="vim -p" # open in tab
alias sv="sudo vim -p"
alias g="bgcmd gedit"
alias sg="bgcmd gksudo gedit"

This is very nice for me because I use my editors quite a bit. One thing I needed though was a command that would create and open a temporary file in gedit. The main reason for this is that at the time I don’t know how to name or place the file properly. Another reason is that at other times I like to have a scratchpad but would lose information if it was just a New File and a crash… happened.

Running just gt will create and open a file named of the current time (MMDDhhmm) and will be saved in the trash folder. If gt is followed by a name (e.g. gt cssbox or cssbox.css) the name will be appended to the current time.

Screenshot from 2013-08-26 08:17:12

The name is helpful if wanting to dig the file out of the trash folder at a later time.

xuserrun – Run a command on the currently-active user’s X.org server display

xuserrun is a bash script to run a command on the active X.org server display.  This is primarily necessary if from within another environment  (different user, console, cron, boot script…).  xuserrun gathers DISPLAY and user environmental variables via systemd and passes them on to be able to dothis. xuserrun is designed for use with only a single user X.org server display. Running it is basic:

xuserrun xclock -digital

Tto put out a notification:

xuserrun notify-send "Hello, Dave."

It’s available for Arch users in the AUR and is also downloadable on github.

No-think link (redone)

When I first used ln, I expected the behavior of ln source link to work. But ln is slightly more complicated than that. While learning it from the beginning, I discovered it does not always resolve the path correctly; additionally, it doesn’t not error if the source is non-existence. I discovered that this was necessary because ln is designed to be versatile. So I understood that one would want links at times to be created that don’t exist yet, but it was unhelpful to me in the way that I felt ln should logically work, and the way that I wanted to use it.

Originally, I would do a link like this and it would be broken:

# ln -s ../dir/file dir/

I’ve attempted other links since then that too (of which I can’t remember now) would also be broken. So the way I discovered to do links correctly without doubt was to use whats called the fully-qualified or absolute path (i.e. a path defined from root [/]):

# ln -s /home/username/Document/file /home/username/Desktop/file

There’s a program called realpath that does this, so I wrote a wrapper script for it and it turned out to be pretty easy. The script resolves the absolute paths, verifies the existence of the source directory and link parent directory, and detects write permissions prompting sudo when needed. It works like this:

# lnk ~/Documents/Resumes/ ~/Desktop
‘/home/username/Documents/Resumes’ -> ‘/home/username/Desktop/Resumes’

By doing:

# lnk ~/Documents/Resumes/ ~/Desktop/Resumes-I-like
‘/home/username/Documents/Resumes’ -> ‘/home/username/Desktop/Resumes-I-like’

a specifically named link will be put on the Desktop.

# lnk /etc/fstab  ~  # or use ~/fstab-edit
‘/etc/fstab’ -> ‘/home/todd/fstab’

If trying to link to a write-protected directory:

# lnk /usr/share/vim/vim73/vimrc_example.vim /etc/vimrc
[sudo] password for $USER:
‘/usr/share/vim/vim73/vimrc_example.vim’ -> ‘/etc/vimrc’

An error message will appear if the source or link parent directory doesn’t exist:

Non-exist: .bashr

Limitations:

* The script will work for just about any instance with the exception of removable media where relative-paths would be better used.

External monitor as Discrete

I use my laptop primarly at home with an external monitor as discrete, meaning that I have the laptop monitor turned off and I only use it. At times this is also called a dedicated monitor. GNOME can be set to disable the laptop monitor and enable the external but it wasn’t able to hotplug the monitor after I returned the laptop, and at times wouldn’t do so after resuming from sleep. Also in the proccess I discovered that the X.org server DPI setting wasn’t being done correctly and that GNOME’s text scaling needed to be adjusted. So I decided to do it in a script and it turned out to be pretty easy.

I wrote the basic script that toggles monitors depending if the external monitor is present, then it detects correct physical size dimension of the screen so the the correct DPI can be set. After this, I added a startup script (.desktop file), a pm-utils script to runafter resuming, and a udev script to detect andset the monitor when plugged in. The udev rule is generic but appears to be working for a lot of people, it relys on Kernel Mode setting (KMS) so doesn’t work for me wiht the catalyst driver, but every thing else works great. I put it on github for any who like to look at it.

The bash script cannot be used right away instead a couple bit will need to be directed:

 The package cannot be installed directly and be expected to work, some edits
 will need to be made.  First, in the resume script '80_discretemon' a username 
 will need to be defined; next, the monitor names will need to be defined as
 created by the driver in 'discretemon'.

Also, the monitors can be defined in xorg.conf but the fix for after resume from sleep, remains.

Section "Monitor"
  Identifier  "0-LVDS"
  Option      "VendorName" "ATI Proprietary Driver"
  Option      "ModelName"  "Acer Aspire Laptop Screen"
  Option      "DPMS" "true"
  Option      "TargetRefresh" "60"
  Option      "Position" "0 0"
  Option      "Rotate" "normal"
  Option      "Disable" "true"
  DisplaySize  344 194 # only works with xrandr disabled.

EndSection

Section "Monitor"
  Identifier  "0-DFP1"
  Option      "VendorName" "ATI Proprietary Driver"
  Option      "ModelName"  "Samsung SyncMaster SA350"
  Option      "DPMS" "true"
  Option      "PreferredMode" "1920x1080"
  Option      "TargetRefresh" "60"
  Option      "Position" "0 0"
  Option      "Rotate" "normal"
  Option      "Disable" "false"
  DisplaySize  476 268 # only works with xrandr disabled.
  Option      "DPI" "102 x 102"
EndSection

Screencasting Done Easy (Desktop Recording)

I got to test out a good number of screencasting applications and I found a good one, and as usual the easiest was the best. I started with recordMyDesktop.

recordMyDesktop

recordMyDesktop is a basic program that works good. The GTK version has a simple UI that sets a border around the area to be recorded. I has sound recording too.

A minor thing but of note is that the window detection area is off when selecting a windows, but the reason I didn’t use recordMyDesktop was because I found the quality wasn’t that good. It could be because it uses .ogv format, or perhaps it had something to do with my system.

This is and example I did with recordMyDesktop and though it’s enlarged (OpenShot doesn’t have the ability to use the original size) the quality I wanted to be better.

Others

I tried Istanbul and a couple others all with about the same recording results. Istanbul hasn’t been developed in several years and though I got excited about xvidcap it hasn’t been developed in years either. xvidcap grabs screenshots and then concatenates them into a video. I got excited because xvidcap’s preview uses Imagemagick’s animate tool to preview the video and it was real nice. Unfortunately very little works in xvidcap anymore but taking the screenshots. To put them together I used:

fmpeg -i out%04d.xwd -r 15 -vcodec huffyuv test.avi

unfortunately the quality was no better than that of the others.

FFmpeg

The great command line tool to encode and decode video ffmpeg can also do screencasts and I read a lot of how people liked it (and I do too). To use it it’s real basic:

ffmpeg -f x11grab -s wxga -i :0.0 -sameq screencast.mpg

The quality isn’t quite what I want it to be, but I’ve seen other people have nice looking screencasts so I think it must be either my video card or my video driver.

This line can be amended some for better quality, performance, and add sound recording. Using the raw, lossless codecs for video and audio improves processor usage for better FPS recording:

ffmpeg -f x11grab -s wxga -i :0.0 -vcodec huffyuv -sameq -acodec pcm_s16le -f alsa -i pulse -ac 2 screencast.avi
  • -s and -i are for size and input. -s will give the dimensions and -i will define the co-ordinates. wxga is a definition of a video resolution standard (available ones are listed in man ffmpeg)
  • -r can be added to define the frame rate. Default is 25 and is good. Only reason really to change it is if frames are dropped during recording (marked with red).
  • -follow_mouse 100 can be added to follow mouse movements. 100 is the border in pixels that must be reached before the area is moved.

ffcast and FFmpeg

ffcast is a program that grabs and passes X.org server dimensions and co-rodinates to other programs. It has built-in support to pass these parameters for some programs including ffmpeg. So the command will now look like this:

ffcast -s ffmpeg -- -vcodec huffyuv -sameq -acodec pcm_s16le -f alsa -i pulse -ac 2 screencast.avi

ffcast’s -s option will prompt for the screen area and then pass the dimensions and co-orodinates to ffmpeg using --.

Now to make this easy, I put this in a bash script, it runs as such:

 screencast <a|f|m|w> - create screencasts (a)rea (f)ull-screen (m)ouse (w)indow

Here’s the bash script:

An example:

Disk and Memory Usage Scripts

At times I like to check my levels of disk and memory usage and it’s more convenient for me to do it from the command line. So, I created a couple scripts for them:

devtop 
 Filesystem      Size  Used Avail Use% Mounted on
 /dev/sda5       9.8G  6.4G  2.9G  69% /
 /dev/sda6       166G   38G  121G  24% /home
memtop 
 PROGRAM                   %MEM    #MEM
 firefox                   10.2    352.98 MB
 gnome-shell               4.1     141.76 MB
 Xorg                      1.5     53.60 MB
 nautilus                  1.2     41.52 MB
 gedit                     1.1     40.59 MB
 gnome-settings-           0.7     26.23 MB
 gnome-terminal            0.6     22.31 MB
 nm-applet                 0.6     21.30 MB
 python2                   0.6     20.89 MB

Saves me a lot of time over having to open a program :).

Convert Videos to PSP



I’m a TED video junkie. I always have videos on my PSP ready to watch. I also like to put YouTube videos on there. I did this enough that I created a script for it that makes putting videos on my PSP real easy:

 pspvidconv <d*> <video(s)> - Convert videos to PSP (d to use directory)

The PSP allows use of a single-depth directory. The directory option (when using d flag) will ask if the user wants to create a new directory, if the answer is no, it will present the existing ones.

Warning: Currently h264 encoding isn’t working. The PSP will report that it is an unknown codec, so mp3g4-xvid is the only option with ffmpeg.

h264enc

Because I’ve found that options and settings change frequently with encoding tools, it is better to have an expert be able to handle them (otherwise, I will spend more time looking options up again). A good program to use is h264enc. It’s a shell script (perl, I believe) and well done; not good for many files as all settings will have to be re-entered but does a good job.

Handbrake

For Handbrake GUI I found this post. I have yet to find any handbrake-cli lines that work.

Markdown to WordPress Format

When I discovered MarkDown I was in love, it was very nice to discover an easy, clean, well-thought-out markup language. I wanted a way to be able take my detailed notes, have them easy to read, and then to be able post them here on the blog that would be in a nice WordPress format. Yesterday, I wrote about discovering pandoc which enables a person to write in markdown and have it converted to HTML. To be able to use the HTML code created by pandoc in WordPress it needed to be slightly edited and cleaned up. Therefore,… I created a bash script for it.

Example conversion from MarkDown format to WordPress format

It works pretty good. Basically it removes tags for paragraphs (<p>), cleans up code blocks, improves formating, as well as doing a few other things. I haven’t gone through all permutations that pandoc conversion can do so a few thing swill be left out, but otherwise it should work for most things. If anything needs to be added please let me know.

Download Package Source Files

I wanted a way to easily download source files for quick editing. Sure abs and numerous AUR-helpers do it but I wanted a simple all-together program/script that could do it, and it turned out to be pretty easy. I wanted to be able to define the repository, and then the package and I did it!

pacpull extra/a52dec

aur/ can be used here as well as the other repositories. I set up /etc/abs.conf and $XDG_CONFIG_HOME/cower/config to use my pkgbuilds directory ~/.arch/pkgbuild (ABSROOT="/home/todd/.arch/pkgbuild", TargetDir = /home/todd/.arch/pkgbuild/aur respectively) so it looks pretty nice:

ls
aur  extra  own  testing

If the repository is wrong it will list the available repositories:

pacpull extar/a52dec
 Not a valid repository:
  core
  extra
  community
  multilib
  testing
  community-testing
  multilib-testing

And it’s nice and it’s basic:

Follow

Get every new post delivered to your Inbox.

Join 52 other followers