I should also say that this is all legacy content being shoehorned into “responsive” courseware. So there are width restrictions now, when before these activities were just popped up in new windows with “100%” widths.

Anyway, when trying to place said swfs into the product, the embed codes for an entire disciple were completely wrong. Things squished, white spaced abounded. Turns out the person who loaded the original metadata set all the widths to the max width for the product (595 px) and then grabbed heights from the questionable html wrappers. But they didn’t in turn adjust this grabbed height in light of a new width. Oooooof. A complete mess!

I needed to start from scratch, but I was looking at several hundred swfs. Grepping the dimensions from the html was out, as they were suspect. As was opening each in Flash, I refuse to do manual labor! I needed a csv with all the filenames and native dimensions. If the width is over 595px, I needed an aspect ratio calculation done.

So swfdump and bash to the rescue!

swfdump is part of the mighty SWFTools library. Still useful as ever even as Flash is slowing dying. Plenty of mileage left as we need to decompile and remake things in HTML5. But back to the subject at hand…

swfdump prints out various information about SWFs, like contained images/fonts/sounds, disassembly of contained code as well as cross-reference and bounding box data.
Usage:

./swfdump [-atpdu] file.swf

Full options here

And here’s my script, should be easy enough to modify to grab other metadata:

shopt -s nullglob
## If there are no files matching *swf then the shell will expand *swf to exactly that, the string "*swf", unless shopt nullglob is set.
for f in *swf
do
echo -ne "$f," >> log.csv ## get the filename for our csv
'/cygdrive/c/Program Files/SWFTools/swfdump.exe' "$f" -XY | awk '{
## cuz awk is meant for processing column-oriented text data
## The variables $1, $2, and so forth are the contents of the first, second, etc. column of the current input line.
if ( $2 > 595 )
{	nwidth = 595;
	nheight = $4/$2*nwidth;
	print nwidth","nheight",ADJUSTED";
}
## perform aspect ratio conversion if needed and make a note in our csv
else
	print $2","$4;
}' >> log.csv
done

Hope this helps!

The exception being the mighty Notepad++. So use that if you’re afraid of the CLI.

Number constructor called as a function simply performs type conversion:

// type conversion
Number("911");        // 911
Number("20px");       // NaN
Number("2e1");        // 20, exponential notation
Number("010");        // 10 (The Number constructor doesn't detect octals)
Number("8,569")       // Nan
Number(8,569)         // 8
Number("8.569")       // 8.569
Number("0xF");        // 15 (But it can handle numbers in hexadecimal notation)
Number("false");      // NaN
Number(true)          // 1

parseInt() performs parsing into an integer*:

// parsing:
parseInt("999");        // 999
parseInt("20px");       // 20 (will stop parsing and drop any part of the string it can't figure out)
parseInt("fi5e")        // NaN (string has to start making some sense right away)
parseInt("10100", 2);   // 20
parseInt("2e1");        // 2
parseInt("010");        // 8 (if parseInt detects a leading zero on the string, it will parse the number in octal base)
parseInt("010", 10);    // 10 (decimal radix used)
parseInt("0xF");        // 15 (can also handle numbers in hexadecimal notation
parseInt("22.5")        // 22 (because the decimal point is an invalid character for an integer)
Number("false");        // NaN
Number(true)            // NaN

*It’s always a good idea to supply a radix to parseInt(value, radix) that way you don’t have accidental octal mode conversions.

The parseFloat() method works in a similar way to parseInt() however the string must represent a floating-point number in decimal form, not octal or hexadecimal.

parseFloat("1234blue"); 	// 1234.0
parseFloat("0xA"); 	// NaN
parseFloat("22.5"); 	// 22.5
parseFloat("22.34.5"); 	// 22.34
parseFloat("0908"); 	// 908
parseFloat("James1980"); 	// NaN

The Unary Plus method (+num) is basically a short cut for casting to a number via Number(). Just a nice shortcut really.

+"911"; // 911
+"20px";       // NaN
//etc etc
num = 1 + +"2"; // num is assigned the value 3
+true; // 1 (just saying...)

Finally, you can attempt some simple math on your string, does the same as the above.

"123" / 1; // 123
"123.987" * 1; // 123.987

You can use num / 1;, num * 1;, 0 + num; or 1 * num;.

If you’re interested in a speed comparison, see this test case (this would only matter for huge numbers of conversions. HUGE.)

I’m also using W3 Total Cache, which appends the param string like “?9d7bd4″ to all my images. This makes sure that if you change your browser cache policies all users immediately see that change.

However, apparently the appended image string allows for the images to be clicked on and stranded. Interferes with whatever is stripping the image links.

Best way to avoid this:

Go to “Browser Cache.”

At the bottom, you have checked the following setting:

Prevent caching of objects after settings change
Whenever settings are changed, a new query string will be generated and appended to objects allowing the new policy to be applied.

This adds an appropriate string to the end of all images called through CDN or otherwise. Uncheck that box and purge/refresh the cache to see the difference.

Yay. Be stranded no more.

No, no, no. Don’t use ls. Use find.

find . -type f

Get certain file types:

find . -name \*.txt

Even better, if you have it, is tree. So boss. In fact, install this now if you don’t have it. Check it.

tree -if .
tree -if directory/

For just files:

tree -if | grep -v \>

Of course, grep out any certain file types if you want.

chown -R bitnami:daemon /opt/bitnami/apps/wordpress/htdocs
find /opt/bitnami/apps/wordpress/htdocs -type f | xargs chmod 664
find /opt/bitnami/apps/wordpress/htdocs -type d | xargs chmod 775
chmod 664 /opt/bitnami/apps/wordpress/htdocs/.htaccess
chmod 664 /opt/bitnami/apps/wordpress/htdocs/wp-config.php

apc.shm_size=64

to

apc.shm_size=64M

I.e. specify “M” for MB or “G” for GB to solve this issue.

1) Go in Terminal (in Apps > Add a local App > Terminal), and copy/paste :

wget -q http://apt.last.fm/last.fm.repo.gpg -O- | sudo apt-key add -

2) Then type :

sudo gedit /etc/apt/sources.list.d/ubuntu.list

3) A text file just opened. At the end of the document, add a new line, and paste :

deb http://apt.last.fm/ debian stable

Then save!

4) Go back in Terminal, and copy/paste :

sudo apt-get update

and

sudo apt-get install lastfm

So the laptop is old and slow, and we don’t really watch movies or TV or play video games. So these heavy duty media server installs are too much, clearly. We want to play music, and maybe be able to hop onto allmusic.com or wikipedia for some fact-checking. We want to use mainly web-based streaming apps like MOG, Spotify and last.fm, be able to listen to NPR and other podcasts, watch/listen to stuff on youtube, and play from our iTunes libraries remotely. Hmm, barring that last requirement, doesn’t it just sound like what I really need is a Chromebook (Google’s netbook? Say, can I install Chrome OS on my machine? No, but you can install Chromium OS and it might not work amazingly well. Read through that post and its one of those that keeps us all afraid of Linux. But check the comments and enter Jolicloud OS!

Joli OS is built on top of Ubuntu, tweaked for netbooks and other computers with limited disk storage, memory, and screen size. Joli OS was designed for easy installation, with Wi-Fi, Bluetooth, and 3G modem support all included. Jolicloud claims the OS supports 98% of netbooks with out-of-the-box compatibility but also works on a very large number of other devices, up to 10 years old: laptops, desktops and tablets. Bottom line is it just worked for me.

The natural comparison in look and function is iOS, the operating system used by the Apple iPad/iPhone, though with the folders and files and CLI of a conventional Linux install tucked deep away. The user interface is built primarily in HTML5 that includes an application launcher and a library of compatible applications with one-click installation and removal.

Boxee, Dropbox, VLC, Skype and Google Chrome are all featured as one-click installs. So are Gmail, Google Maps, Wikipedia, MOG and Facebook. Many of the applications available to Jolicloud users are, in fact, websites but really, of course, better thought of as applications. Open any of these up, and you get the website, e.g. Gmail, as program in a heavily tweaked full-screen browser, minus most of the browser toolbars and menus. You don’t need that browser stuff though for these web apps, so when you browse, you can just browser and not waste your tabs on the web apps.

The distinction is further broken down in your one-click dashboard apps (a mixture of mostly web-apps and some local programs) and “native” apps like gedit, terminal, gnome settings, etc. If you know what you’re doing you can add whatever to your installation and put shortcuts in your “local apps” folder on you application launcher/dash, but not your dash proper, which is annoying. For example, if you install last.fm as an app through Joli, it just gives you the web app, whilst there is a perfectly good linux port of the superior last.fm scrobbler program. So I added it through apt and I wish I could get it directly onto the main launching area rather than having to add it as a shortcut in a folder on my main area. But there are some workarounds like adding it to the little upper “taskbar” and I don’t foresee adding many other programs, since Joli makes intelligent choices about whether to give me the web app or real app otherwise.

Anyway, blathering on. Some tips and tricks for this setup soon!