Add Ligatures

I've consulted a few sources (listed below) and this is what I can gather:

• In general, long s is used initially and medially
• In general, short s is used at the end of a word
• I can't find a pattern in how long s behaves in ligatures
• There seem to be different conventions for different languages

Perhaps you can clarify the rules?

http://en.wikipedia.org/wiki/Long_s
http://babelstone.blogspot.com/2006/06/rules-for-long-s.html
http://babelstone.blogspot.com/2006/07/long-and-short-of-letter-s.html
http://www.straightdope.com/classics/a1_110.html

For reference:

A ligature hack on php smartypants
http://pages.cs.wisc.edu/~wolfson/ligatures.html

Fancypants implements ligatures in Scheme
http://www.jeroenvandoorn.nl/index.php/2006/12/20/ligatures-in-fonts

Added preliminary support for ligatures. Please review code or install and test!

Does it work?
Does it break anything?
Are there any other ligatures you'd like to see? (I'd like to see "Th", but I can't figure out where it is in unicode)

Medial S support coming soon...

Code is available on CVS now, and should be available from the project page when the packaging script gets around to it, sometime tomorrow (July 10, 2007).

Features can be configured in an admin panel, it's just really hard to find!

First, go to Admin -> Site Configuration -> Input formats

Then, click on the configure link next to the Input format that Typogrify is a part of.

Then, click on the Configure tab at the top of the admin panel (the options are View, Configure, and Rearrange). (The URL should look like www.yoursite.com/admin/settings/filters/#/configure, where # is the filter number you're editing.)

There you have the option to turn each typogrify filter on or off, and to enable ligatures and arrows (they're off by default!).

Ligature support varies from font to font and browser to browser... good luck!

That foreach bug has something to do with the ligature system, but I can't figure out what. Let me poke around my test installation to see if I can figure it out.

Here's a shot in the dark: is the file unicode-conversion.php in the modules/typogrify directory?

Thanks, fm, for your thorough review! Here are my thoughts:

The foreach() Bug

Still hunting... I'll let you know. Can you tell me anything more about it? Do you always get 6 repitions of the error?

Ligatures and Arrows

I'd say this feature isn't broken, but to be on the safe side, I'll leave those features off by default. Here's why:

I think browser support for Ligatures and Arrows is spotty. For example, the ft ligature and arrows all display properly on Safari on the Mac (I think it might be in some version of Caslon -- it looks weird). There may be some subset of ligatures that display in all browsers, and maybe those could be turned on by default?

(In my opinion, it's best to use the ligatures and arrows only in situations where you know a lot about the browser that will be used to access the site. I actually built the feature for use on in-gallery kiosks at my museum -- I know they'll be using Safari and they'll have Adobe Garamond Pro installed!)

(Oh, and there's the search engine friendliness issue -- does google see an fi ligature as exactly the same thing as an ascii fi character pair?)

Amp Wrapping

I think I misrepresented the feature in the project page. I'll fix that! Here's what actually happens:

The module currently wraps all ampersands in the amp class. The style sheet uses the fancy ampersand only within <h1> - <h6> tags.

Unfortunately, Drupal makes it difficult to apply filters to headlines. To get showy ampersands in post titles, you'll need to modify your theme.

Other Typogrify Features

Glad to hear they seem to be working properly! That widow management thing is pretty subtle... you'll notice that you never see one word alone at the end of a paragraph, it'll always be at least two words.

Thanks again for all your help! I'll let you know when I get that medial S thing working.

Sorry you've become the guinea pig! I can't for the life of me reproduce the bug on my personal machine, but here's another shot at a fix.

I know it has something to do with the arrays of ligatures and arrows passed into the foreach loops() there on lines 100 and 108. So I've added some code that double checks that those arrays are (1) actually arrays and (2) not empty.

The changed code looks like this (this code replaces lines 99 - 113 of typogrify.module):

// Build a list of ligatures to convert
if ((is_array($ligature_map)) && (count($ligature_map) > 0)) {
foreach ($ligature_map as $pair => $ligature) {
$setting = 'typogrify_use_' . $pair . '_ligature_' . $format;
if (variable_get($setting, 0) == 1) {
$characters_to_convert[] = $pair;
}
}
}

// Build a list of arrows to convert
if ((is_array($arrow_map)) && (count($arrow_map) > 0)) {
foreach ($arrow_map as $ascii => $unicode) {
$setting = 'typogrify_use_unicode_for_' . $ascii . '_' . $format;
if (variable_get($setting, 0) == 1) {
$characters_to_convert[] = $ascii;
}
}
}

If you get a chance to test this code, that would be great! (I've removed indenting because it can cause a weird bug when you copy and paste code out of safari and into your text editor.) Let me know how it goes. Fingers crossed!

I've attempted to fix this by making the array that the foreach is traversing into a global, and by double checking that the arrays are indeed what foreach is expecting them to be. Since I can't reproduce the bug, I can't be sure it's squashed. Let me know if this works.

The new code is available in CVS right now; should be available through the project page when the packaging script gets around to it.