I'll send this info to Ben though. Perhaps it will be fixed in 2.61.

• HTML Entities by name do not work. There is a comprehensive list of named HTML entities at http://www.w3.org/TR/html401/sgml/entities.html . When any of these are encountered, the & is converted to & - this is disturbing when you encounter languages that need to use these, such as « («) and » (») in French and Norwegian.
• Conversion of national characters to HTML Entities: This should really be optional. I author with ISO-8859-1 and UTF-8 because I do not neccesarily want my documents littered with ås and øes.
• If *strong* and /emphasis/ occurs at the beginning of a new line, they are interpreted literally instead of being converted to their markup equivalent.

1) Try as I might, I can't get the cite tag to work. It just shows up as double question marks on either side of the text, even in the published entry.

2) I can't get lists to work either. They do seem to work in the MT preview screen, but not in the published entry. Maybe it's something screwy in my css? But HTML-produced numbered and un-numbered lists seem to work fine, I think. Hmm.

The URL discovery code misses a few of the legal top level domains. IANA currently lists, in addition to the ones textile.pm catches: .aero, .coop, .museum, .name, and .pro. While few of these are heavily used, they're all in use. http://mos.boston.museum/ works, for example.

I kinda doubt this needs to be fixed, but I thought I'd point it out in case you're as anal about domain names as I am.

Again, thanks for the great contribution.

So I made a small change in textile.pm, line 177: adding a single dash to the character class [^\s], so that it now reads [^\s-] on both occurences.
I hope this is (a) allowed, (b) legible and (c) useful. I can't get the whol eline to display here, but will send it to you privately if you'd like.

One bug report:

The URL wrapper appears too aggressive. In a line with both "Quoted Text" and a "URL":http://www.bradchoate.com/ the anchor appears before the first open-quote, not the one closest to the URL...

O...

*this is a test*

*this is* a test

this is *a test*

this *is a* test

Apart from that, looks pretty groovy.

I upgraded to MT 2.6 and installed this plug-in alongside the already excellent "Smartypants" plug-in I'd been using already. Went into the preferences for MT and enabled it for both the entries and for my comments. Tried to do a full site rebuild and....

... I got an Internal 500 error. A little investigation reveals it only errors when trying to rebuild my individual entries archives. Monthly, Category, and all indexes rebuild just fine. Yet when I do a new entry the individual archive page is created without issue. Turning off text formatting in the preferences allows me to once again re-build all of my individual entries. Turning on Textile in my entries alone and leaving Comments set to line breaks also works, but the moment I tell MT to use Textile on comments it gives me a 500 error when trying to rebuild all of the individual archives.

Don't know why, but I can reproduce it at will. Let me know if you come up with some suggested fixes.

Also, to the person who suggested a UBB Code version of Textile: I think that's an excellent idea and it shouldn't be too difficult to do using a variant on Textile.

You might want to update your documentation.

But I would have got it running a lot quicker if there have been a nice prominent comment in the Installation section saying that you have to set the Test Formatting option in the entry (and in the blog setup?) to "Textile".

I spent ages wondering what permissions were wrongly set, before I found that.

It could be mentioned somewhere, either in the MT2.6 user doc or in your release notes, but I managed to completely overlook it and only found it after an hour or two fiddling with the set up!

1. Don't use 'target' and let your users decide when to navigate to a new window.
2. Use regular <a href=...> tags for when you need the 'target' attribute.
3. Hack MT-Textile and add support for it. The license certainly permits that.

Love the concept and thanks for writing it -- I look forward to easier blogging!

in Germany, it is better to link to a new window, if you are linking to another site - people can sue you for enclosing their content in your frames. So sorry, but I'm sticking with popups as my site is a German domain ;-)

I moded the plugin, which had the nice side effect after rebuilding the site, that the existing links done with Textile each link to a new window!

Thanks again for the plugin - it makes life much easier.

in Germany, it is better to link to a new window, if you are linking to another site - people can sue you for enclosing their content in your frames. So sorry, but I'm sticking with popups as my site is a German domain ;-)

I moded the plugin, which had the nice side effect after rebuilding the site, that the existing links done with Textile each link to a new window!

Thanks again for the plugin - it makes life much easier.

My one question is if there is any way to control positioning (align=) or borders or titles on images?

As soon as I remove textile.pl from my plugins folder, though, the save works fine.

Arg! Brad?

Your documentation for links is:
"linktext[space](alt)":url.
If you leave in the space, it adds a space to the end of the link. I left it out and it seemed to work fine.

Let me know if there's anything wrong with this approach.

I've been changing a few things around in the .pm file to suit me better but I don't know how to set it so an empty alt="" is put in img tags if no alt text is specified. Suggestions?

"A sample link":http://www.google.com/: a good link to check out.

Putting a space between the url and the colon works of course, but that leaves the space in the displayed blog entry, not good. Making the colon part of the linktext works, but that's kind of ugly as well. Other delimiters work fine such as a full stop or a closing ) bracket.

Is there some trick to signal to the plugin that the url is over and that what follows is normal text?

Like me, you may have been _desperately_^[1]^ trying

gives me my sup but not my em?

The offending comment described above:
http://nontroppo.org/blog/archives/2002/10/16/predictable.html

Thanks

1.2 sounds as if it can be made to do exactly what I want. Thank you for all of them, anyway. Even 1.0 has made blogging hugely easier for me. So I realy didn't want to come over sarky and demanding.

Incidentally, one thing I noticed today: a string of the form ??Is this right? Certainly not!?? doesn't seem to get converted by 1.0 -- the '?' in the middle seems to throw it off. I didn't see that in the changelog, so it may not have been reported.

like <b>so* and then later on in the same sentance do it *again<b> and <b>again*

Not pretty. I changed the regexp from (.+) to (.+?) and it fixed it. Enjoy.

--- textile.pm.orig 2003-02-17 12:26:36.000000000 -0500

+++ textile.pm 2003-03-06 11:05:49.000000000 -0500

@@ -194,10 +194,10 @@

                     # macros -- applied to non 'pre' blocks:

                     foreach my $f (keys %qtags) {

                         my $r = $qtags{$f};

-                        $text =~ s!(^|\s|>)$f\b(.+)\b($punctre*)$f(($punctre{0,2})(\s|$)|$2$3$4!gm;

+                        $text =~ s!(^|\s|>)$f\b(.+?)\b($punctre*)$f(($punctre{0,2})(\s|$)|$2$3$4!gm;

                     }

 

-                    $text =~ s!\^(.*)\^!$1!gm;

+                    $text =~ s!\^(.+?)\^!$1!gm;

 

 

                     # (tm) -> ™

Thanks

I'd like to create an HTML::Textile, mainly for my own use in conjunction with HTML::Mason. Basically, I'd like to write all my page content in Textile markup, and have Mason translate it to HTML on the fly.

I assume I'd also need to ask Mr. Allen if I can use the name.

!http://ooo/pic1.jpg! foobar !http://xxx/pic2.jpg!

is expect to see (pic1) foobar (pic2) but is parsed as "!http://ooo/pic1.jpg! foobar!" and "http://xxx/pic2.jpg!".
If I put the follow line, it works fine. (notice the space before `!')

!http://ooo/pic1.jpg ! foobar !http://xxx/pic2.jpg !

I think rewrite re slightly may solve the problem.

Thanks.

What's the problem?

Looking at the source code, looks like & is getting changed to &amp; before it gets there, which complicates the pattern slightly.

Suggested patch:

--- textile.pm~ Mon Feb 17 09:26:36 2003
+++ textile.pm  Tue May 20 09:40:44 2003
@@ -183,9 +183,9 @@
                     $text =~ s!(^|\s|>)\!(\([A-Za-z0-9_\-]+\))?([^\s\(]+)((\([^\)]+\)|[^\!]+)*)\!(:$urlre)?!$1._repl($repl,_mkimage($ctx,$3,$4,$6,$2))!gem;
  
                     # ABC(Aye Bee Cee) -> acronym
-                    $text =~ s|([A-Z][A-Z0-9]+)\((.+?)\)|_repl($repl,qq{$1})|ge;
+                    $text =~ s{([A-Z](?:[A-Z0-9]|&)+)\((.+?)\)}{_repl($repl,qq{$1})}ge;
                     # ABC -> 'capped' span
-                    $text =~ s/(^|[^"][>\s])([A-Z][A-Z0-9\., ]+)([^$2}).$3/ge;
+                    $text =~ s/(^|[^"][>\s])([A-Z](?:[A-Z0-9\., ]|\&)+)([^$2}).$3/ge;
  
                     # simple replacements...
                     # _blah_ -> emphasis

"hola@hola.com":mailto:hola@hola.com

Any solution to this?

Thanks for any information.

Only "Convert Line Breaks on by default: (?) ".

MT and other plugins are feeling well...

Whats wrong?

$text =~ s!(^|\s|>)\!(\([A-Za-z0-9_\-]+\))?([^\s\(]+)((\([^\)]+\)|[^\!]+)*)\!(:$urlre)?!$1._repl($repl,_mkimage($ctx,$3,$4,$6,$2))!gem;

to:

$text =~ s!(^|\s|>)\!(\([A-Za-z0-9_\-]+\))?([^\s\(\!]+)((\([^\)]+\)|[^\!]+)*)\!(:$urlre)?!$1._repl($repl,_mkimage($ctx,$3,$4,$6,$2))!gem;

Perhaps it helps somebody. =)

For hlb: Did you check the list of HTML tags allowed in your blog configuration?

How do I code for a mailto: link? "text":mailto:address@foo.com doesn't work; nor does "text":"mailto:address@foo.com"

And what, if anything, do I need to do to my RSS templates once I've got this running, so that the content comes out right?

Thanks for a great tool! I finally have a weblog authoring solution that I really like (in combination with NewsGator and the MovablePoster plugin).

When I format with Textile, it wraps my DIV in a P tag, and I therefore get an error (document type does not allow element "div" here; missing one of "object", "applet", "map", "iframe", "button", "ins", "del" start-tag) when trying to validate (XHTML 1.0 Trans).

How can I make it so it doesn't wrap my DIV in the P tag?

Thanks!

It appears that using +text+ doesn't wrap text in <ins>-tags if text contains a html-link. Ie

+Blabla <a href="link">blah</a> blah.+

Will not wrap the text in -tags, like it does at http://www.textism.com/tools/textile/ . Using the Textile-format for links works fine though, so it is no biggie. Keep up the good work.