I tried to italicize an I but it didn't work...

a word a word a word

a word a word a word

Everyone feel free to test any other combinations...

-- KevinKinnell - 13 Apr 2000

This is in fact an (undocumented) known bug. Thanks for submitting it. A work around is to enter some invisible text to have more then one character. Example: Enter: " __<nop>X__ " to get: " X " (I know this is a Q&D cludge)

Needs to be fixed...

-- PeterThoeny - 19 Apr 2000

There are also weird effects happening with the end of the file. For example, right now the string after my signature is italicizing the first test but not the second. If anything is put after the end of the second _test_, even a carriage return, then it is italicized properly.

-- JohnAltstadt - 25 May 2000

It seems like a space/newline is required around the escapes. I've noticed this when I try to do italics or bold before a period or comma. For example "*test*" vs. " test ". This would seem to account for the end-of-page problem also. A known bug?

-- JohnBelmonte - 25 May 2000

The effect at the end of the file is a bug. Needs to be fixed.

The rendering rule depends on white space, e.g. beginning and end of line is fine, but not a period or a comma. This is a feature. Example:

• italic OK because of space before comma ,
• italic not OK because no space before period.

The rendering rule could be extended to allow common stop signs like a trailing period, comma or question mark.

-- PeterThoeny - 25 May 2000

If we pick a one character tag for strong/em instead of the star-underscore/underscore-star or double-underscore, we could just get rid of the whitespace delimiters in the rules for text face manipulation. That would allow a lot more freedom of expression. What about using ^ as the strong/em tag? -- kk - 11 Jun 00

Maybe the white space rule should be added to the notes in TextFormattingRules? I don't think it's obvious.

-- JohnBelmonte - 28 May 2000

I think it is essential to keep the whitespace delimiter rule. If we do not we will have many unwanted side-effects, i.e.

• This under_scored_text would loose the underscores.
• This equation would not display correctly: y = 2*x + 5*a

-- PeterThoeny - 20 Jun 2000

Uh oh, I hadn't thought of that. Well, learning to use the occasional <tt> or <em> or whatever is probably not a big inconvienience for most people. I think JohnBelmonte is probably right about adding some documentation about this stuff to the text formatting page.

test and test

-- KevinKinnell - 21 Jun 2000

Fixed single char problem. Commited to TWikiAlphaRelease, version 02 Dec 2000.

You can patch your TWiki installation if you wish. Change the # Emphasizing part in wiki.pm to:

# Emphasizing
  # PTh 25 Sep 2000: More relaxing rules, allow leading '(' and trailing ',.;:!?)'
  s/([\s\(])__([^\s]+?|[^\s].*?[^\s])__([\s\,\.\;\:\!\?\)])/$1<STRONG><EM>$2<\/EM><\/STRONG>$3/go;
  s/([\s\(])\*([^\s]+?|[^\s].*?[^\s])\*([\s\,\.\;\:\!\?\)])/$1<STRONG>$2<\/STRONG>$3/go;
  s/([\s\(])_([^\s]+?|[^\s].*?[^\s])_([\s\,\.\;\:\!\?\)])/$1<EM>$2<\/EM>$3/go;
  s/([\s\(])=([^\s]+?|[^\s].*?[^\s])=([\s\,\.\;\:\!\?\)])/$1<CODE>$2<\/CODE>$3/go;

-- PeterThoeny - 25 Nov 2000