Hi Gavin, On Wed, 8 Jan 2003 00:25:25 +1100 Gavin Sinclair <[EMAIL PROTECTED]> wrote:
> (Left|Right|Center|Full)-justification has little to do with spaces > and everything to do with newlines. No, of course this is wrong. See this line and this and this line too No imagine <Alt>+L would be applied to these three line and it would only care about '\n'. What's the result? See this lineand thisand this line too Bad ... really bad. The re-justification has to take care of spaces, because it has to insert one for every '\n'. But now take these lines: This line is a shorter one. While this will be a very long, long line that's not wrapped correctly at e.g. 72 characters counted from beginning of line. And another shorter line. No imagine re-justification only replacing every '\n' with ' ': This line is a shorter one. While this will be a very long, long line that's not wrapped correctly at e.g. 72 characters counted from beginning of line. And another shorter line. Result: one long line. So this could only be an intermediate step to the final result. What's next? Simple: The Bat! has to wrap each line at XX (e.g. 72) characters. But it shouldn't wrap in the middle of a word, but only between them AND not after XX characters. So The Bat! has to spplit up these one very long line into single words, take the first one and check for "length() <= 72". If this is true it adds the next word and checks again ... until this conditio is false. No the last appended word is stripped off and the result is written down a "one line". The just stripped word is taken as starting point for the next line and the following words are appended until the line length would grow to much ... and so on. But "what is a word?" Words distinguished by "non-word boundaries". This non-word boundaries are defined as "everything but [a-zA-Z0-9_]". Therefore the space is a simple boundary. Now in the final "line building" the words are "glued" together by spaces. As all spaces are "stripped" while splitting the one long line into words the double spaces are lost. There's an exception for word-glueing: punctation characters. If the next word is a punctation character it's appended without trailing space, to avoid them beeing singled out on a line like this : Of course, the whole process could be arbitrary more complex, but that (or something very similar) is the basic method. If one now wan't to introduce a "double space between sentences keeper" he'd have to write a far more complex algorithm. And this might have two disadvantages: 1. it migth slow down thing sensible 2. One (the coder) might have to implement a completely "self written" function to achieve wrapping. The one given in The Bat! might be one the component used as editor brings with it, so it's optimized for speed and usuability. Implementing an own opens the gate for introducing new bugs and significant slow down. All in all I don't see the down side of The Bat! stripping the double spaces to single ones. This is still e-mail and not a typewriting competition and it seems we're all able to read texts even if sentences are isolated by single spaces. -- Regards Pit ________________________________________________ Current version is 1.62 | "Using TBUDL" information: http://www.silverstones.com/thebat/TBUDLInfo.html
