[reportlab-users] Handling of too long entries in Table

[Andy Robinson]

> Sorry for the typo and good luck searching! ;-)
I saw that, and thought "funny, I must be forgetting my
school German..."

 
> Second, at first sight URLs could be treated like that as the only
> allowed splitting points would be at slashes and some other characters
> like dots, questionmarks and some other ones. But URLs also differ
> from a normal language because their file paths very often contain
> "normal language" words like in
> 
>    http://www.reportlab.com/docs/graphics_reference.pdf
> 
> where you might consider it ok to split between "graph" and "ics",
> too. Hence you need to respect the URL language *plus* the "real"
> language... And you'd need to detect which real language that is...


We're down to 3 levels deep, why not have n?
So, in the general case we have a 'language'
wth a 'split routine'.  The outer language is
usually called 'text' and splits on whitespace.
Each sub-section may have its own language, which
can be either explicitly tagged or guessed from the
content.  And that may have its own split algorithm.

One more step towards a Grand Unified Theory of
document layout....

- Andy