[reportlab-users] Platypus idea "splitting preferences"

[Andy Robinson]

2010/1/14 Roberto Alsina <ralsina at netmanagers.com.ar>:

> I am thinking there are flowables that would prefer not to be split if

> possible, for example small literal blocks (say, a code fragment), but *if

> needed* can be split anyway (for example if they are over a page long, or

> would leave a half-empty page if not split).

>

> Would reportlab be interested in this kind of functionality? I am thinking of

> adding a property to the flowables describing this and hacking the place where

> frames consume flowables so it's taken into account.


Definitely.


> Any suggestions about how the property should look like, or how it should be

> handled are welcome, since I am still pretty fuzzy on what I really mean ;-)


Can you give some examples of the flowables which this would apply to?

I am aware of table cells being a problem, because someone might put
pages worth of content into a table cell at the moment, and we can't
split inside one row.  But I am not sure how we could do the
splitting.

If we ever get to ReportLab 3000, it ought to have some kind of
fuzziness - the ability to ask for multiple split points, and/or
backtracking so it can come up with the least ugly solutions.    I
have plenty of ideas on that, but none are even remotely
backwards-compatible.

- Andy