[reportlab-users] Tons of table questions

[Andy Robinson]

2009/3/6 Dinu Gherman <gherman at darwin.in-berlin.de>:

> Not sure I understand this. Does the current LongTable show

> non-linear performance or not?


I believe (without having detailed benchmarks) that changing
a 0 to a 1 in rl_config.py will improve performance, and is unlikely
to adversely affect any real-world scripts.   I am asking if anyone
else is concerned, before making the change.    If someone's
document relied on the table algorithm correctly spotting
an extra-wide table cell on row 500 of a long table, and allowing for
it right from the beginning, this change would impact them.

Since nobody has answered I now take it that this change is safe,
so will probably do it today (I'm not somewhere I can code now).

IMHO further improvements beyond that need a benchmarking script,
with a bunch of different options (given or calculated column widths,
simple versus more complex cell content), as well as a theoretical
analysis which might miss part of the problem.


> In any case I'm missing some-

> thing like a "next action" (in GTD parlance)...


I have added it to the tracker.  I cannot say when we will do anything
about it, nor what, but if anyone wants to contribute a benchmarking
script then it would probably happen sooner or encourage others
to hack on it.

- Andy