[reportlab-users] Another tables patch

[Robin Becker]

Gary Poster wrote:
> Hello.  Please find attached an additional patch.  This one is against 
> the SVN trunk, revision 2539 (current as of this writing).  It is 
> intended to be applied by cd-ing into the platypus directory and then 
> the usual "patch -p0 <platypuspatch".  As you can see, it mainly patches 
> tables.py, but has a small change to paragraph.py.
> 
> This patch does three things.
> 
> First, the patch I sent yesterday included one diff that didn't make it 
> to the checkin yesterday: without it, the join code breaks.  Without the 
> changes, the code also barely avoids stomping on the following 
> arguments, if provided, such as linecount.  Today's patch reinstates the 
> fixes.  The change is the last one listed in the tables.py portion of 
> the patch.  (I'm not even sure if specifying a join in a table actually 
> has a visible effect because of the way the lines are drawn, but the 
> code is there so it might as well be correct.)
> 
> Second, today's patch includes code to make a much better guess about 
> column widths in the case of a table with both non-fixed-width 
> flowables, such as paragraphs, and columns without absolute widths 
> specified (None or '*'; maybe code should also check for percentages).  
> It respects percentage requests as much as possible while doing its best 
> to give columns the minimum width they need.  Widths of '*', if present, 
> are considered greedy in comparison to widths of None.  Another part of 
> the change is that if the total table width is greater than the 
> available space it no longer raises an error: I think this is an 
> improvement, but I could be mistaken.  In total, this part of the patch 
> a very nice improvement and works well in my code.  It exercises the 
> Flowable.minWidth method to a greater degree than before, so it might 
> expose some problems in other flowables...which leads to the last change.
> 
> Third, the patch includes simple code for paragraphs.py that handles the 
> edge case of minWidth for a paragraph with no content.
> 
> I hope you can use the patch.
> 
> Thanks
> 
> Gary
sorry to have messed up the last one. I have merged that bit separately as I 
don't have time to check the rest today.
-- 
Robin Becker