[reportlab-users] Small errors in reportlab code

[Andy Robinson]

>>BTW, wouldn't it be better to ignore useless Unicode and UTF-8 decoding

>>errors by using e.g.

>>

>>  s.decode('utf8','ignore')

>>

>>instead of

>>

>>  s.decode('utf8')


In general, when I'm building reporting systems, I want to know about 
corrupt input as soon as possible.  I believe our libraries should be as 
strict as possible about this:  if you say it's UTF8, it should be UTF8, 
and anything else should cause an exception.

Otherwise you end up with ? marks or missing data in printed literature 
for somebody else a long time later, and it's very hard to pinpoint 
where the error was introduced.

If you have to live with corrupt input, it's only two lines to write 
your own "pre-filter" function which will accept bad input.

Potentially we could make this user-configurable in future with a 
default of 'strict'.

- Andy