[reportlab-users] Re: The upcoming Reportlab 1.18 release and pat ches.

[Robin Becker]

In article <2546C8F5D3D2D411857F00508BDCA4E8517DDB@dalbar_mail.dalbar.co
m>, Marc Stober <MStober@DALBAR.com> writes
>Hi John,
>
>Here is another problem in 1.17 we discussed a few weeks ago (below). The
>fonts.py file in the lib directory can just be rolled back to the previous
>version (from ReportLab 1.16) if you and Andy agree that's the correct one.
>Let me know if you need any help or information.
>
>Thank you,
>Marc Stober
>mstober@dalbar.com
.... I'm leaving this bug for the future Marc. The real difficulty is
not the aliases (which I admit are not properly controlled), but the
lack of a family concept. I believe that the attempt to cross register
was an attempt by Andy to make some sense of what paragraphs etc do when
you put in <b> or <i> tags etc. I believe the solution is to take out
the automatic mapping attempt and to make the default transition an
identity unless explicitly set. The registration then becomes more
complex if we want to use family names in paragraphs and expect the <b>
and <i> tags mean something. Typically you'd have to register a font as
part of a 'family' with bold/italic properties. The alternative is to
make things more automatic and hope that we can pick up information from
font files etc. I'm not the font expert and Andy is away at OSCON so I'm
leaving this one for now. Sorry.

As for those who were eagerly awaiting instant CVS I'm afraid SF is
badly hosed right now wrt pserver access so as we use that for our daily
builds it won't be right tomorrow. The Realease_1_18.xxx tings are right
though as they were done handomatically.
-- 
Robin Becker