Re: Symbol and Dingbat hackery

To reply to myself, I did not realise that the problem with Symbols and
Dingbats is not what we do with them, but the fonts themselves, which
are broken; in order to be able to get rid of those hacks, we would need
to start using a different font for our bulleted lists. I have looked at
coverage provided by different fonts, and I am not sure it is entirely
feasible; on *nix FreeSerif and OpenSymbol provide pretty much all the
symbols we need, but on Win32 the more exotic symbols (ticks, hands,
implies and boxes) are provided for by the broken Symbol/Dingbat fonts,
  and so we would end up with a different set of hacks, and I do not see
much point in doing that.

Tomas

Tomas Frydrych wrote:
>
> I think we should take the oportunity in the 2.6 cycle to get rid of the
> Symbol and Dingbat hackery that plagues our graphics classes. My
> suggestion is that:
>
> * We increase the AW document version to 1.2 and start storing all
> character values in Unicode.
>
> * We add code somewhere (where ever it makes most sense, probably
> PD_Document), to translate Symbol and Dingbat codes to Unicode on import
> of AW doc version < 1.2.
>
> * There might need to be some changes to the win32 base class should we
> continue to support Win9x in 2.6, although I think Uniscribe deals with
> this internally, not entire sure.
>
> * There might need to be some changes to our importers and exporters,
> have not investigated that yet, but with possible exception of rtf/doc,
> this should amount to removal of hacks.
>
> I have got the Unix side nearly ready, just need to deal with the AW
> legacy issue.
>
> Tomas
>
Received on Tue Mar 21 11:42:53 2006