Alexander Larsson wrote:
>
> I've added latin1 encoding to all fonts in exported eps files.
> It's not very nice currently, as i do encodings for all fonts, not just
> the onces used in the exported diagram. But it works.
>
> Marcin, could you try if ogonkify works for you now?
>
Using
ogonkify -ATH -N
(-ATH means: correct Courier, Times and Helvetica fonts, -N means: the
postscript is "Netscape-like")
I translated eps generated by dia to eps which correctly contained
polish characters. I tried normal text and some UML symbols. As ogonkify
supports many encodings (ISO-Latin-[1-9] and some other), this means dia
went international. Thanks Alex!
One word of warning: ogonkify in the form like above works well for
Courier, Times and Helvetica fonts. Other will not be "corrected". It
seems dia uses only those fonts on default symbols, so this is not large
problem. As future concept: it would be nice to specify in some
configuration file/window which fonts are to be used for UML symbols
etc.
* Problem with << and >>
Experimenting with different symbols I noticed, that the stereotypes
look rather funny - instead of
<< stereotype >>
I get
^ '
T stereotype t
(T-dash and t-acute are single symbols).
It looks so on screen and in ogonkified eps. In default eps there are
correct << and >> characters.
It seems the problem is caused by the fact that the iso-latin-2 fonts
doesn't contain << and >> on the places one would expect.
Maybe those symbols could be painted as glyphs or just as double <,
double > ?
Regards (and thanks)
-- Marcin Kasperski Marcin.Kasperski<at>softax.com.pl
-- marckasp<at>friko6.onet.pl
-- Moje pogl�dy s� moimi pogl�dami, nikogo poza mn� nie reprezentuj�.
-- (My opinions are just my opinions.)
