On Thu, 16 Feb 2012 13:08:09 +0100, Hans Hagen
no, although it's no big deal to provide that (of course there is then the matter of utf being more complete than the target)
---- src_enc = "utf8" tgt_enc = "cp1250"
str = regimes.translate(str, src_enc, tgt_enc) ----
is there a reason not to stick to utf?
Hans
Well - I'm working with a .cld document (with UTF encoding). There are some functions which typeset texts. And there is also a part which creates a .scr file. .Scr files are sequences of AutoCAD commands - their contents are passed directly to AutoCAD command prompt. When AutoCAD is creating a text entity, it reads the input stream (in our case: the .scr file) BYTE-PER-BYTE. When bytes represent a text to be drawn, unknown bytes (= bytes that don't have any graphical representation in AutoCAD font file ("shape" file in AutoCAD's terminology)) are shown as "?". Of course, valid representation of language-specific-characters (like "čřž..." in Czech) requires an appropriate .shx (= "shape compiled") file. Anyway, when AutoCAD is to write "č", it requires just ONE BYTE to be passed to it; so 2-byte UTF representation gives bad result (= "??"). So back to the origin, when I call the .cld's function that writes a command to the .scr file, I need to convert a UTF string back to CP 1250. Would it be possible to provide this? NB: There are two examples of .scr files; CP1250.scr works well in AutoCAD, the latter draws "????ST" instead of "ČÁST". Kind reagrds, Lukas -- Ing. Lukáš Procházka [mailto:LPr@pontex.cz] Pontex s. r. o. [mailto:pontex@pontex.cz] [http://www.pontex.cz] Bezová 1658 147 14 Praha 4 Tel: +420 244 062 238 Fax: +420 244 461 038