On Thu, 02 Sep 1999 18:58:46 +0100, Duncan Simpson wrote:
> Given the only way to fix bugs in word is to upgrade, which costs money,
>conspiracy theorists would say I was right :-) IMHO they might not be entirely
>wrong. Word 98 is avuabale without a NDA but the file format itself is just as
>dire (not worse than word 6, primarly because doing wose is virtually
>impossible).
>
As it seems, Word 8 is now de facto the international standard for
document exchange, even in the internet and within specialized academic
contexts. LaTeX 'market share' is rapidly decreasing. Even Linux people
seem to adhere more and more (see the new xyzOffice suite efforts for
Linux). So LyX has to change gears and gain speed...
>Unfortunately 99% of the time people need to do this the only avialable format
>is word binary junk. Hence word2x, catdoc and MSWordview---you do not imagine
>the authors of these beasts actually like word format, do you? I can only
>speak for word2x's author myself, of course.
Do you plan Word 8 support?
Some time ago I compiled an old version of word2x on emx. Compiling was
relatively easy, only minor quirks. Adding something like this; I think
not all were necessary:
#ifdef __cplusplus
extern "C" {
#endif
#ifdef __EMX__
/* Compare S1 and S2, ignoring case. */
int stricmp(__const__ char *, __const__ char *);
#ifndef strcasecmp
#define strcasecmp stricmp // emx uses M$ fn name instead of posix
name??!
#define _strcasecmp stricmp
#endif
/* Compare S1 and S2, ignoring case. */
int memicmp(__const__ void *, __const__ char *, size_t);
#ifndef memcasecmp
#define memcasecmp memicmp // emx uses M$ fn name
#endif
#if !defined(__STRICT_ANSI__) && !defined(_POSIX_SOURCE)
/* Compare no more than N chars of S1 and S2, ignoring case. */
int strnicmp(__const__ char *, __const__ char *, size_t);
#ifndef strncasecmp
#define strncasecmp strnicmp // emx uses M$ fn name
#define _strncasecmp strnicmp
#endif
#endif /* !defined(__STRICT_ANSI__) */
#endif // __EMX__
#ifdef __cplusplus
}
#endif
Those funny fn names - despite emx' aim to facilitate Unix ports - is
always a reason for headache.
BTW: Porting LyX we have the very same nuisance, recompiling the emx
Lib C with new aliases would help, but is non-standard.
Could you add this (as far as it is still necessary) to your code base
(to config.h or so?)?
I would like to release an updated emx version, perhaps a bit
synchronized with a LyX release for OS/2, if this hasn't already been
done (there was an old release from Russia, IIRC a bit idiosyncratic).
Would you agree?
A usability problem: Non ASCII-character handling was a major problem
for me; do you plan to add some support for the different M$-code pages
and LaTeX encodings (as used with Babel)?
Greets,
Arnd
