Windows binary saga

[Anton Lundin]

On 22 September, 2015 - Lubomir I. Ivanov wrote:

> On 22 September 2015 at 15:05, Lubomir I. Ivanov <neolit123 at gmail.com> wrote:
> >
> > so apart from NO_DOCS=0, the only thing i'm not building at the moment
> > is FTDISUPPORT=1, but that's probably doable as well.
> >
> 
> alright,
> 
> so libftdi builds (without the EEPROM part as i don't want to install
> the other libconfuse dependency thing), but the local serial_ftdi.c
> doesn't compile.

The libftdi bits are mostly ment to be used on android where you can't
use kernel-drivers for ftdi. Anyway, the code should be cleaned up.

> 
> Anton, if you are the maintainer of serial_ftdi.c, please review the
> attached patch so that Dirk can apply it
> 
> timeradd()/timersub() are a BSD thing, not POSIX and Windows has never
> heard of them! :-(
> i've added replacements, but that's completely untested.
> 

Yea, maybe we should replace that wart with something less ugly.

> on this:
> #define ERROR ...
> such a common name is a bad idea. conflicts with the existent
> definition in "wingdi.h", which the GDI people took pretentiously.
> i'm renaming it to *ERROR_MSG().
> 

Those macros was just some bits from when the code was ment to end up in
libdivecomputer and the macros was just ugly replacements for the
libdivecomputer versions.

> there are also a ton of "const char* -> char *" and similar warnings
> (+ bad coding style), but i'm not touching that.
> 

Yea, the hole file should be cleaned up, but it sorta-kinda-works right
now.

Anyhow. I'll take a look at your patch later, and see how/what/if we
actually would like to do with that ftdi code.


//Anton - trying to straighten out bluetooth code right now.


-- 
Anton Lundin	+46702-161604