Alexander Wallace <[EMAIL PROTECTED]> writes:

> Now, hopefully it works with the ugly hacks....

The one worrying me the most is the SIGPIPE stuff. It basically means
the process will receive a SIGPIPE on controller failures, which sort
of defeats the very purpose of the software...


> Is there something you need from me to try to make this work on
> solaris by default?

Yes! Please open a new CAROB jira issue with the detailed and
exhaustive list of the hacks you currently use. That will help A LOT.


> as I mentioned to you before, i had to define and implement isinf and
> std:signbit ... I certainly hope my implementation is correct...

These, as well as the uintN_t types, are part of C99 and NOT of
C++98. That's the reason they are a recurrent source of pain. You
could try to play with #includes <math.h/cmath> <stdint.h/cstdint>,
using inspiration from what we did for MINGW. Also check inttypes.h,
sys/types.h, include files found on the internet...


> Next, I will try compiling libMySequoia... So i'm as lost again :)
> but will hopefully find the way.... I'm sure I'll need your help!


Just realized libmysequoia is linking against a dynamic carob by
default... hopefully it's not too hard to change that.

On the portability front libmysequoia is using autoconf, which is
supposed to solve issues automagically.




_______________________________________________
Carob mailing list
[email protected]
https://forge.continuent.org/mailman/listinfo/carob

Reply via email to