[DGD] Hydra 1.2.8

[Felix A. Croes]

diff -c 15ebec4df74da034f55a42c472b72b7a4a8828d2/src/Changelog ead14b04f366591a369215c2f4a15288e02117a9/src/Changelog
*** 15ebec4df74da034f55a42c472b72b7a4a8828d2/src/Changelog
--- ead14b04f366591a369215c2f4a15288e02117a9/src/Changelog
***************
*** 172,175 ****
   - Prevent a buffer overflow in kf_restore().
   - Store times in the callout table with the same offset as in objects.
   - Switch to 64 bit (48 bit seconds, 16 bit milliseconds) time internally.
! 
--- 172,191 ----
   - Prevent a buffer overflow in kf_restore().
   - Store times in the callout table with the same offset as in objects.
   - Switch to 64 bit (48 bit seconds, 16 bit milliseconds) time internally.
!  - Fixed a bug in the handling of floats in kf_sum().
!  - The size of an array is now held in a value of type ArrSize.
!  - Add O_INHERITED and O_INSTANTIATED to status(obj).
!  - Fixed two instances where bad code was generated for the new builtin kfuns.
!  - Distinguish between the LPC integer type and other internal integer types
!    that used to have the same size.
!  - Hydra can now be compiled for 64 bit LPC integers and 96 bit LPC floats.
!    The old (and for Hydra, obsolete) instructions 0 and 1 are used for I_INT8
!    and I_DOUBLE, respectively.  Big floats converted to strings have 18 digits
!    of precision; for small floats, the number of digits is increased to 9.
!    Trigoniometric functions have not yet been adapted.
!  - The internal precision for small floats was given as 10 digits, when it
!    should have been 11 all along.
!  - Allow snapshots created with small ints/floats to be restored by a server
!    with big ints/floats, but not vice versa.  Integers will be expanded as
!    signed values, and floats will be converted to the bigger format.
!  - Add a (dummy, for the moment) argument to shutdown().

    http://ftp.dworkin.nl/hydra/

Regards,
Felix Croes