Closed GoogleCodeExporter closed 8 years ago
Or, loops can be designed as follows to make them cleaner:
for(ctr=number of notes/beats;ctr>0;ctr--)
{
(operate on note/beat number ctr-1)
}
Original comment by raynebc
on 3 Nov 2010 at 4:54
r524 restores some variables as int in order to prevent any malfunctioning.
Eventually, the entire code base would probably need to be looked at in order
to remove the use of signed variables to track things that are safer to use as
unsigned variables (counters, realtime positions, etc). This should probably
be put on hold until at least the 1.7 stable release has been made.
Original comment by raynebc
on 3 Nov 2010 at 8:18
Some other functions like eof_get_beat() use -1 as a return value for an error
status. Ultimately, this would need to change, or unsigned variables should be
disused in various places to improve compatibility.
Original comment by raynebc
on 3 Nov 2010 at 8:54
Original comment by raynebc
on 3 Nov 2010 at 9:59
Up to now, I may have changed some variables to unsigned when they were
designed to work as signed. I'll need to look for decremental counters that
were being checked for negativity (ie. >= 0).
Original comment by raynebc
on 27 Nov 2010 at 12:47
A majority of the signed loop counters have been converted to unsigned where
possible. There are various other places where signed numbers representing
beats, etc. are returned by functions or passed as function parameters that I
have left alone. It's hard to predict where changing a signed to an unsigned
will break the program, as issue 210 demonstrates. The rest of the conversions
will probably need to occur over time.
Original comment by raynebc
on 30 Nov 2010 at 10:53
I rarely encounter sections of EOF's code that are vulnerable to
signed/unsigned conflicts anymore, so I'll consider the majority of them
resolved.
Original comment by raynebc
on 9 Jan 2011 at 10:15
Original issue reported on code.google.com by
raynebc
on 3 Nov 2010 at 4:40