TA_TRIMA
Brought to you by:
fortier
I believe the inner or outer ranges from which separate
SMA's are calculated are incorrect resulting in out of
bounds read/write.
I ask for TRIMA(4665,4665,10...) and then internal to
TRIMA calls SMA(0,4661,...) for outer and SMA
(4660,4665,...) for inner. Perhaps there is some offset
missing in the outer (either the zero should be offset
upward or the 4661 downward).
Thanks,
Chris
crokusek@hotmail.com
Logged In: YES
user_id=1123
Thanks Chris.
Yes, it is broken.
I could fix it now, but instead I think I will revisit the whole
algorithm. It is currently innefficient for user doing the
calculation for one (or just a few) price bar.
I should be able to provide a working implementation in CVS
by March 17 (I have a busy week).
Very good bug report, keep these coming !
\Mario
Side Note: I am surprise that my regression testing did not
catch this. I will see to add a test for this...
Fix for this problem. This will be included in V0.0.8
Logged In: YES
user_id=1123
Please find attach a new TA_TRIMA.c file
The Fix is also now check-in in CVS.
The FIX will be part of release V0.0.8
Thanks Chris.
Logged In: YES
user_id=1123
Please find attach a new TA_TRIMA.c file
The Fix is also now check-in in CVS.
The FIX will be part of release V0.0.8
Thanks Chris.
Contributed patch, already applied. TA-Lib has moved to GitHub and this SourceForge tracker is retired.
This ticket carried a contributed patch (
ta_TRIMA.c) and was resolved asclosed-fixedon SourceForge in 2003. It is accounted for in the final SourceForge review: https://github.com/TA-Lib/ta-lib/issues/30No further support is provided on SourceForge. Please send patches as pull requests on GitHub instead:
Thank you for the contribution.