Hi,

brlcad has left the following comment at Investigate and fix source of rounding error https://www.google-melange.com/gci/task/view/google/gci2014/5880145319559168:


reassuring


This is very reassuring! Looking at the values, most are either identical, or matching within double-floating point precision (probably 1ulp off) at 15 digits after the decimal.

As this was an issue you had to investigate in detail, that implies a comment is warranted to document and warn others about the sensitivity issue. You do have a comment that says calculating the angle is complicated, but you don't explain what that means in lay terms. Basically, there needs to be some warning that says these calculations are sensitive to trigonometric accuracy so care must be taken to ensure calculated rays do not have or accumulate error.

Looking at the line counts, this does increase the counts by about 20 lines which is pretty minor and probably okay to commit now that the results are matching. However, the data remarshalling does make me think that returning an xrays list is not the best approach. in fact, ideally the _gen_ functions shouldn't be allocating memory dynamically too, giving the user the option of creating memory dynamic or on the stack and passing that in to be used.


Greetings,
The Google Open Source Programs Team


---
You are receiving this message because you are subscribed to Investigate and fix source of rounding error. To stop receiving these messages, go to: https://www.google-melange.com/gci/task/view/google/gci2014/5880145319559168.

------------------------------------------------------------------------------
New Year. New Location. New Benefits. New Data Center in Ashburn, VA.
GigeNET is offering a free month of service with a new server in Ashburn.
Choose from 2 high performing configs, both with 100TB of bandwidth.
Higher redundancy.Lower latency.Increased capacity.Completely compliant.
http://p.sf.net/sfu/gigenet
_______________________________________________
BRL-CAD Tracker mailing list
brlcad-tracker@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/brlcad-tracker

Reply via email to