PR #252818 When I construct a Filter feature out of a patch scan

huangyhg

When I construct a Filter feature out of a patch scan, I get an application error.
<< Don Ruggieri  --  01/22/08  11:40:29>>
data is in 01222008_01a.zip and again in 01222008_01b.zip.  When I construct a Filter feature out of a patch scan, I get an application error.
Steps to reproduce:
While testing Constructions as part of the overall v4.3R test plan, on section requires -
Construct a Gaussian Filter set from scan data.  Adjust the tolerance and use Polar and Linear.
I created  patch scan on the top of the curved area near the left side of the hexagon block, consisting of about 7 rows of data every 1.5mm in density.  I opened the ConstructFilterFeature diialog and selected this patch scan as input, leaving all of the other settings as is.
Results:
When I hit the Create button, I get an application error.
<<END>>

huangyhg

<< Changes made by Don Ruggieri -- 02/14/08  11:35:21>>
Status:  RESOLVED to CLOSED
<<END>>

<< Dan Zwick  --  01/24/08  13:17:45>>
I've also uploaded these files to 4.2 and 4.2R.
<<END>>

<< Dan Zwick  --  01/23/08  13:51:06>>
I uploaded the corresponding files to 4.3B.
<<END>>

<< Dan Zwick  --  01/23/08  11:00:59>>
I uploaded to 4.3R so far. Files uploaded:
V43R\CONSFEAT\FILTEROB.CPP
V43R\SCAN\BASESCAN.CPP
and dependencies.
<<END>>

<< Changes made by Dan Zwick -- 01/23/08  11:01:58>>
Action:  Dan Zwick to Don Ruggieri, Status:  OPEN to RESOLVED
<<END>>

<< Dan Zwick  --  01/22/08  20:46:35>>
I fixed this problem and another one as well. It looks like this was code that I modified back in 4.2 in FilterPoints() to get not just the points but the vectors and tip radii. In addition to the bug that was causing the crash (a result of a new test that someone else added recently), the code that was getting the tip radii was not working correctly. It was getting all of the tip radii each time instead of just the ones associated with each basic scan. This probably never caused a problem because nobody ever used more than one tip for a patch scan, but, with Ryan's help, I fixed that as well.

I will upload as soon as I can build both debug and release configurations without errors.
<<END>>

<< Dan Zwick  --  01/22/08  13:50:32>>
Progress report: Ignoring for the moment the fact that doing a Gaussian polar filter on a patch scan is completely nonsensical and anyone who does it should be strung up by the thumbs, the problem seems to be that at some point in FilterPoints() the number of points in a basic scan is 35 but the number of vectors is only 34. I'll try to find out how this came about.
<<END>>

<< Dan Zwick  --  01/22/08  12:55:51>>
I'll drop everything I'm doing and devote myself to this PR.
<<END>>

<< Changes made by Tim Wernicke -- 01/22/08  10:53:11>>
Action:  Andy Roberts to Dan Zwick, Assigned:  Andy Roberts to Dan Zwick, Priority:  to Stop Rel.
<<END>>

huangyhg

1. V43R\CONSFEAT\FILTEROB.CPP
pObj->get_points((ObjIsScan ? SCAN_HITDATAFORROW + BALLCENTER : BALLCENTER), MEAS, PARTMM3, global_align, &NumPntsInRow, pArrMeasPnts, global_workplane);
pObj->get_points((ObjIsScan ? SCAN_HITDATAFORROW + TIPRADIUS : TIPRADIUS), MEAS, PARTMM3, global_align, &NumTipRadii, pArrPrbRadii, global_workplane);
pObj->get_points((ObjIsScan ? SCAN_HITDATAFORROW + VECTOR : VECTOR), MEAS, PARTMM3, global_align, &NumVecs, pArrVecs, global_workplane);

2. V43R\SCAN\BASESCAN.CPP
case SCAN_HITDATAFORROW + TIPRADIUS: