After testing my original approach on a few more parts, I wanted to test the Edger2 red line tool as suggested by MagFors.
I open a file I just happen to have lying around; 3069bp84.dat in LDPE. I switch to No Bacface Culling view mode and select top view. 3069bp84.dat is a flat pattern on a 1x2 tile. Visual verification with pictures was effortless.
https://library.ldraw.org/parts/50020
For testing purposes, I deliberately edit the x-coordinate of the color 16 quads sharing the 3.35 0 -2.75 coordinate to create a thin gap. I leave the purple surfaces unaffected.
I check with Edger2 with default settings that it doesn't report anything. My method of selecting the vertex notices the mismatch and the popup doesn't show up.
Now I'm ready to test the Edger2 red lines trick. Didn't find it at first, but the Tools menu contains a second Edger2 that allows changing the options. The default setting seems to be "Exclude Unmatched Edges" which is probably good for subfiles.
I change the dropdown to "Include Unmatched Edges" and run the tool. It gives me six red lines. Four are where I introduced the vertex mismatch. That worked, but what are those other two?
They are at the top where the pattern has introduced an extra vertex on a hard edge (line type 1). This is acceptable. Consequently, not all red lines are errors.
Running the tool on 2335p04.dat, I find that it reports all T-junctions both in the pattern in the the base part. Those are all fine.
https://library.ldraw.org/parts/50302
Running the tool on 2345p01.dat, I find that it generates shorter cond lines along the fold where the parent part has one long cond line. Additionally, the red lines appear over the same lines. These are also fine.
https://library.ldraw.org/parts/50303
The tool comes in really handy when there is a large pattern to review; like in 60581p05.dat having some 3000 lines. LDPE and Edger2 are slow, but in the end it produces 4 red lines, all of which are T-junctions in the parent part. The pattern is perfect.
https://library.ldraw.org/parts/39995
I did also some tests with the y-coordinate, but the cond lines aren't that reliable in detecting small displacements. Given that I can be pedantic about the y-direction only on flat patterns, perhaps I should tolerate as much deviations as on 3D shapes - i.e. adopting the cond line approach: If the pattern doesn't generate new cond lines, then its fine.
It seems both suggestions will improve my efficiency in reviewing patterned parts. Thanks MagFors for the great suggestions!
I open a file I just happen to have lying around; 3069bp84.dat in LDPE. I switch to No Bacface Culling view mode and select top view. 3069bp84.dat is a flat pattern on a 1x2 tile. Visual verification with pictures was effortless.
https://library.ldraw.org/parts/50020
For testing purposes, I deliberately edit the x-coordinate of the color 16 quads sharing the 3.35 0 -2.75 coordinate to create a thin gap. I leave the purple surfaces unaffected.
I check with Edger2 with default settings that it doesn't report anything. My method of selecting the vertex notices the mismatch and the popup doesn't show up.
Now I'm ready to test the Edger2 red lines trick. Didn't find it at first, but the Tools menu contains a second Edger2 that allows changing the options. The default setting seems to be "Exclude Unmatched Edges" which is probably good for subfiles.
I change the dropdown to "Include Unmatched Edges" and run the tool. It gives me six red lines. Four are where I introduced the vertex mismatch. That worked, but what are those other two?
They are at the top where the pattern has introduced an extra vertex on a hard edge (line type 1). This is acceptable. Consequently, not all red lines are errors.
Running the tool on 2335p04.dat, I find that it reports all T-junctions both in the pattern in the the base part. Those are all fine.
https://library.ldraw.org/parts/50302
Running the tool on 2345p01.dat, I find that it generates shorter cond lines along the fold where the parent part has one long cond line. Additionally, the red lines appear over the same lines. These are also fine.
https://library.ldraw.org/parts/50303
The tool comes in really handy when there is a large pattern to review; like in 60581p05.dat having some 3000 lines. LDPE and Edger2 are slow, but in the end it produces 4 red lines, all of which are T-junctions in the parent part. The pattern is perfect.
https://library.ldraw.org/parts/39995
I did also some tests with the y-coordinate, but the cond lines aren't that reliable in detecting small displacements. Given that I can be pedantic about the y-direction only on flat patterns, perhaps I should tolerate as much deviations as on 3D shapes - i.e. adopting the cond line approach: If the pattern doesn't generate new cond lines, then its fine.
It seems both suggestions will improve my efficiency in reviewing patterned parts. Thanks MagFors for the great suggestions!