[Bf-committers] Fae of triangle cannot be converted to quadrangular face

PerfectionCat sindra1961reborn at yahoo.co.jp
Fri Jan 4 03:05:28 CET 2013

As for everybody, this year thanking you in advance.

In addition, I make the following sentences by machine translation.
There may be the part which does not understand a meaning, but, in that case, please forgive it.

Because Bug Tracker stops, I report it here.
Because I did not confirm it, to see all source codes, interpretation may be wrong.
I am glad then when I have you teach right interpretation.

You recognized this problem at point in time of version 2.63, but because the opportunity which is mentioned this problem was lost, you apologize the fact that you can send report.
In addition, this is not something which shows improvement method.
Inspecting cause, because it reports that it is ascertained it does not pass.

When by the fact that the function of NGON is taken in, converting the triangular face to the quadrangular face, it has become the cause trouble occurring.
Quadrangular face with CTL+T is designated simply as face of triangle, next that resetting to quadrangular face with ALT+J works normally.
But, it converts quadrangular face to face of triangle with CTL+T, next it cuts in the form which cuts off face of two triangles in half alongside the diagonal of the opposite of original quadrangular face.
It is the expectation which becomes the kind of shape which 4 divides quadrangular face alongside the diagonal.
Selecting face of the triangle which it adjoins, when it converts to quadrangular face with ALT+J, it is the expectation where problem is found.
Executing with whichever combination, it does not work normally.

That I think whether it does not mean that the value which is set to the variable limit which is used with the function edbm_tris_convert_to_quads_exec of editmesh_tools.c is inadequate.
Setting PAI as the value which does not provide restriction in trial, when it executes similar operation, in order past version to work in the same way, it is thought.


However, the behavior which you can think that it is inadequate with when it is several is shown.

Plane is converted to face of triangle.
Next, exceeding the base of face of other triangle one vertex alongside the diagonal, it arranges.
Finally, it converts to quadrangular face.

Actually when it executes, it is clear, but the side on the diagonal going out, new quadrangular face is formed with four vertex.
The air like the result of being reconstructed when normal information has missed does this.


I have observed to only the value which variable limit can take, but I do not inspect whether that value how is handled in inside to.
This report of investigation being vague, serves honest purpose, whether or not is, you do not understand in me.
But, if this becomes the opportunity of addressing the issue, it is fortunate.

The description regarding the inadequate behavior of the part where it put with above-mentioned “+” is visible when you verify including subdivision surface modifier, in order to be processed appropriately.
Therefore, you think that description should be deleted, but that it is the funny case because you think boldly without deleting described, it leaves.

I put the image of the sample to the following URL.


The environment that I confirmed:

Blender version 2.65.5 Rev 53539
Windows XP/32bit


More information about the Bf-committers mailing list