Fix an endless loop.

[mark-v-d]

The following program created an infinite loop, this patch fixes that

---

#<_initial>=20
#<_final>=19.726
#<_champher>=0.5
#<_zmax>=0
#<_zmin>=-25
#=2

G54
G7
G18
G21
T2M6G43

o100 sub
o101 if [#<_champher> gt 0]
G1 X[#<_final>-2.0001*#<_champher>]
Z#<_zmax>
X#<_final>C#<_champher>
o101 else
X#<_final>
o101 endif
W-5
U0.5
Z#<_zmin>
X#<_initial>
o100 endsub

M3 G96 S200 d1400
G95 F0.12
G52 z[0.1-#]
;m7
G71 Q100 Z[#<_zmax>+0.8+#] X#<_initial> D# I# R0.3
G52 z0
;m7
g42
G70 Q100 Z[#<_zmax>+1] X#<_initial> D# P1
g40
M9
G0X[#<_initial>+5]

M2

[andypugh]

There seems to be a conflict with other changes to interp_g7x.cc

Your change:

	if(std::abs(imag(location)-x)>tolerance || (*p)->ep()==location) {

Existing code in Master:

        // We have added our move commands for the last path, we end the
        // recursive function before calculating and adding intersections.
	if (p==std::prev(end()))
	    return;

	if(std::abs(imag(location)-x)>tolerance) {
>>>>>>> master

Is it possible that this is already fixed?

[mark-v-d]

I'll have a look soon, but I'm not sure it solves the same problem. I will also improve my patch, since == on floating point numbers is not a good idea. andypugh ***@***.***> wrote: andypugh left a comment (LinuxCNC/linuxcnc#3790) There seems to be a conflict with other changes to interp_g7x.cc ``` <<<<<<< 2.9 if(std::abs(imag(location)-x)>tolerance || (*p)->ep()==location) { ======= // We have added our move commands for the last path, we end the // recursive function before calculating and adding intersections. if (p==std::prev(end())) return; if(std::abs(imag(location)-x)>tolerance) { >>>>>> master ``` Is it possible that this is already fixed?

…

-- Reply to this email directly or view it on GitHub: #3790 (comment) You are receiving this because you authored the thread. Message ID: ***@***.***>

[petterreinholdtsen]

I corrected the base for this pull request to 2.9, and got rid of the conflict.

[petterreinholdtsen]

Given that the submitters branch was based on 2.9, I assume it was ment to be targeting 2.9, not master.

Given the example program, I recommend adding a test to demonstrate the problem and hence also demonstrate that the patch fixes the problem. Can you provide code in tests/ to demonstrate the endless loop?