[Octave-bug-tracker] [bug #39552] quiver/quiver3 ignores scale with only one point to plot

[Nicholas Jankowski]

Follow-up Comment #19, bug #39552 (project octave):

yes, i see you were already on the inputParser issue. 

I can trigger a segfault on dev as well, even if I strip out my changeset. 
so, it doesn't seem my >= that did it. 

on 8.2.0 (on windows):

>> quiver(1,2,3,4,5) #no error, expected plot with one big arrow

>> close all

>> quiver(1,2,3,4,5,6) #same plot produced but with error
error: set: invalid number of arguments
error: called from
    __quiver__ at line 335 column 10
    quiver at line 92 column 10

>> close all

>> quiver(1,2,3,4,5,6,7)  #same plot produced but with error
error: set: argument 0 must be a property name


there's no input validation on quiver, so it makes sense that nonsensical
inputs for 6 and 7 will cause issues with set.

trying the same on dev (still rolled by from my patch below):

>> quiver(1,2,3,4,5) #no error, expected plot with one big arrow
clo
>> quiver(1,2,3,4,5,6) #figure produced before segfault
fatal: caught signal Segmentation fault -- stopping myself
Segmentation fault (core dumped)

>> quiver(1,2,3,4,5,6,7) #error, expected plot with one big arrow
fatal: caught signal Segmentation fault -- stopping myself
Segmentation fault (core dumped)


both 6 and 7 input cases, the error happens on the final call to set:

quiver line 102:

  unwind_protect_cleanup
    if (! isempty (oldfig))
      set (0, "currentfigure", oldfig);
    endif
  end_unwind_protect


i'm not sure why the 6 input case is reporting the error inside ___quiver___,
when i stepped through the last successful set call  was in __quiver__ at 335,
but no error and it finished ___quiver__ and stepped back up to quiver to hit
line above. error occured on that call to set. 

What would make previous errors segfault on dev? Does set use inputparser? 
i'll try stripping out rik's changes and see if the segfault repeats.


    _______________________________________________________

Reply to this item at:

  <https://savannah.gnu.org/bugs/?39552>

_______________________________________________
Message sent via Savannah
https://savannah.gnu.org/