Phantom changes regions boundaries (breaking view.extract_scope method)

[vovkkk]

Expected behavior

view.extract_scope method returns same result.

Actual behavior

view.extract_scope method returns different results depending on phantom presence/absence.

Steps to reproduce

1. create new view, set Python syntax via palette, type class and autocomplete

   class ClassName(object):
       """docstring for ClassName"""
       def __init__(self, arg):
           super(ClassName, self).__init__()
           self.arg = arg

2. open console and execute
   a. import pprint
   b. pprint.pprint([[i, view.extract_scope(i)] for i in range(24)])
   c. phantoms = sublime.PhantomSet(view)
   d. phantoms.update([sublime.Phantom(sublime.Region(22), 'yo', sublime.LAYOUT_INLINE)])
   e. pprint.pprint([[i, view.extract_scope(i)] for i in range(24)])

3. compare results of 2.b and 2.e (0-14 are omitted)

point	b	e
15	(15, 22)	(15, 22)
16	(16, 22)	(15, 22)
17	(16, 22)	(15, 22)
18	(16, 22)	(15, 22)
19	(16, 22)	(15, 22)
20	(16, 22)	(15, 22)
21	(16, 22)	(15, 22)
22	(16, 23)	(22, 24)
23	(22, 24)	(22, 24)

Note, size and layout of phantom does not affect the results in any manner

Environment

• Operating system and version:
  • Windows 8.1
  • Mac OS ...
  • Linux ...
• Sublime Text:
  • Build 3126

[keith-hall]

I think that view.extract_scope has always been notoriously dodgy anyway, and it is recommended to manually check and compare the scope at each position yourself...

[vovkkk]

After some more testing I figured that phantom actually splits a scope.
Just a note, if you decide to change title or something

[deathaxe]

API: Fixed view.expand_to_scope stopping at phantoms

Maybe fixed by ST 4134?