Repository navigation
Support endless ranges in where - #34906
Conversation
There was a problem hiding this comment.
I think this will give us a syntax error on 2.5. Can you conditionalize this test so we don't get build failures on 2.5?
There was a problem hiding this comment.
Sure. The built actually failed for precisely this reason. I'll push an update tomorrow.
There was a problem hiding this comment.
D'oh, what I did won't work for syntax error 🙈
I can either:
- Come up with something that would work for syntax errors.
- Leave the check as it is and replace
0..with0..nil(they're equivalent). In Rubyies pre-2.6 the latter results in anArgumentErrorexception.
@tenderlove what do you think?
3a67c12 to
20b108a
Compare
There was a problem hiding this comment.
May I suggest this?
| node = attribute.between(0..) | |
| node = attribute.between(eval("0..")) # Use eval for compatibility with Ruby < 2.6 parser. |
This way Ruby parser in version < 2.6 will just overlook it.
There was a problem hiding this comment.
Great idea. Thanks!
905d11e to
fbccf87
Compare
sikachu
left a comment
sikachu
left a comment
There was a problem hiding this comment.
Would you mind checking CodeClimate? Looks like in the test, line 642 and 645 should both use double quote and not single quote.
This commit adds support for endless ranges, e.g. (1..), that were added in Ruby 2.6. They're functionally equivalent to explicitly specifying Float::INFINITY as the end of the range.
fbccf87 to
7110dbe
Compare
|
@tenderlove @sikachu thanks for comments! The build is green now. Please have another look. |
sikachu
left a comment
sikachu
left a comment
There was a problem hiding this comment.
I think this is good to merge. @tenderlove, please!
And support endless ranges for `not_between` like as `between`. Follow up #34906.
Support endless ranges in where rails/rails#34906
3 participants
Summary
This commit adds support for endless ranges, e.g.
(1..), that were added in Ruby 2.6. They're functionally equivalent to explicitly specifyingFloat::INFINITYas the end of the range.Specifically:
is equivalent to
Other Information
Only the end of a range can be omitted. Trying
(..1)results in a syntax error so there's no need to add support for implicit-Float::INFINITY.Alternative Implementation
We could modify
otherat the beginning ofbetweenand replacenilwithFloat::INFINITY(or even support-Float::INFINITY) so that there'd be no changes to conditions inifstatements below + we would be able to supportnil..1(equivalent to-Float::INFINITY..1).Please let me know if you find this preferable and I'll update the PR.