[Compatibility] Add String#byteindex and String#byterindex #3043
Conversation
oracle-contributor-agreement
bot
added
the
OCA Verified
itarato
force-pushed
the
feature/PA-3039-string-byteindex-byterindex
branch
2 times, most recently
from
657d2a8 to
7f21400
Compare
|
Also it will be useful to run these Ruby 3.2 related specs on CI. To do so the spec files are supposed to be listed in the |
itarato
force-pushed
the
feature/PA-3039-string-byteindex-byterindex
branch
from
7f21400 to
61b5b60
Compare
itarato
force-pushed
the
feature/PA-3039-string-byteindex-byterindex
branch
2 times, most recently
from
ccb2732 to
2cfaa3a
Compare
Done |
itarato
force-pushed
the
feature/PA-3039-string-byteindex-byterindex
branch
from
2cfaa3a to
4ff1b34
Compare
itarato
force-pushed
the
feature/PA-3039-string-byteindex-byterindex
branch
from
4ff1b34 to
63e2718
Compare
andrykonchin
added
the
in-ci
| finish_adjusted = Primitive.byte_index_to_character_index(self, finish) | ||
| finish_adjusted += str.size | ||
| finish_adjusted = size if finish_adjusted > size | ||
| finish = Primitive.character_index_to_byte_index(self, finish_adjusted) |
There was a problem hiding this comment.
Can't this be something like finish += str.bytesize ?
I don't understand why this is necessary or what it does though.
There was a problem hiding this comment.
I guess I could have made the comment above better. Let's say the call is
("x" * 10).byterindex("xxx", 5)Ruby will start the lookup matching the pattern on index 5:
xxxxxxxxxx
xxx
StringByteReverseIndexNode on the other hand is matching the pattern at the end (non inclusive) on index 2:
xxxxxxxxxx
xxx
What makes this offset adjustment non trivial is the difference in encoding. Since str might have different char length, finish += str.bytesize can fall on a non codepoint boundary. Conceptually we need to adjust str.size characters in bytes. Is there a better way to do this?
There was a problem hiding this comment.
Ruby will start the lookup matching the pattern on index 5:
These docs are pretty confusing: https://docs.ruby-lang.org/en/master/String.html#method-i-byterindex
Integer argument offset, if given and non-negative, specifies the maximum starting byte-based position in the string to end the search:
But indeed:
> ("x" * 10).byterindex("xxx", 10)
=> 7
> ("x" * 10).byterindex("xxx", 9)
=> 7
> ("x" * 10).byterindex("xxx", 8)
=> 7
> ("x" * 10).byterindex("xxx", 7)
=> 7
> ("x" * 10).byterindex("xxx", 6)
=> 6
> ("x" * 10).byterindex("xxx", 5)
=> 5
> ("x" * 10).byterindex("xxx", 4)
=> 4
So conceptually it's like a maximum limit to the returned value.
In my understanding, adding finish = Primitive.min(finish + str.bytesize, self.bytesize) is correct then.
can fall on a non codepoint boundary.
Why does it matter? I think if TruffleString can handle it we can just ignore that.
If it can't we could keep increasing finish by 1 until that position is a character head.
There was a problem hiding this comment.
Why does it matter? I think if TruffleString can handle it we can just ignore that.
I had the assumption it matters. If we don't care and the code neither (tested, it doesn't), than I guess it's fine. Fixed.
andrykonchin
removed
the
in-ci
itarato
force-pushed
the
feature/PA-3039-string-byteindex-byterindex
branch
2 times, most recently
from
de5d8a7 to
7eb9724
Compare
Add String#byterindex Add tests
itarato
force-pushed
the
feature/PA-3039-string-byteindex-byterindex
branch
from
7eb9724 to
45d9756
Compare
Source: #3039
String#byteindex and String#byterindex have been added. [Feature #13110]