Fix #1623: Fix an infinite loop in j.i.RandomAccessFile#readLine #2100
Conversation
|
@WojciechMazur Thank you for the timely review & approval. Always nice to have a second set of eyes, especially Not trying to push the river, but rather to understand the process. Do I need to do anything to get this merged or just |
| var end = length() // standard practice: 1 past last valid byte. | ||
| if (pos >= end) { | ||
| null // JDK 8 specification requires null here. | ||
| } else { | ||
| val builder = new StringBuilder |
There was a problem hiding this comment.
This isn't specifically part of the PR but I think we should be using java.lang.StringBuilder here to remove the Scala lib dependency.
There was a problem hiding this comment.
Good suggestion, good eyes! Thank you. Review leads to better code all around.
Since that line is new code, I think that making the change you suggest does not
break the "unity of purpose" required of a PR.
Easy enough to write to current practices and not introduce technical debt.
Other Scala lib dependencies may have crept in but I know of no easy way
to find them all at once. At least this one is gone. Others as they are found
(We will have to have a planned, but "Turn off Predef" bugfix week sometime).
We correct the condition which lead to a user reporting an infinite loop. Previously, two-byte Chars were being used where 1-byte Bytes should have been. This PR updates and supersedes now closed PR scala-native#1624. The latter has extensive notes which may aid future developers.
LeeTibbert
force-pushed
the
PR_j_i_RandomAccessFileLoop_I1623_V2
branch
from
ddb07c0
to
b9205c4
Compare
sjrd
changed the title
…eadLine (scala-native#2100) We correct the condition which lead to an infinite loop. Previously, we read two-byte Chars instead of 1-byte Bytes.
…eadLine (scala-native#2100) We correct the condition which lead to an infinite loop. Previously, we read two-byte Chars instead of 1-byte Bytes.
We correct the condition which lead to an infinite loop.
Previously, two-byte Chars were being used where 1-byte Bytes
should have been.
This PR updates and supersedes now closed PR #1624. The latter
has extensive notes which may aid future developers.