RFR: 8282429: StringBuilder/StringBuffer.toString() skip compressing for UTF16 strings [v3]

Claes Redestad redestad at openjdk.java.net
Tue Mar 22 10:12:40 UTC 2022


On Tue, 22 Mar 2022 06:54:35 GMT, Xin Liu <xliu at openjdk.org> wrote:

>> src/java.base/share/classes/java/lang/AbstractStringBuilder.java line 1008:
>> 
>>> 1006:         this.count = newCount;
>>> 1007:         putStringAt(start, str);
>>> 1008:         if (end - start > 0) {
>> 
>> regardless of value of `end - start` you could also skip setting `maybeLatin1 = true` if:
>> - `str.coder() == UTF16`
>> - `this.coder == LATIN1`
>
> hi, @cl4es, 
> you are correct for this and the comment at setCharAt(), but I don't think it's necessary to check all cases. this attribute is just like a hint. if this.coder == LATIN1, that it doesn't matter if maybeLatin1 is true. 
> 
> if our attitude is checking all cases, it will become too complex and error-prone. deleteCharAt() and setLength() also need to check.  if so, it will pollute code more. I incline to set this attribute conservatively in all deleting methods.

Right, we don't need to check every case and I agree with favoring simplicity at the expense of some false positives. Perhaps the `if (end - start > 0) {` test here isn't pulling its weight either and we should just unconditionally set `maybeLatin1 = true` even if we're not actually replacing anything (which is very much a corner-case).

-------------

PR: https://git.openjdk.java.net/jdk/pull/7671


More information about the core-libs-dev mailing list