I suspect the case of a single character either never occurred or when
it did, it didn't matter.
IIRC, the "remove trailing whitespace" is called to trim any whitespace
after the period that ends the first sentence, which is unlikely to be a
single period! In addition, excess whitespace is often "ignored" in
HTML, unless it is within a <pre> block or equivalent.
Anyway, it's been interesting to discover these issues as a side-effect
of what was intended as just a code cleanup!
-- Jon
On 12/18/19 7:32 AM, Hannes Wallnöfer wrote:
Yes, I intended it in the way Claes explained it, and I also assumed this was a
bug.
In the case of removeTrailingWhitespace even something like "x " would return
unchanged, because the for-loop stops before reaching the first character. I wonder if
this case of a single non-whitespace text never occurs, or if just nobody ever noticed it?
Hannes
Am 18.12.2019 um 16:18 schrieb Jonathan Gibbons <[email protected]>:
Claes,
Thanks for the clarification. It sounds like the old behavior was a bug.
I note that I did do before/after comparison for the main docs build
and for all the many runs of javadoc performed by the jtreg tests,
and all compared OK.
-- Jon
On 12/18/19 7:11 AM, Claes Redestad wrote:
Sorry for jumping in: I initially misread Hannes' observation in the
same way, but I think he meant that " ".strip() will return "", while
removeTrailingWhitespace(" ") will return " ", which is a more
subtle difference in behavior.
/Claes
On 2019-12-18 15:46, Jonathan Gibbons wrote:
Hannes,
Thanks for the review.
My reading of the code for String.strip.... is that the original string is
returned if there is no whitespace to be removed ... in other words, the new
methods should have the same behavior as the removed methods.
-- Jon
On 12/18/19 4:56 AM, Hannes Wallnöfer wrote:
Looks good.
I noticed that the removed methods would return the original string if they
didn’t encounter a non-whitespace character, but I guess that wasn’t actually
part of their intended behaviour.
Hannes
Am 16.12.2019 um 23:47 schrieb Jonathan Gibbons <[email protected]>:
Updated webrev based on Ivan's suggestion to use one of stripLeading(),
stripTrailing(), or strip() instead of the old methods, which are now removed
-- Jon
http://cr.openjdk.java.net/~jjg/8236030/webrev.01/
On 12/16/2019 01:33 PM, Jonathan Gibbons wrote:
Ivan,
Great suggestion, and sets up the possibility to use strip() when both leading
and trailing whitespace should be removed.
The only slight change in semantics would be that these methods work in terms
of code-points, not characters, and javadoc has not (yet?) been adapted to use
code-points throughout.
-- Jon
On 12/16/2019 01:23 PM, Ivan Gerasimov wrote:
Hello!
Can String.stripLeading()/stripTrailing() methods be used instead, or is there
a reason to avoid them?
With kind regards,
Ivan
On 12/16/19 1:10 PM, Jonathan Gibbons wrote:
Please review a tiny change to eliminate a string copy into a temporary
character buffer when removing leading or trailing whitespace from a string.
The affected methods are called within the main code to translate doc comments
to HTML. This is noreg-cleanup/no-reg-trivial, so no new additional tests.
Webrev below, but the patch is also included here:
---
a/src/jdk.javadoc/share/classes/jdk/javadoc/internal/doclets/formats/html/HtmlDocletWriter.java
Mon Dec 16 15:33:03 2019 -0500
+++
b/src/jdk.javadoc/share/classes/jdk/javadoc/internal/doclets/formats/html/HtmlDocletWriter.java
Mon Dec 16 13:07:18 2019 -0800
@@ -1604,22 +1604,19 @@
}
private String removeTrailingWhitespace(String text) {
- char[] buf = text.toCharArray();
- for (int i = buf.length - 1; i > 0 ; i--) {
- if (!Character.isWhitespace(buf[i]))
- return text.substring(0, i + 1);
+ int i = text.length() - 1;
+ while (i >= 0 && Character.isWhitespace(text.charAt(i))) {
+ i--;
}
- return text;
+ return i == text.length() - 1 ? text : text.substring(0, i + 1);
}
private String removeLeadingWhitespace(String text) {
- char[] buf = text.toCharArray();
- for (int i = 0; i < buf.length; i++) {
- if (!Character.isWhitespace(buf[i])) {
- return text.substring(i);
- }
+ int i = 0;
+ while (i < text.length() && Character.isWhitespace(text.charAt(i))) {
+ i++;
}
- return text;
+ return i == 0 ? text : text.substring(i);
}
/**
-- Jon
JBS: https://bugs.openjdk.java.net/browse/JDK-8236030
Webrev: http://cr.openjdk.java.net/~jjg/8236030/webrev/
