8349183: [BACKOUT] Optimization for StringBuilder append boolean & null

[jaikiran]

Can I please get a review of this change which backs out the commit that was introduced for https://bugs.openjdk.org/browse/JDK-8333893?

The comment in the PR review of that issue #19626 (comment) explains what the issue is with the change that was integrated. Furthermore, one part of that original change introduced a few internal methods. These internal methods were then used in few other places within the JDK through https://bugs.openjdk.org/browse/JDK-8343650. As a result, this backout PR also reverts the change that was done in JDK-8343650.

The backout was done as follows, using git revert against the 2 relevant commits:

git revert 74ae3c688b37e693e20eb4e17c631897c5464400
git revert 5890d9438bbde88b89070052926a2eafe13d7b42

The revert of 5890d9438bbde88b89070052926a2eafe13d7b42 wasn't clean and I had to resolve a trivial conflict in StringLatin1.java.

tier1, tier2 and tier3 testing is currently in progress with this change. Once this is integrated into mainline, a corresponding backport will be done to jdk24 branch with the requisite approvals.

---

Progress

• Change must be properly reviewed (1 review required, with at least 1 Reviewer)
• Change must not contain extraneous whitespace
• Commit message must refer to an issue

Issues

• JDK-8349183: [BACKOUT] Optimization for StringBuilder append boolean & null (Sub-task - P2)
• JDK-8349239: [BACKOUT] Reuse StringLatin1::putCharsAt and StringUTF16::putCharsAt (Bug - P4)

Reviewers

• Claes Redestad (@cl4es - Reviewer)
• Chen Liang (@liach - Reviewer)

Reviewers without OpenJDK IDs

• @abdelhak-zaaim (no known openjdk.org user name / role)

Reviewing

Using git

Checkout this PR locally:
$ git fetch https://git.openjdk.org/jdk.git pull/23420/head:pull/23420
$ git checkout pull/23420

Update a local copy of the PR:
$ git checkout pull/23420
$ git pull https://git.openjdk.org/jdk.git pull/23420/head

Using Skara CLI tools

Checkout this PR locally:
$ git pr checkout 23420

View PR using the GUI difftool:
$ git pr show -t 23420

Using diff file

Download this PR as a diff file:
https://git.openjdk.org/jdk/pull/23420.diff

Using Webrev

Link to Webrev Comment

[bridgekeeper]

👋 Welcome back jpai! A progress list of the required criteria for merging this PR into master will be added to the body of your pull request. There are additional pull request commands available for use with this pull request.

[openjdk]

@jaikiran This change now passes all automated pre-integration checks.

ℹ️ This project also has non-automated pre-integration requirements. Please see the file CONTRIBUTING.md for details.

After integration, the commit message for the final commit will be:

8349183: [BACKOUT] Optimization for StringBuilder append boolean & null
8349239: [BACKOUT] Reuse StringLatin1::putCharsAt and StringUTF16::putCharsAt

Reviewed-by: redestad, liach

You can use pull request commands such as /summary, /contributor and /issue to adjust it as needed.

At the time when this comment was updated there had been 1 new commit pushed to the master branch:

• bb837d2: 8342775: [Graal] java/util/concurrent/locks/Lock/OOMEInAQS.java fails OOME thrown from the UncaughtExceptionHandler

Please see this link for an up-to-date comparison between the source branch of this pull request and the master branch.
As there are no conflicts, your changes will automatically be rebased on top of these commits when integrating. If you prefer to avoid this automatic rebasing, please check the documentation for the /integrate command for further details.

➡️ To integrate this PR with the above commit message to the master branch, type /integrate in a new comment.

[openjdk]

@jaikiran The following labels will be automatically applied to this pull request:

• core-libs
• hotspot-compiler

When this pull request is ready to be reviewed, an "RFR" email will be sent to the corresponding mailing lists. If you would like to change these labels, use the /label pull request command.

[mlbridge]

Webrevs

• 01: Full - Incremental (23d01f0a)
• 00: Full (d406bc1d)

[wenshao]

This concurrency problem also exists in the UTF16 scenario, so why only change to Latin1 here?

[jaikiran]

This concurrency problem also exists in the UTF16 scenario, so why only change to Latin1 here?

Do you mean there are additional commits that have been done in the JDK which introduce a similar issue related to array writes beyond their limit that need to be backed out?

[cl4es]

@wenshao this re-instates the checkBoundsBeginEnd(i, end, value); in the UTF16 case that was removed by the issue being backed out, so we get back to a state where we have appropriate bounds checks on all array accesses.

An alternative to this backout would be to add checkBoundsBeginEnd(i, end, value); to all the putCharsAt methods, though it's unclear if that would undo the performance advantage. Better then to backout and - if possible - redo with a closer examination of the performance with a safer construct.

[liach]

We have this explicit check for null and true, false has another bound check. This backout version should be safe.

[liach]

Also for append(char): it uses putCharSB which does a check and calls unchecked putChar. putCharsAt is always checked.

[jaikiran]

Thank you Claes and Chen for the reviews. tier1, tier2 and tier3 testing is nearing completion without any failures. Could one of you approve that it's OK to integrate this trivial backout without waiting for 24 hours (the review process allows it https://openjdk.org/guide/#trivial-changes)? I will be integrating this as soon as the tier testing completes.

[wenshao]

Can you wait for me for a while? I am looking for other solutions that do not require a fallback.

[jaikiran]

Can you wait for me for a while? I am looking for other solutions that do not require a fallback.

The updated fix/change (if any) doesn't have to be rushed and you can take longer to work on it with additional help and reviews from others.

Once this backout is integrated into mainline, it will be backported to jdk24 (which is to be released in a few days). The risk of waiting for an additional/different fix is higher compared to a backout. So we intend to go ahead with the backout.

[wenshao]

#23423
I submitted another PR where I reproduced the issue locally and this PR fixed the issue.

[JesperIRL]

Thank you Claes and Chen for the reviews. tier1, tier2 and tier3 testing is nearing completion without any failures. Could one of you approve that it's OK to integrate this trivial backout without waiting for 24 hours (the review process allows it https://openjdk.org/guide/#trivial-changes)? I will be integrating this as soon as the tier testing completes.

A backout is always considered a trivial change. See https://openjdk.org/guide/#backing-out-a-change

[jaikiran]

/issue add JDK-8349239

[openjdk]

@jaikiran
Adding additional issue to issue list: 8349239: [BACKOUT] Reuse StringLatin1::putCharsAt and StringUTF16::putCharsAt.

[wenshao]

Don't rush to roll back, the current rollback solution still has problems

[jaikiran]

Thank you all for the help on this one.

/integrate

[openjdk]

Going to push as commit 618c5eb.
Since your change was applied there has been 1 commit pushed to the master branch:

• bb837d2: 8342775: [Graal] java/util/concurrent/locks/Lock/OOMEInAQS.java fails OOME thrown from the UncaughtExceptionHandler

Your commit was automatically rebased without conflicts.

[openjdk]

@jaikiran Pushed as commit 618c5eb.

💡 You may see a message that your pull request was closed with unmerged commits. This can be safely ignored.

[wenshao]

The problem still exists, we need to complete another PR #23427