-
Notifications
You must be signed in to change notification settings - Fork 2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Grails 2.3.3 upgrade and implementation for all test types #603
Conversation
@lhotari I pulled your latest and it seems that it is still getting errors: Log - https://gist.github.com/msmith-techempower/7571534 |
I'll test and fix the problem tomorrow and send a new PR. -Lari |
@lhotari I don't think you need a new PR; if you update your fork's master branch, then I should be able to pull and test out your changes with this existing PR. Basically, just keep making commits, and post a comment back here (tag me) when you are ready for me to test again. |
- Mysql JDBC driver causes a lot of SQLExceptions internally which cause a lot of overhead
- enabling holdResultsOpenOverStatementClose helps - see com.mysql.jdbc.PreparedStatement:2265 (5.1.27 version source)
- It is not acceptable to retrieve all required rows using a SELECT ... WHERE id IN (...) clause.
OK. I set up vagrant for testing that the benchmark can be run. I sent a separate PR #621 about that. |
I found the Resin error message
This is a Grails bug http://jira.grails.org/browse/GRAILS-10577 which should be fixed in Grails 2.3.3 which is planned to be released today. I'll update the PR to Grails 2.3.3 and test it. |
Hi, Grails 2.3.3 has been published so this PR is ready to be merged. Lari |
@lhotari I pulled your latest and ran the tests. Everything looks good except that your Here is your log output - https://gist.github.com/msmith-techempower/7590958 As you can see, there's that odd I bet that's a unicode comma. The output is supposed to be "Hello, World!". OKAY - turn that into an ascii comma and upper-case your 'w' in world, and I will accept this PR. |
@lhotari I just thought of one more thing you need to add to your pull request - Remove the line in the |
@msmith-techempower
there is "json_verify" in yajl-tools on Ubuntu (apt-get install yajl-tools) |
I optimized JSON output in Grails 2.3.3. Here's the change in Grails if anyone is interested: The goal of the change was: "Improve "render obj as JSON" performance by doing minor refactorings and streaming the JSON output without doing unnecessary String conversions." The string value escaping was changed to use a streaming version of character escaping which is already used for escaping string values in javascript values embedded in HTML templates (script elements). It replaces a lot of characters to prevent different kind of XSS exploits that are possible in values that are placed in embedded HTML templates (script elements). The escaping of these characters doesn't make sense in JSON output, so I agree that this is a minor glitch. |
@lhotari My comment here is getting off-topic for this pull request... it's more of a question/suggestion for Grails. It has nothing to do with the correctness of the benchmark. At a quick glance, it looks like your JavaScriptEncoder.escapeCharacter is missing some cases:
Just some friendly suggestions. I wrote a similar class for our internal framework and remember having to deal with those. |
@michaelhixson Thanks for the suggestions! |
@msmith-techempower I have now fixed the JSON output by backporting the fix from grails-core 2.3.x branch. It's in commit 61e4a03 . I think this is now ready for Round 8 benchmarks. Thanks for the help!
|
@lhotari Wait, are you saying that in the normal release of Grails there is no way to send an ascii comma in a string -- that it will be replaced with a unicode comma? Unrelated - why is there a percent symbol at the end of your json? EDIT: Sorry if that came off impolite -- I'm honestly just curious about these things. I tested out your latest and accepted the PR. |
Grails 2.3.3 upgrade and implementation for all test types
The percent sign is from curl -i mode. |
@lhotari I had a hunch that that was some curl noise at the end. Glad to hear that the escaping issue is fixed in the next round. |
This contains Grails 2.3.3 upgrade and implementation for all test types.