Bug 164064 – testharnessreport.js should sanitize the results before printing them

• Status: RESOLVED
• Resolution: FIXED
• Component: Tools / Tests

[Chris Dumez]

testharnessreport.js should sanitize the results before printing them. We currently have 3 copies of this script and only 1 does the sanitization.
Short term, let do the sanitization in all of them. Longer term, we should merge these and have a way to keep them in sync.

[Chris Dumez]

Created attachment 293038 [details]
Patch

[youenn fablet]

If it proves to be difficult to remove the testharness report.js copies, we could add checks.
We do that for testharness.js when running rwt.
In the case of test harness report.js, style checker could do it.

[Darin Adler]

Comment on attachment 293038 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=293038&action=review

Thanks so much for fixing this!

> LayoutTests/http/tests/w3c/resources/testharnessreport.js:56
> +            // Escape null characters, otherwise diff will think the file is binary.

This comment is not the reason this matters to us. Seems like a comment written by whoever made this change in Blink and so only mentions the reason that affects them.

> LayoutTests/http/tests/w3c/resources/testharnessreport.js:58
> +            // Escape carriage returns as they break rietveld's difftools.

This comment is even more Blink-specific. I don’t even know what “rietveld” is.

[WebKit Commit Bot]

Comment on attachment 293038 [details]
Patch

Clearing flags on attachment: 293038

Committed r207995: <http://trac.webkit.org/changeset/207995>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[Chris Dumez]

(In reply to comment #3)
> Comment on attachment 293038 [details]
> Patch
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=293038&action=review
> 
> Thanks so much for fixing this!
> 
> > LayoutTests/http/tests/w3c/resources/testharnessreport.js:56
> > +            // Escape null characters, otherwise diff will think the file is binary.
> 
> This comment is not the reason this matters to us. Seems like a comment
> written by whoever made this change in Blink and so only mentions the reason
> that affects them.
> 
> > LayoutTests/http/tests/w3c/resources/testharnessreport.js:58
> > +            // Escape carriage returns as they break rietveld's difftools.
> 
> This comment is even more Blink-specific. I don’t even know what “rietveld”
> is.

Sorry about this. I went quick and dirty. I am planning to sync-up all 3 scripts so they are identical in a follow-up when I have a little bit more time. I'll clean up the comments as well.