TestResultsServer should keep old test results

[Adam Barth]

TestResultsServer should keep old test results

[Adam Barth]

Created attachment 100155 [details]
Patch

[Ojan Vafai]

Comment on attachment 100155 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=100155&action=review

> Tools/ChangeLog:8
> +        Having historical data will help use do failure archeology.

s/use/us

> Tools/TestResultServer/handlers/testfilehandler.py:107
> +        if callback_name:

If callback_name is optional, it should default to None above, right?

> Tools/TestResultServer/handlers/testfilehandler.py:133
> +        file = db.get(key)

Not entirely sure what this does. What are valid values for "key"?

> Tools/TestResultServer/templates/showfilelist.jsonp:4
> +{% for file in files %}
> +  "{{ file.key }}",
> +{% endfor %}

Do we want to use a richer format for this? For example, do we want to include the metadata we have about each file? Basically the data we currently show at http://test-results.appspot.com/testfile.

[Adam Barth]

> > Tools/TestResultServer/handlers/testfilehandler.py:133
> > +        file = db.get(key)
> 
> Not entirely sure what this does. What are valid values for "key"?

AppEngine datastore keys.

> > Tools/TestResultServer/templates/showfilelist.jsonp:4
> > +{% for file in files %}
> > +  "{{ file.key }}",
> > +{% endfor %}
> 
> Do we want to use a richer format for this? For example, do we want to include the metadata we have about each file? Basically the data we currently show at http://test-results.appspot.com/testfile.

I'll add some more structure here so that it is future-proof.

[Adam Barth]

Created attachment 100159 [details]
Patch

[Adam Barth]

Committed r90665: <http://trac.webkit.org/changeset/90665>