Bug 64410 - ApplicationCache Quota Output is Flakey
Summary: ApplicationCache Quota Output is Flakey
Alias: None
Product: WebKit
Classification: Unclassified
Component: Tools / Tests (show other bugs)
Version: 528+ (Nightly build)
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Joseph Pecoraro
Depends on:
Reported: 2011-07-12 16:48 PDT by Joseph Pecoraro
Modified: 2011-07-13 15:58 PDT (History)
5 users (show)

See Also:

[PATCH] Starting Point (836 bytes, patch)
2011-07-12 17:51 PDT, Joseph Pecoraro
no flags Details | Formatted Diff | Diff
[PATCH] Truncate space needed to ten thousands (11.19 KB, patch)
2011-07-12 21:40 PDT, Joseph Pecoraro
ap: review+
Details | Formatted Diff | Diff

Note You need to log in before you can comment on or make changes to this bug.
Description Joseph Pecoraro 2011-07-12 16:48:45 PDT
I'll skip these for now and then attempt fixing this cleanly for the different ports that implement them.
Comment 2 Joseph Pecoraro 2011-07-12 17:09:18 PDT
Skipped the tests temporarily in:

I didn't think printing out the exact number of bytes was safe, but
I thought the same code was shared in WebCore.

First, I need to fix the printf format in DumpRenderTreeQt, then
I should make printing the number of bytes more robust.
Comment 3 Joseph Pecoraro 2011-07-12 17:48:57 PDT
Adding some Qt developers involved in adding ApplicationCache Quotas for Qt.

Now, back to the drawing board.
Comment 4 Joseph Pecoraro 2011-07-12 17:51:45 PDT
Created attachment 100598 [details]
[PATCH] Starting Point

For starters, I put the newline in the wrong spot and then replaced it with a space.
This patch puts the newline in the right spot. But that still doesn't handle the
# of bytes being different, which is probably brittle to begin with. I'm going
to investigate that a bit more.
Comment 5 Joseph Pecoraro 2011-07-12 20:44:31 PDT
estimatedSizeInStorage includes some HTTP response data, which of course
could vary with each request and across platforms. I can't include the # directly.
Comment 6 Joseph Pecoraro 2011-07-12 20:56:08 PDT
Skipped on platform/mac as well.

I had missed one test on run-webkit-tests after a rename as well.
I'd still like some value in these tests. Maybe approximating to
10000 bytes would work? Instead of 31003*bytes something "like
~30000 bytes" would work. Or just not printing any number...
Comment 7 Joseph Pecoraro 2011-07-12 21:40:47 PDT
Created attachment 100623 [details]
[PATCH] Truncate space needed to ten thousands
Comment 8 Joseph Pecoraro 2011-07-12 23:37:20 PDT
Comment on attachment 100623 [details]
[PATCH] Truncate space needed to ten thousands

View in context: https://bugs.webkit.org/attachment.cgi?id=100623&action=review

bdash pointed out a grammar mistake.

> Tools/DumpRenderTree/mac/UIDelegate.mm:173
> +        // Rounding up or down down not really matter for these tests. Its

I will fix this "up or down down" => "up or down" when landing.

> Tools/DumpRenderTree/qt/DumpRenderTreeQt.cpp:1046
> +        // Rounding up or down down not really matter for these tests. Its

I will fix this "up or down down" => "up or down" when landing.
Comment 9 Joseph Pecoraro 2011-07-13 14:42:15 PDT
Landed in <http://trac.webkit.org/changeset/90946>.
Watching the bots.
Comment 10 Joseph Pecoraro 2011-07-13 14:48:23 PDT
Caught this early: <http://trac.webkit.org/changeset/90951>

That "\n" is going to be the end of me. I've already moved it, removed it,
and now forgotten to add it. Sorry about all of the churn related to this.
Comment 11 Joseph Pecoraro 2011-07-13 15:58:14 PDT
This now looks good on Qt and Mac. Yay!