Summary: | Output more descriptive PASS messages for list layout tests | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Daniel Bates <dbates> | ||||||
Component: | Tools / Tests | Assignee: | Daniel Bates <dbates> | ||||||
Status: | RESOLVED FIXED | ||||||||
Severity: | Normal | CC: | darin | ||||||
Priority: | P2 | ||||||||
Version: | 528+ (Nightly build) | ||||||||
Hardware: | All | ||||||||
OS: | All | ||||||||
Bug Depends on: | 42803 | ||||||||
Bug Blocks: | |||||||||
Attachments: |
|
Description
Daniel Bates
2010-07-24 17:27:58 PDT
Created attachment 62514 [details] Patch I thought it would be easier to review the changes to the files LayoutTests/fast/lists/resources/dump-list.js, LayoutTests/fast/lists/w3-css3-list-styles-alphabetic-expected.txt, and LayoutTests/fast/lists/w3-css3-list-styles-numeric-expected.txt in a separate patch than to post an updated patch for Bug #42803. My plan is to land Bug #42803 (as is). Then this bug (Bug #42938). Then, eventually Bug #36734. I am open to suggestions on how to make the PASS messages more descriptive. Notice, we only output the PASS message if the list item marker matches the list item text for each <li> in the list. Without loss of generality, consider the following list: <ol start="5"> <li>5</li> <li>6</li> </ol> Then the PASS message would be: "PASS list markers matched list items for list that starts from 5 and has 2 items." I thought to output the PASS message on a per list basis instead of a per <li> basis so as to make the output easier to read (since there would be less lines to read). Comment on attachment 62514 [details]
Patch
I guess this is OK. Better would be to include the strings of the markers that were tested. The danger of a test that compares two invisible things is that somehow they could both be wrong.
(In reply to comment #2) > (From update of attachment 62514 [details]) > I guess this is OK. Better would be to include the strings of the markers that were tested. The danger of a test that compares two invisible things is that somehow they could both be wrong. Would it be better to print PASS/FAIL for each <li>? It would be verbose but it seems like this would be more beneficial. Created attachment 62527 [details]
Patch
Updated patch to output PASS message for each <li>. Also, separated out common code in dump-list.js and LayoutTests/lists/w3-css3-list-styles-numeric.html into function testListItemMarkerEquals.
Comment on attachment 62527 [details] Patch Clearing flags on attachment: 62527 Committed r64259: <http://trac.webkit.org/changeset/64259> All reviewed patches have been landed. Closing bug. |