[webkit-dev] webkit tests fail results running on Windows (probably font-related)

nick yamane nick.yamane at openbossa.org
Tue Jun 25 18:09:56 PDT 2013 

Vladimir,

Do you have any PNG generated for that test when running with --pixel ?

I was running into a similar problem when using a custom WEBKITOUTPUTDIR..
I suspect that is bug on RWT/webkitpy.

Br,

*Nick Diego Yamane
Nokia Institute of Technology*


On Tue, Jun 25, 2013 at 2:03 PM, Vladimir Bushin <vladimirb at bsquare.com>wrote:

>  Nick,****
>
> ** **
>
> Yes:****
>
> ** **
>
> Tools/Scripts/run-webkit-tests –no-http
> Layout/css1/box_properties/margin_top.html****
>
> ** **
>
> Output:****
>
> *
> /cygdrive/c/Code/webkit-public/tmp/layout-test-results/css1/box_properties/margin_top-actual.txt
> *
>
> 1 layer at (0,0) size 785x1****
>
>  1layer at (0,0) size 785x1753****
>
> 22 RenderView at (0,0) size 785x600****
>
> 3 layer at (0,0) size 785x1****
>
> 4  RenderBlock {HTML} at (0,0) size 785x1****
>
> 5  RenderBody {BODY} at (8,8) size 769x1 [bgcolor=#CCCCCC]****
>
> 6  RenderBlock {P} at (0,0) size 769x****
>
> 7  RenderText {#text} at (0,0) size 355x****
>
>  3layer at (0,0) size 785x1753****
>
>  4 RenderBlock {HTML} at (0,0) size 785x1754****
>
>  5 RenderBody {BODY} at (8,8) size 769x1738 [bgcolor=#CCCCCC]****
>
>  6 RenderBlock {P} at (0,0) size 769x21****
>
>  7 RenderText {#text} at (0,0) size 355x20****
>
> 88 text run at (0,0) width 355: "The style declarations which apply to
> the text below are:"****
>
> 9  RenderBlock {PRE} at (0,3) size 769x120****
>
>  9 RenderBlock {PRE} at (0,34) size 769x120****
>
> 1010 RenderText {#text} at (0,0) size 392x120****
>
> 1111 text run at (0,0) width 384: ".zero {background-color: silver;
> margin-top: 0;}"****
>
> 1212 text run at (384,0) width 0: " "****
>
> ** **
>
> 2424 text run at (392,90) width 0: " "****
>
> 2525 text run at (0,105) width 200: "P, UL {margin-bottom: 0;}"****
>
> 2626 text run at (200,105) width 0: " "****
>
> 27  RenderBlock {HR} at (0,16) size 769x2 [border: (1px inset #000000)]***
> *
>
> 28  RenderBlock {P} at (0,17) size 769x [bgcolor=#C0C0C0]****
>
> 29  RenderText {#text} at (0,0) size 203x****
>
>  27 RenderBlock {HR} at (0,167) size 769x2 [border: (1px inset #000000)]**
> **
>
>  28 RenderBlock {P} at (0,177) size 769x21 [bgcolor=#C0C0C0]****
>
>  29 RenderText {#text} at (0,0) size 203x20****
>
> 3030 text run at (0,0) width 203: "This element has a class of zero."****
>
> 31  RenderBlock {P} at (0,19) size 769x [bgcolor=#C0C0C0]****
>
> 32  RenderText {#text} at (0,0) size 232x****
>
>  31 RenderBlock {P} at (0,198) size 769x21 [bgcolor=#C0C0C0]****
>
>  32 RenderText {#text} at (0,0) size 232x20****
>
> 3333 text run at (0,0) width 232: "This element also has a class of zero."
> ****
>
> 34  RenderBlock {P} at (0,2) size 769x [bgcolor=#00FFFF]****
>
> 35  RenderText {#text} at (0,0) size 759x****
>
>  34 RenderBlock {P} at (0,267) size 769x42 [bgcolor=#00FFFF]****
>
>  35 RenderText {#text} at (0,0) size 759x41****
>
> 3636 text run at (0,0) width 759: "This element should have a top margin
> of half an inch, which will require extra text in order to make sure that
> the margin"****
>
> 37  text run at (0,) width 155: "isn't applied to each line."****
>
> 38  RenderBlock {P} at (0,3) size 769x [bgcolor=#00FFFF]****
>
> 39  RenderText {#text} at (0,0) size 742x****
>
>  37 text run at (0,21) width 155: "isn't applied to each line."****
>
>  38 RenderBlock {P} at (0,334) size 769x42 [bgcolor=#00FFFF]****
>
>  39 RenderText {#text} at (0,0) size 742x41****
>
> 4040 text run at (0,0) width 742: "This element should have a top margin
> of 25 pixels, which will require extra text in order to make sure that the
> margin"****
>
> 41  text run at (0,) width 155: "isn't applied to each line."****
>
> 42  RenderBlock {P} at (0,4) size 769x [bgcolor=#00FFFF]****
>
> 43  RenderText {#text} at (0,0) size 745x****
>
>  41 text run at (0,21) width 155: "isn't applied to each line."****
>
> {…. Cropped …}****
>
> ** **
>
> For some reason the second coordinate is not always getting printed in the
> tree…****
>
> ** **
>
> Thanks,****
>
> vlad****
>
> ** **
>
> *From:* Nick Diego [mailto:nick.diego at gmail.com]
> *Sent:* Tuesday, June 25, 2013 12:18 PM
> *To:* Vladimir Bushin
> *Cc:* Dirk Pranke; webkit-dev at lists.webkit.org
> *Subject:* Re: [webkit-dev] webkit tests fail results running on Windows
> (probably font-related)****
>
> ** **
>
> Vladimir, ****
>
> Does the same problem happen with "release" builds ?****
>
> --****
>
> ** **
>
> On Tue, Jun 25, 2013 at 1:17 AM, Vladimir Bushin <vladimirb at bsquare.com>
> wrote:****
>
> Dirk,****
>
>  ****
>
> I’m running with and without “-p” option and in both cases have trouble
> passing the tests. Please look below where I post the difference in the
> tree-dump.****
>
> So in this case the command is looking like this:****
>
>  ****
>
> Tools/Scripts/run-webkit-tests --debug –no-http
> LayoutTests/css1/box_properties/margin_top.html****
>
> Thanks!****
>
> Vlad****
>
>  ****
>
> *From:* dpranke at google.com [mailto:dpranke at google.com] *On Behalf Of *Dirk
> Pranke
> *Sent:* Monday, June 24, 2013 1:26 PM
> *To:* Vladimir Bushin
> *Cc:* webkit-dev at lists.webkit.org
> *Subject:* Re: [webkit-dev] webkit tests fail results running on Windows
> (probably font-related)****
>
>  ****
>
> I'm not sure that any of the webkit.org ports actually run with
> pixel-tests enabled, so the port likely has outdated baselines and getting
> failures doesn't surprise me at all.****
>
>  ****
>
> -- Dirk****
>
>  ****
>
> On Mon, Jun 24, 2013 at 11:23 AM, Vladimir Bushin <vladimirb at bsquare.com>
> wrote:****
>
> Hi All,****
>
>  ****
>
> I’m working on submitting numerous patches for Webkit and need to figure
> out how to run all the regression tests.****
>
>  ****
>
> I have most of the webkit tests fail running on Windows with typically the
> following results:****
>
>  ****
>
> >Tools/Scripts/run-webkit-tests --debug –no-http –p
> LayoutTests/css1/box_properties/margin_top.html****
>
>  ****
>
> ****
>
>  ****
>
> The difference tree:****
>
> 1 layer at (0,0) size 785x1****
>
>  1layer at (0,0) size 785x1753****
>
> 22 RenderView at (0,0) size 785x600****
>
> 3 layer at (0,0) size 785x1****
>
> 4  RenderBlock {HTML} at (0,0) size 785x1****
>
> 5  RenderBody {BODY} at (8,8) size 769x1 [bgcolor=#CCCCCC]****
>
> 6  RenderBlock {P} at (0,0) size 769x****
>
> 7  RenderText {#text} at (0,0) size 355x****
>
>  3layer at (0,0) size 785x1753****
>
>  4 RenderBlock {HTML} at (0,0) size 785x1754****
>
>  5 RenderBody {BODY} at (8,8) size 769x1738 [bgcolor=#CCCCCC]****
>
>  6 RenderBlock {P} at (0,0) size 769x21****
>
>  7 RenderText {#text} at (0,0) size 355x20****
>
> 88 text run at (0,0) width 355: "The style declarations which apply to
> the text below are:"****
>
> 9  RenderBlock {PRE} at (0,3) size 769x120****
>
>  9 RenderBlock {PRE} at (0,34) size 769x120****
>
> 1010 RenderText {#text} at (0,0) size 392x120****
>
> 1111 text run at (0,0) width 384: ".zero {background-color: silver;
> margin-top: 0;}"****
>
> 1212 text run at (384,0) width 0: " "****
>
>  ****
>
> 2424 text run at (392,90) width 0: " "****
>
> 2525 text run at (0,105) width 200: "P, UL {margin-bottom: 0;}"****
>
> 2626 text run at (200,105) width 0: " "****
>
> 27  RenderBlock {HR} at (0,16) size 769x2 [border: (1px inset #000000)]***
> *
>
> 28  RenderBlock {P} at (0,17) size 769x [bgcolor=#C0C0C0]****
>
> 29  RenderText {#text} at (0,0) size 203x****
>
>  27 RenderBlock {HR} at (0,167) size 769x2 [border: (1px inset #000000)]**
> **
>
>  28 RenderBlock {P} at (0,177) size 769x21 [bgcolor=#C0C0C0]****
>
>  29 RenderText {#text} at (0,0) size 203x20****
>
> 3030 text run at (0,0) width 203: "This element has a class of zero."****
>
> 31  RenderBlock {P} at (0,19) size 769x [bgcolor=#C0C0C0]****
>
> 32  RenderText {#text} at (0,0) size 232x****
>
>  31 RenderBlock {P} at (0,198) size 769x21 [bgcolor=#C0C0C0]****
>
>  32 RenderText {#text} at (0,0) size 232x20****
>
> <the end of the tree is cropped…>****
>
>  ****
>
> As far as I can see the Actual text has couple additional pixels of a
> margin.****
>
> The first line “The style declarations….” Is just couple pixels below the
> Expected line.****
>
> The other lines (of the default font) are looking a little thinner than
> the Expected lines but do not seem to have the additional margin.****
>
> The rest of the text is looking similar but with the additional margin.***
> *
>
>  ****
>
>  ****
>
> I have Windows 7 Enterprise SP1 64 bit.****
>
> Safari version: 5.1.7 (7534.57.2)****
>
> Webkit branch: latest from git://gitorious.org/webkit/webkit.git****
>
> Compiler version: VS 2005 (8.0.50727.867)****
>
> CPU: Intel Xeon E5530 ****
>
>  ****
>
>  ****
>
>  ****
>
> I have cloned the original webkit from:****
>
> git://gitorious.org/webkit/webkit.git****
>
>  ****
>
> the fonts were taken from:****
>
> git://gitorious.org/qtwebkit/testfonts.git****
>
>  ****
>
> and I set the environment variable WEBKIT_TESTFONTS to point to the folder
> with the fonts. The ‘set’ command displays the right path after restart of
> Cygwin console.****
>
> Didn’t help.****
>
>  ****
>
> Also in order to fix the problem I tried to get fonts from ‘leopard’ and
> convert them following these instructions:****
>
> http://trac.webkit.org/wiki/BuildingOnWindows#Font-metric-relatedfailures*
> ***
>
> It didn’t help****
>
>  ****
>
> tried to substitute system font folder with the folder with fonts from
> gitorious. Didn’t help either****
>
>  ****
>
> The most useful info in archive was this:
> https://lists.webkit.org/pipermail/webkit-dev/2011-April/016427.html****
>
> and the last line there that it might be a bug.****
>
>  ****
>
> However WebKit BuildBot builds on Windows somehow.****
>
> I’m wondering what is in this folder:
> WEBKIT_TESTFONTS=C:\cygwin\home\buildbot\slave\WebKitTestFonts****
>
>  ****
>
>  ****
>
> Thanks for all the suggestions!****
>
> Vlad****
>
>
> _______________________________________________
> webkit-dev mailing list
> webkit-dev at lists.webkit.org
> https://lists.webkit.org/mailman/listinfo/webkit-dev****
>
>  ****
>
>
> _______________________________________________
> webkit-dev mailing list
> webkit-dev at lists.webkit.org
> https://lists.webkit.org/mailman/listinfo/webkit-dev****
>
>
>
>
> --
> ------------------------------
> Nick Diego Yamane
> +55(92)91645045 ****
>
> _______________________________________________
> webkit-dev mailing list
> webkit-dev at lists.webkit.org
> https://lists.webkit.org/mailman/listinfo/webkit-dev
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-dev/attachments/20130625/9b8f7f55/attachment.html>

More information about the webkit-dev mailing list