BPWG checker task force Teleconference

25 Jun 2008


See also: IRC log


abel, miguel, Dom, francois

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5006 bug 5006 "Does a "tidied" element or attribute exist? "

abel: the goal of this was to inform the user whether the test was run on a tidied version of the document or not
... but that's not required by mobileOK
... so it's up to us to decide

dom: I think we should move it as "enhancement" and re-assign it
... I'll take it

abel: sounds good

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5015 bug 5015 Need working example for accessing line number in XSL

abel: in Java, we can access the line number
... but that's not available in XSLT
... since it doesn't work on a DOM tree
... I don't think we can do much about it

<francois> http://saxonica.com/documentation/extensions/functions/linenumber.html

dom: I think Saxon has an extension to get the line number
... but it didn't work when Roland tried, as far as I remember
... let's close it for now and see if we have time for this in the future

abel: agreed

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5421 bug 5421 URI resolution doesn't take <base href> into account

abel: does this still need more work?

dom: the Java code is OK - remains to check that this works correctly for XSLT

abel: we have checked the Java code and it is indeed fixed
... but what happens when there is more than one base tag?

dom: I think using the first <base> is ok
... ; (nothing that more than one <base> is not valid)
... I still need to look at the some of the XSLT cases where further URI resolution is needed

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5590 bug 5590 Record more detail about source of an image/object retrieval

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5774 bug 5574 Object Rule Processing

abel: we're not sure about 5590

-> http://lists.w3.org/Archives/Public/public-mobileok-checker/2008Jun/0005.html Dom's message re 5590

abel: from what we understand, we need to annotate the objects that are really rendered in moki
... not sure about having to annotate whether images come from style sheets

dom: [trying to summarize his message]

miguel: we think it would be better if we filter objects and images that are tasted or rendered in the moki

<objectInfo rendered="false" tasted="true" />

<object data="foo.png"><img src="foo.gif" /></object>

<object data="foo.png" type="image/png"><img src="foo.gif" /></object>

<Zakim> francois, you wanted to ask for clarification about "rendered eq false and tasted eq false"

loadtype="tasted" loadtype="rendered"

loadtype="tasted" vs loadtype="rendered"

http://www.w3.org/TR/mobileOK-basic10-tests/#PAGE_SIZE_LIMIT reads "In the following, include in the total only those objects retrieved under the 3.15.1 Object Element Processing Rule whose type attribute is not specified, and those whose content type is either "image/jpeg" or "image/gif" irrespective of whether the type attribute is specified. "

dom: I have implemented in XSLT - doable with XPath, more difficult through the DOM AFAICT

[detection of tasted in XSLT: http://dev.w3.org/cvsweb/2007/mobileok-ref/src/org/w3c/mwi/mobileok/basic/xslt/includedResources.xsl ]

[detection of rendered: http://dev.w3.org/cvsweb/2007/mobileok-ref/src/org/w3c/mwi/mobileok/basic/xslt/ObjectsOrScriptTest.xsl ]

dom: any idea about schedule on this bug?

abel: we're targeting mid-July, right?
... we'll start working on this beginning of next week - hopefully we can finish it in ~10 days
... if you guys can help with some of the other bugs, it would be great
... esp. with 5775 grammar validation

dom: sounds good; francois and I will try to tackle 5775

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5591 Handling of Meta Refresh

abel: think we should close it

dom: probably more part of UI than library
... I'll take it and move it to the Web UI component

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5762 Media rules check in linked CSS (StylesheetUseTest)

Abel: I think we need to review better the StylesheetUse test

Miguel: that test tries to check that at least one style sheet is used for handheld devices
... in the XSLT, we check the media attribute of the <link> tag
... but we don't check the content of the style sheet

dom: so, if the entire style sheet is enclosed in @media screen { }, the checker won't notice anything

-> http://www.w3.org/TR/mobileOK-basic10-tests/#STYLE_SHEETS_USE "If all styles are restricted to media types other than "handheld" or "all" by means of @media at-rules, warn "

scribe: indeed, the mobileOK spec says we should warn on this
... so yes, I agree we should be checking this as well

miguel: so we need to change the preprocessor and the XSLT
... we'll be dealing with it

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5763 xml-stylesheet processing instructions

abel: this one requires changes to the Java code too
... we'll take that one too - hopefully won't take too long

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5764 i18n

abel: we can do the Spanish translations, you guys could do the French one
... but obviously, this has a lower priority

dom: indeed; but let's leave it for after the release

-> http://www.w3.org/Bugs/Public/show_bug.cgi?id=5765 documentation

abel: same for that one

<francois> [I would add that having more semantics in the info returned by the checker should also have more priority than i18n IMO]

abel: hopefully there isn't that much work to do on that one

dom: let's meet next week

abel: ok
... what about the vodafone's contributor?
... I got a mail from Oscar last week
... I gave him information about our ongoing work, but I didn't get a response

dom: dan confirmed that he was allowed to work, and nominated him to the group
... ping him again by email, and copy francois and I, maybe?

<francois> [I will contact him as well to have him subscribe to the mailing-list]

abel: ok, will do and inform him of our call next week




[discussions on getting the test suite to work]

dom: I think there must be a bug in the preprocessor
... the length of the content detected shouldn't vary from a platform to another
... (typically, this could make a page fail on Windows and gets a mobileOK on linux - which doesn't make sense)

miguel: the problem is not in the preprocessor, but in tomcat

dom: ok, I hadn't understood this
... do you know whether we can configure tomcat not to behave differently?

miguel: it's related to line-feed vs carriage return characters
... we can check if removing carriage return characters help

dom: ok... would still be good to fix at some point, but I can guess we can live with it for the time being

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log)
$Date: 2008/06/26 08:57:33 $