From connolly at w3.org Mon Jan 8 11:08:57 2007 From: connolly at w3.org (Dan Connolly) Date: Mon Jan 8 11:09:05 2007 Subject: [uf-dev] DTSTART needs value=date for all-day events Message-ID: <1168283337.4882.57.camel@dirk> In the hCalendar tests, I see DTSTART:20060315 in, for example, component-vevent-rdate.ics . The default value type for DTSTART is DATE-TIME, but 20060315 is not a DATE-TIME; you can use DATE values in DTSTART, but you have to be explicit: DTSTART;VALUE=DATE:20060315 RFC2445 had some misleading examples, but they were fixed in 2006-10-04 edit http://tools.ietf.org/wg/calsify/draft-ietf-calsify-rfc2445bis/draft-ietf-calsify-rfc2445bis-03.html#DATE_TIME_START I don't have a patch for x2v nor the tests yet. But the following test data seems to be bad: 01-component-vevent-dtstart-date.ics:DTSTART:19970903 03-component-vevent-dtend-date.ics:DTSTART:19970903 05-calendar-simple.ics:DTSTART:20051005 08-component-vevent-multiple-classes.ics:DTSTART:20051005 09-component-vevent-summary-in-img-alt.ics:DTSTART:20060306 14-calendar-anniversary.ics:DTSTART:19971102 15-calendar-xml-lang.ics:DTSTART:20051005 16-calendar-force-outlook.ics:DTSTART:20051005 17-component-vevent-description-value-in-subelements.ics:DTSTART:20060118 calendar-del.ics:DTSTART:20051005 calendar-html-lang.ics:DTSTART:20051005 calendar-lang-sub-lang.ics:DTSTART:20051005 component-vevent-description-value-in-subelements-2.ics:DTSTART:20060118 component-vevent-rdate.ics:DTSTART:20060118 component-vevent-rdate.ics:DTSTART:20060215 component-vevent-rdate.ics:DTSTART:20060315 I'm working from: changeset: 334:6028b022c67f tag: tip user: brian suda http://suda.co.uk/ date: Tue Nov 28 23:38:40 2006 +0000 summary: added test for trailing slash in rel-tag http://hg.microformats.org/tests -- Dan Connolly, W3C http://www.w3.org/People/Connolly/ D3C2 887B 0F92 6005 C541 0875 0F91 96DE 6E52 C29E From msamuel at google.com Mon Jan 8 13:10:18 2007 From: msamuel at google.com (=?UTF-8?Q?=E2=98=BBMike_Samuel?=) Date: Mon Jan 8 13:10:30 2007 Subject: [uf-dev] hcalendar test suites Message-ID: <19b880020701081310p5ad10ddax52288fa75e336740@mail.gmail.com> I'm writing an hcalendar extractor in javascript. Does anyone have a testsuite of hcalendar examples? Would anyone be interested in the end-product once I've gotten it working? (Before anyone asks, I would try x2v but I need something that works in browsers that don't have full XSLT/js integration) cheers, mike From rbach at rbach.priv.at Mon Jan 8 14:52:23 2007 From: rbach at rbach.priv.at (Robert Bachmann) Date: Mon Jan 8 14:51:07 2007 Subject: [uf-dev] hcalendar test suites In-Reply-To: <19b880020701081310p5ad10ddax52288fa75e336740@mail.gmail.com> References: <19b880020701081310p5ad10ddax52288fa75e336740@mail.gmail.com> Message-ID: <45A2CB27.5040901@rbach.priv.at> Hello and welcome to this list! Mike Samuel schrieb: > I'm writing an hcalendar extractor in javascript. Does anyone have a > testsuite of hcalendar examples? Would anyone be interested in the > end-product once I've gotten it working? Have a look at the tests hg repository: Go to and click on "manifest" to see the directory listing. In the hcalendar directory you'll find the hcalendar test cases (.html and .icf files). HTH, -- Robert Bachmann (OpenPGP KeyID: 0x4A5CCF10) From brian.suda at gmail.com Tue Jan 9 08:32:45 2007 From: brian.suda at gmail.com (Brian Suda) Date: Tue Jan 9 08:32:52 2007 Subject: [uf-dev] DTSTART needs value=date for all-day events In-Reply-To: <1168283337.4882.57.camel@dirk> References: <1168283337.4882.57.camel@dirk> Message-ID: <21e770780701090832o66b55a50o9c653a2613686407@mail.gmail.com> On 1/8/07, Dan Connolly wrote: > RFC2445 had some misleading examples, but they were fixed > in 2006-10-04 edit > http://tools.ietf.org/wg/calsify/draft-ietf-calsify-rfc2445bis/draft-ietf-calsify-rfc2445bis-03.html#DATE_TIME_START --- hm, i'll look into this. This might impact alot of things? if X2V correctly determines that it is a DATE and not a DATE-TIME, then out little trick of "add one more day because iCal is not inclusive of dates" might changes. I'll make the update to the XSLT and then we'll have to test this in different calendaring applications. -brian -- brian suda http://suda.co.uk From brian.suda at gmail.com Tue Jan 9 09:00:08 2007 From: brian.suda at gmail.com (Brian Suda) Date: Tue Jan 9 09:00:17 2007 Subject: [uf-dev] DTSTART needs value=date for all-day events In-Reply-To: <1168283337.4882.57.camel@dirk> References: <1168283337.4882.57.camel@dirk> Message-ID: <21e770780701090900u1c24e6d5v8003649459a922bf@mail.gmail.com> On 1/8/07, Dan Connolly wrote: > I don't have a patch for x2v nor the tests yet. --- i have pushed a new version of the XSLT to HG that checks the string-length of the date to determine if it is a DATE-TIME or just a DATE and adds the parameter as needed. I have not done anything with the test cases yet. I did try to 'round-trip' the data with Apple's iCal. Even i convert a hCal page to .ics (with the VALUE param) and import it into iCal i get the correct dates. Then export the iCal calendar to .ics and i get the same dates and date-times back. Even with the DTEND being specified as just DATE the date still needs to be one day higher. -brian -- brian suda http://suda.co.uk From ryan at technorati.com Tue Jan 9 11:56:09 2007 From: ryan at technorati.com (Ryan King) Date: Tue Jan 9 11:56:09 2007 Subject: [uf-dev] DTSTART needs value=date for all-day events In-Reply-To: <21e770780701090900u1c24e6d5v8003649459a922bf@mail.gmail.com> References: <1168283337.4882.57.camel@dirk> <21e770780701090900u1c24e6d5v8003649459a922bf@mail.gmail.com> Message-ID: <3A016F4E-65BE-4799-854F-27D29CF21AA3@technorati.com> On Jan 9, 2007, at 9:00 AM, Brian Suda wrote: > On 1/8/07, Dan Connolly wrote: >> I don't have a patch for x2v nor the tests yet. > > --- i have pushed a new version of the XSLT to HG that checks the > string-length of the date to determine if it is a DATE-TIME or just a > DATE and adds the parameter as needed. > > I have not done anything with the test cases yet. > > I did try to 'round-trip' the data with Apple's iCal. Even i convert a > hCal page to .ics (with the VALUE param) and import it into iCal i get > the correct dates. Then export the iCal calendar to .ics and i get the > same dates and date-times back. > > Even with the DTEND being specified as just DATE the date still needs > to be one day higher. If we can get this to work well across our implementations (iCal, Exchange, Entourage, Sunbird), then let's update the tests to reflect this. -ryan -- Ryan King ryan@technorati.com From ryan at technorati.com Wed Jan 10 10:28:28 2007 From: ryan at technorati.com (Ryan King) Date: Wed Jan 10 10:28:37 2007 Subject: [uf-dev] parsing question Message-ID: <4C6CE80A-87C9-4B20-B78A-684C64C56D12@technorati.com> How should I handle the following in hCard: :D -ryan -- Ryan King ryan@technorati.com From brian.suda at gmail.com Thu Jan 11 01:13:23 2007 From: brian.suda at gmail.com (Brian Suda) Date: Thu Jan 11 01:13:34 2007 Subject: [uf-dev] parsing question In-Reply-To: <4C6CE80A-87C9-4B20-B78A-684C64C56D12@technorati.com> References: <4C6CE80A-87C9-4B20-B78A-684C64C56D12@technorati.com> Message-ID: <21e770780701110113l4665b9deo1bd4feacb9809d1c@mail.gmail.com> On 1/10/07, Ryan King wrote: > How should I handle the following in hCard: --- i think the valuw would be '//' The class="email" on the DIV is not an 'a' or 'abbr' (it is a div) so paring rules say use the node value. In this case the node value is a script node, which contains only ' //' As a parser, you should NOT look inside comments or > > > :D > > -ryan > -- > Ryan King > ryan@technorati.com > > > > _______________________________________________ > microformats-dev mailing list > microformats-dev@microformats.org > http://microformats.org/mailman/listinfo/microformats-dev > -- brian suda http://suda.co.uk From kmarks at technorati.com Thu Jan 11 01:28:08 2007 From: kmarks at technorati.com (Kevin Marks) Date: Thu Jan 11 01:28:15 2007 Subject: [uf-dev] parsing question In-Reply-To: <21e770780701110113l4665b9deo1bd4feacb9809d1c@mail.gmail.com> References: <4C6CE80A-87C9-4B20-B78A-684C64C56D12@technorati.com> <21e770780701110113l4665b9deo1bd4feacb9809d1c@mail.gmail.com> Message-ID: <50223b3d41a502f9d7625bc37a4e1a44@technorati.com> On Jan 11, 2007, at 1:13 AM, Brian Suda wrote: > On 1/10/07, Ryan King wrote: >> How should I handle the following in hCard: > > --- i think the valuw would be '//' > > The class="email" on the DIV is not an 'a' or 'abbr' (it is a div) so > paring rules say use the node value. > > In this case the node value is a script node, which contains only ' //' > > As a parser, you should NOT look inside comments or > That's my take? any other thoughts. > I think that also matches the author's intent - the point of that kind of gibberish is to avoid machine-harvesting of email addresses, so pulling it out in a microformat parser would undo that. From connolly at w3.org Thu Jan 11 07:23:13 2007 From: connolly at w3.org (Dan Connolly) Date: Thu Jan 11 07:23:17 2007 Subject: [uf-dev] parsing question In-Reply-To: <21e770780701110113l4665b9deo1bd4feacb9809d1c@mail.gmail.com> References: <4C6CE80A-87C9-4B20-B78A-684C64C56D12@technorati.com> <21e770780701110113l4665b9deo1bd4feacb9809d1c@mail.gmail.com> Message-ID: <1168528993.25509.49.camel@dirk> On Thu, 2007-01-11 at 09:13 +0000, Brian Suda wrote: > On 1/10/07, Ryan King wrote: > > How should I handle the following in hCard: > > --- i think the valuw would be '//' > > The class="email" on the DIV is not an 'a' or 'abbr' (it is a div) so > paring rules say use the node value. > > In this case the node value is a script node, which contains only ' //' > > As a parser, you should NOT look inside comments or References: <4C6CE80A-87C9-4B20-B78A-684C64C56D12@technorati.com> <21e770780701110113l4665b9deo1bd4feacb9809d1c@mail.gmail.com> <1168528993.25509.49.camel@dirk> Message-ID: <21e770780701110759t1d3872b0l1f1af03cdc73fa8c@mail.gmail.com> On 1/11/07, Dan Connolly wrote: > have you tried XSV to see what it currently does? I have just tried and you are correct, X2V does actually extract what is inside i tested with: the other problem with the original example is that the '<' are not escaped so as XML xsltproc throws plenty of errors, but changing the CDATA to just "a+b ;" does actually work. I then tried to put the CDATA inside a script element it still correctly extracts 'a+b ;' Finally, to replicate the original example, i commented out the CDATA in the script element: The extracted text is now "// a + b ;" with the '//' The '//' is a comment within the script element, which has no meaning in XML, it is just text so it is still extracted. Is this the expected behavior? if so, X2V works :) -brian -- brian suda http://suda.co.uk From ryan at technorati.com Thu Jan 11 10:21:08 2007 From: ryan at technorati.com (Ryan King) Date: Thu Jan 11 10:21:11 2007 Subject: [uf-dev] parsing question In-Reply-To: <50223b3d41a502f9d7625bc37a4e1a44@technorati.com> References: <4C6CE80A-87C9-4B20-B78A-684C64C56D12@technorati.com> <21e770780701110113l4665b9deo1bd4feacb9809d1c@mail.gmail.com> <50223b3d41a502f9d7625bc37a4e1a44@technorati.com> Message-ID: On Jan 11, 2007, at 1:28 AM, Kevin Marks wrote: > On Jan 11, 2007, at 1:13 AM, Brian Suda wrote: >> On 1/10/07, Ryan King wrote: >>> How should I handle the following in hCard: >> >> --- i think the valuw would be '//' >> >> The class="email" on the DIV is not an 'a' or 'abbr' (it is a div) so >> paring rules say use the node value. >> >> In this case the node value is a script node, which contains only >> ' //' >> >> As a parser, you should NOT look inside comments or > >> That's my take? any other thoughts. >> > I think that also matches the author's intent - the point of that > kind of gibberish is to avoid machine-harvesting of email > addresses, so pulling it out in a microformat parser would undo that. Thanks for all the input guys. I posted this partially because it frustrated and disgusted me and wanted you to share in my pain. :D I'm acutally thinking now that we should just ignore content inside