Re: [WSG] Firefox DOM and whitespace (bug?)
On Aug 4, 2005, at 2:45 PM, Patrick Ryan wrote: User agents have always (and will always) ignore whitespace in their display of XHTML. Should the DOM ignore it too? I recognize that's kind of backward logic, but it's certainly the practical view. Useragents better *not* ignore whitespace in their parsing of XHTML, since any element can be styled with white-space:pre; They only ignore the whitespace if the CSS tells them to, and even then it needs to be stored in the DOM somehow because you may change what styles get applied at any moment. In short, the whitespace is and always will be important to keep track of, even if it is routinely discarded during rendering. It would be interesting to create a test case to see if IE parses the DOM differently based on whether the element in question preserves whitespace, and if that parsing changes when the white-space property changes -- Ben Curtis : webwright bivia : a personal web studio http://www.bivia.com v: (818) 507-6613 ** The discussion list for http://webstandardsgroup.org/ See http://webstandardsgroup.org/mail/guidelines.cfm for some hints on posting to the list getting help **
[WSG] Firefox DOM and whitespace (bug?)
I recently ran across an issue (I would call it a bug?) in firefox's DOM. I wrote a rather lengthy bit on it here: http://www.agavegroup.com/?p=32 But in short, firefox considers whitespace (tab, space, new line) to be nodes in the DOM. I've browsed the W3C spec, as well as the Mozilla DOM spec and I can't come up with anything that demonstrates how this should be handled. But it seems to me white space should be entirely ignored in the DOM. I just wanted to see if anyone else has run into this, and hear some thoughts. IE handles this differently (no surprise there...) and in this case, better. Is this a recent Firefox bug or proper behavior (that must be scripted around...). I'd be interested in any other thoughts/ideas. Thanks, Patrick ** The discussion list for http://webstandardsgroup.org/ See http://webstandardsgroup.org/mail/guidelines.cfm for some hints on posting to the list getting help **
Re: [WSG] Firefox DOM and whitespace (bug?)
On 8/4/05, Patrick Ryan [EMAIL PROTECTED] wrote: I recently ran across an issue (I would call it a bug?) in firefox's DOM. ... But it seems to me white space should be entirely ignored in the DOM. ... Is this a recent Firefox bug or proper behavior (that must be scripted around...). I'd be interested in any other thoughts/ideas. http://www.mozilla.org/docs/dom/technote/whitespace/ Regards, Rimantas -- http://rimantas.com/ ** The discussion list for http://webstandardsgroup.org/ See http://webstandardsgroup.org/mail/guidelines.cfm for some hints on posting to the list getting help **
Re: [WSG] Firefox DOM and whitespace (bug?)
On Aug 4, 2005, at 1:39 PM, Patrick Ryan wrote: I recently ran across an issue (I would call it a bug?) in firefox's DOM. ... But in short, firefox considers whitespace (tab, space, new line) to be nodes in the DOM. ... But it seems to me white space should be entirely ignored in the DOM. ... IE handles this differently (no surprise there...) and in this case, better. Is this a recent Firefox bug or proper behavior (that must be scripted around...). Firefox is right, I believe, because the DOM is defined like this: The DOM presents documents as a hierarchy of Node objects that also implement other, more specialized interfaces.[1] The whitespace is part of the document, therefore the DOM must present it within the hierarchy of Nodes. In this case, it is a Text node (defined on the same page). The proper way to parse a Nodelist is to not assume you know what is next, but to test what Nodetype the next child is, and then tailor your operation to fit (e.g., skip the whitespace and gimme the next node). Admittedly, the IE model would make some of my scripts easier to write, but then we'd lose the capability of the DOM to work with XML such that *everything* is a node, and HTML would be a special case that ignores whitespace. Enough special cases, and the standard ain't so standard. So I think we need to keep coding with tests for Nodetype. 1. http://www.w3.org/TR/DOM-Level-2-Core/core.html -- Ben Curtis : webwright bivia : a personal web studio http://www.bivia.com v: (818) 507-6613 ** The discussion list for http://webstandardsgroup.org/ See http://webstandardsgroup.org/mail/guidelines.cfm for some hints on posting to the list getting help **
Re: [WSG] Firefox DOM and whitespace (bug?)
Thank you for the excellent reponses. I can't believed I missed the whitespace document (http://www.mozilla.org/docs/dom/technote/whitespace/) I see the point, and from a one standard fits all perspective, it makes good sense. I do have to wonder though: The point of defining a DOM is to give you a structure to work with. And it's OK to create sort of sub DOMs. For example the XML DOM is basically a child of the SGML DOM. In other words while standard, the DOM was changed for XML (and therefore XHTML) to better suit XML. User agents have always (and will always) ignore whitespace in their display of XHTML. Should the DOM ignore it too? I recognize that's kind of backward logic, but it's certainly the practical view. Anyway, as they say - learn something new every day. Thanks for the reponses. -- Patrick Ryan http://www.agavegroup.com On 8/4/05, Ben Curtis [EMAIL PROTECTED] wrote: On Aug 4, 2005, at 1:39 PM, Patrick Ryan wrote: I recently ran across an issue (I would call it a bug?) in firefox's DOM. ... But in short, firefox considers whitespace (tab, space, new line) to be nodes in the DOM. ... But it seems to me white space should be entirely ignored in the DOM. ... IE handles this differently (no surprise there...) and in this case, better. Is this a recent Firefox bug or proper behavior (that must be scripted around...). Firefox is right, I believe, because the DOM is defined like this: The DOM presents documents as a hierarchy of Node objects that also implement other, more specialized interfaces.[1] The whitespace is part of the document, therefore the DOM must present it within the hierarchy of Nodes. In this case, it is a Text node (defined on the same page). The proper way to parse a Nodelist is to not assume you know what is next, but to test what Nodetype the next child is, and then tailor your operation to fit (e.g., skip the whitespace and gimme the next node). Admittedly, the IE model would make some of my scripts easier to write, but then we'd lose the capability of the DOM to work with XML such that *everything* is a node, and HTML would be a special case that ignores whitespace. Enough special cases, and the standard ain't so standard. So I think we need to keep coding with tests for Nodetype. 1. http://www.w3.org/TR/DOM-Level-2-Core/core.html -- Ben Curtis : webwright bivia : a personal web studio http://www.bivia.com v: (818) 507-6613 ** The discussion list for http://webstandardsgroup.org/ See http://webstandardsgroup.org/mail/guidelines.cfm for some hints on posting to the list getting help ** ** The discussion list for http://webstandardsgroup.org/ See http://webstandardsgroup.org/mail/guidelines.cfm for some hints on posting to the list getting help **
