Closed Bug 148488 Opened 21 years ago Closed 21 years ago
More HTML validation fixes
This also fixes some of the html in the perl files. I've left out stuff which is yet to be templatised or will be otherwise rewritten (reports.cgi, *edit*.cgi, queryhelp.cgi) and would have been a pain to do, for not much gain. gmuck isn't as accurate here, obviously - it picks up on <HTML> where HTML is a file handle, for example. Why are we using http-equiv redirects instead of 302 responses, btw?
Attachment #85889 - Attachment is obsolete: true
Yes, those scripts probably should be commented. The extended syntax was supported for html4.01, and ISTR it being required for html4.01 strict complience. Do browsers really have problems with it?
>The extended syntax was supported for html4.01, and ISTR it being >required for html4.01 strict complience. Do browsers really >have problems with it? The "extended syntax" (meaning nowrap="nowrap") is actually the basic markup, and attribute minimizations ("nowrap") are just SGML syntactic sugar. The different HTML dialects are just DTD's; the SGML declaration is the same, and thus the rules for SGML structural elements are identical between doctypes. XHTML is based on XML which doesn't have the concept of attribute minimization, and thus all X(HT)ML based markup must use fully written attributes. So in short, in HTML you can have it done either way regardless of the DTD used. In X(HT)ML you always have to write out the values. About the browser incompatibilities: I don't know what the current situation practically is. At least some time ago I remember having a problem with some older browser not implementing checked="checked" properly. But my key point is: I don't see a reason to partially move towards a new markup unless we have a clearly set goal of doing it all. Query forms etc. are filled with minimized attributes, so we won't reach XHTML compatibility with these changes anyway - and HTML 4.01 Strict can be done even without touching the nowrap/checked/readonly/selected/whatever thingies.
The HTML2.0 spec says: "Attributes such as ISMAP and COMPACT may be written using a minimized syntax (see 188.8.131.52 "Omitted Attribute Name" in [SGML]). The markup: <UL COMPACT="compact"> can be written using a minimized syntax: <UL COMPACT> (9) " Where footnote 9 is: "Some historical implementations only understand the minimized syntax." If HTML 2.0 refers to historical implementations, I'd be surprised if bugzilla works on such versions. Note that even in html2.0, teh minimisation stuff was only a shortcut. I'll check with hixie, though.
Hixie says we don't have to worry about browsers which don't support the long for for attributes - they're unlikely to be supporting forms anyway, and noone uses them.
Attachment #85891 - Attachment is obsolete: true
Comment on attachment 87520 [details] [diff] [review] v3 >- <option value="[% resolution FILTER html %]" [% selected IF resolution == "FIXED" %]> >+ <option value="[% resolution FILTER html %]" [% selected="selected" IF resolution == "FIXED" %]> Make that 'selected="selected"' IF resolution == "FIXED". I know, the previous code didn't work either... but why not fix this now, since it's such a triviality. r=jouni with the above.
Comment on attachment 87520 [details] [diff] [review] v3 r=gerv for 2.17 only. I can't be bothered to check the entire patch, but if you break something, we are bound to spot it before 2.18. Gerv
Checked in to trunk
Status: NEW → RESOLVED
Closed: 21 years ago
Resolution: --- → FIXED
QA Contact: matty_is_a_geek → default-qa
You need to log in before you can comment on or make changes to this bug.