- From: Jan Willem Knopper <jwk@stack.nl>
- Date: Tue, 19 Oct 2004 14:44:39 +0200
- To: "Jukka K. Korpela" <jkorpela@cs.tut.fi>
- Cc: Jan Willem Knopper <jwk@stack.nl>, www-validator@w3.org
Jukka K. Korpela wrote: > On Tue, 19 Oct 2004, Jan Willem Knopper wrote: Thank you for your fast reply, > > > In a form I used an input type="file" (inside a table). The form lacked > > the enctype="multipart/form-data". > > It is somewhat debatable whether that's against the specifications. > In practical terms, you should surely use that enctype for a form that > contains a file input field. > > But it surely doesn't violate any _formalized_ requirement in the HTML > specification. > As I read it back it says for the enctype attribute: The value "multipart/form-data" should be used in combination with the INPUT element, type="file". I read this the first time as if you use INPUT type=file, then..., but I see that this is debatable and I agree that there is no formalized requirement. > > I got no errormessage from the validator. > > That's unavoidable, since the mistake is not a reportable markup error. > Your markup does not violate the formal rules in the DTD, which is all > that the validator knows, basically. > > It even _could not_ be a reportable markup error. In SGML (and still less > in XML) you cannot express, in a DTD, a requirement that an attribute must > be present if a certain other attribute is present - still less that an > attribute with a specific value must be present depending on the value of > another attribute. If an attribute is made REQUIRED, then it is > unconditionally required. > So it seems it is not an error and cannot be reported. Maybe there should be a message in the 4.01 specification about using enctype at the file-select control type (17.2.1), but I guess thats another mailing list. Thanks for your quick response, Jan Willem Knopper
Received on Tuesday, 19 October 2004 12:45:04 UTC