[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
Re: [xml-dev] Error and Fatal Error
- From: John Cowan <cowan@mercury.ccil.org>
- To: Joe Fawcett <joefawcett@hotmail.com>
- Date: Sun, 17 Jul 2011 14:15:52 -0400
Joe Fawcett scripsit:
> So what's the interpretation of the spec where it has these two
> categories, one of which is recoverable? What tools does that apply to?
To start with, there are relatively few cases of errors which are not
fatal errors (other than validity errors, which are generally treated
as non-fatal by validating parsers, and of course are not even visible
to non-validating parsers). A non-fatal error is marked by the use of
the phrase "an error", or by one of the modal verbs MUST or MUST NOT or
their equivalents.
The relevant occurrences of "an error" are these, AFAICT:
1) An xml:space attribute contains an unsupported value.
2) The DTD is non-deterministic.
3) An attribute value contains a reference to an undefined entity.
4) A system identifier contains "#".
5) A reference to an unparsed entity appears in an entity declaration.
Most appearances of MUST are within validity constraints (see above),
within well-formedness constraints (in which case they are fatal errors,
since violation of a WFC is fatal), or are constraints on what the XML
processor does. The remaining MUST (NOT)s in documents are AFAICT:
6) Names MUST begin with NameStartChar and contain NameChars.
7) "--" MUST NOT occur in comments.
8) A DOCTYPE declaration MUST appear, if at all, before the root element.
9) PE references MUST NOT appear within declarations in the internal subset.
10) The predefined entities, if declared, MUST be declared correctly.
--
The Imperials are decadent, 300 pound John Cowan <cowan@ccil.org>
free-range chickens (except they have http://www.ccil.org/~cowan
teeth, arms instead of wings, and
dinosaurlike tails). --Elyse Grasso
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]