Re: Optional options w/o defaults on built-in steps

ht@inf.ed.ac.uk (Henry S. Thompson) writes:

> We have 67 of these (!) -- 31 after merging duplicates:
>
> Here's my recommended disposition, but please review this, as some of
> these are not completely straightforward. . :

I changed the definition of in-scope variables in the XPath section to
refer explicitly to "specified options" and added a note about
unspecified ones.

I've made the changes detailed below, except for the ones where I've
raised questions.

I also updated the serialization options section, though I think it
still needs a bit of editorial cleanup.

> p:count
>  limit Add select='0'

Done.

> p:escape-markup
>  cdata-section-elements Add select='""' [same as XSLT2]

Done.

>  escape-uri-attributes  Add select="'false'"

Done.

>  include-content-type   Add select="'true'" plus remark that should
>                         be ignored unless method='xhtml' or method="'html'"

Done.

>  media-type Add prose saying "if not specified, per method:
>                                  xml: text/xml

application/xml? See below.

>                                  html or xhtml: text/html

xhtml = application/xhtml+xml?

>                                  text: text/plain
>                                  otherwise: implementation defined"

>  omit-xml-declaration "'true'"

Done.
>  standalone             Add select="'omit'", to fit with o-x-d='true'

Done.

>  undeclare-prefixes Leave as is (any default might cause errors wrt version)

> p:exec
>  args Add select='""'

Done

>  byte-order-mark  Add prose saying "if not specified, per encoding:
>                                       UTF-16: true
>                                       all others: false"

Done.

>  encoding Add prose saying "if not specified, implementation-defined,
>                             but in the case of the xml and xhtml
>                             methods it must be either UTF-8 or UTF-16."

Done.

>  normalization-form "'none'"

Done.

> p:load
>  dtd-validate Add select="'false'"

Done

> p:namespace-rename
>  elements-only Add select="'false'"

Done.

> p:store
>  href Add select="''" (implied by prose) and add prose lower down as follows:
>                              "The value of the href option may be
>                               relative, in which case it will be
>                               resolved against the base URI of its
>                               p:option element or parent before use."

I don't think this one is right.

  <p:store href="''"/>

would make the href="" attribute absolute wrt the base URI of p:store.

  <p:store/>

uses the base URI *of the input document* as the URI.

> p:unescape-markup
>  content-type Wrong as is -- change to select="'text/xml'", and in prose

Why? I understood the I18N folks to dislike text/xml because of encoding
issues.

>  charset Add select="'UTF-8'" and remove the prose to that effect,
>                               also correcting "using the specified
>                               encoding" to "using the encoding named
>                               by the value of the charset option".

Done.

> Also note that the prose for p:directory-list, p:load,
> p:make-absolute-uris, p:xslt and p:xsl-formatter needs to be changed
> by the addition of the words "or parent" or similar to phrases along
> the lines of "if relative, [it] will be resolved against the base URI
> of its p:option element", to allow for the shorthand form of the
> relevant option.

For p:directory-list, I think it's actually a mistake to specify a
default value. While it's fair to ask implementations to know the
difference between default and not-specified, I don't think it's fair
to know the difference between specified and defaulted. So saying that
"." has a special semantics wrt base URI resolution seems wrong. I
removed the default and made the prose deal with the unspecified
value.

                                        Be seeing you,
                                          norm

-- 
Norman Walsh <ndw@nwalsh.com> | 'I have done that,' says my memory. 'I
http://nwalsh.com/            | cannot have done that'--says my pride,
                              | and remains adamant. At last--memory
                              | yields.-- Nietzsche

Received on Thursday, 18 September 2008 14:54:52 UTC