- From: Christian Grün <christian.gruen@gmail.com>
- Date: Mon, 19 May 2014 13:34:39 +0200
- To: John Lumley <john@saxonica.com>
- Cc: EXPath ML <public-expath@w3.org>
Hi John, this may be late, but I wanted to point out that the output of the existing archive:entries() function - with the file paths as text node - could already be used as input for new archives. If I see it correctly, the archive:entry-names() function only got necessary because the paths were now moved to a value attribute. What was the main intention behind this change? Best, Christian On Mon, May 19, 2014 at 12:59 PM, John Lumley <john@saxonica.com> wrote: > An updated editor's draft of this specification (12 May 2014) is available > from: > > http://expath.org/spec/archive/editor > > (A diff version from the previous (3 December 2013) version is available but > not yet on the website). The major changes are: > > Shifting to a fully serialization-like structured entry description, e.g. > <arch:entry><arch:name value="picture.jpg"/>..... > Map-based functions are an optional interface, in the same namespace as the > element-based versions and with '-map' suffix to the name. > Map-based functions can use 'position' properties to report on and control > archive order. > All map-based functions use single maps, rather than map sequences, as > arguments. > A new function arch:entry-names() just yields the names of all the entries > in sequence (simpler than the structured one, and useful when using maps.) > Two convenience functions arch:text() and arch:xml() are defined, to > generate xs:base64Binary for archive content sequences. > > Issue(s) arising: > > I have not yet used structured entry descriptions in arch:create() and > arch:update() - to do so would require polymorphism (i.e. type detection) on > the arguments, as the arity would probably remain the same between > structured entry and string versions. To create with different compressions > on a per-entry basis (as might be needed for EPUB) would require something > like this. The alternative is to use different function names for the two > versions. [On the other hand you could be modern and use maps, where you can > do this as easily as falling off a log....] > > Further work: > > Some of the examples need further alteration; minor formatting changes > I have not yet started changing the qt3 tests on GitHub (or indeed our Saxon > implementation) until the shift to structured entry descriptions is pretty > clear. > > Reactions would be welcome. I hope that with the specification as now > altered, we can move to some reasonably quick convergence and actually get > this specification completed. > > > > John Lumley > -- > John Lumley MA PhD CEng FIEE > john@saxonica.com > on behalf of Saxonica Ltd
Received on Monday, 19 May 2014 11:35:32 UTC