XSLT WG meeting 2016-04-14: Summary of technical discussions and decisions

The XSL WG met on Thursday, 14 April 2016

7. Spec bugs

  10 Open Bugs as of 13 April 2016 - 9:57 EDT

29558  [xslt30] Incorrect namespace for JSON
   marked straigtforward in minutes but still not closed.

§§ Marked resolved/fixed on 14 April 2016.

29550  [XSLT30] typo in Note in 3.11.1 Locating Stylesheet Modules
    editorial - no discussion necessary

§§ Marked resolved/fixed on 14 April 2016.

29513  [XSL30] Definition of "potentially consuming" in Glossary not complete
    editorial - no discussion necessary

§§ Remains open pending further testing of stylesheet changes.

29507  [xslt30] A problem case for streamed grouping
    Discussed April 7 with decision as follows:
    RESOLVED: to resolve bug 29507 by accepting the proposal in comment 3.

§§ Marked resolved/fixed on 14 April 2016.

Bugs Not Previously Discussed
--------------------------------------------

29502  [xslt 3.0] Streamability of xsl:fork

§§ The proposal was accepted. Since the meeting, the changes have been applied.

29499  [XSLT30] Global xsl:variable/xsl:param not in streamability rules       2016-02-22

§§ We started to understand what exactly the spec says in this area. Perhaps we need a third suggestion as to what the API should do: if the initial match selection is a streamed node, then leave the global context item as "absent".

§§ We will need to come back to this next week.

29492  [XSLT30] streamability of xsl:attribute-set may not be complete         2016-02-20

§ We reviewed this in detail.

§ The phrase "considered as a sequence constructor" appears redundant.

§ Usage transmission should be usage absorbtion in 19.8.6.

§ We don't know the context posture statically. Creates a problem? Abel suggests simplifying it by requiring streamable attribute sets to be motionless; it can then be assessed against "striding" for assessment purpose.

§ For xsl:element (etc) we need to say that if the attribute-set isn't declared streamable, then the use-attribute-sets operand is roaming/free-ranging.

We AGREE that the bug identifies problems,

ACTION 2016-04-14-001 on MHK (bug 29492) to review Abel's suggestions and revise as necessary.

29479  [XSLT30] Streaming and non-well-formed documents        2016-03-03

MK felt that we shouldn't attempt to make the behaviour of different processors identical in this case; we need to allow processor to abandon parsing when they've seen all the data they need, but it's very difficult to mandate that they must do so.

MSM concurred.

Abel concurred.

AGREED to make no change to the specification. (Close as WORKS-FOR-ME).

Bug is now closed.

29478  [XSLT30] Allow selecting all component types inside xsl:accept/xsl:expose       2016-02-23

§§ DECISION: we accepted the proposal in the bug report.

§§ (The changes have now been applied and a couple of test cases have been created).

29472  [XSLT30] Optionally allow xsl:stream to process a document *without* streaming  2016-03-10

§§ Abel presented the problem. MK presented his proposal in comment #6 and the rationale for it. Abel presented his response in comment #7. MK concurred with Abel. We discussed a bit about what the options strict / fallback / extended / optimistic were intended to mean, but the details will appear in MK's proposal.

§§ ACTION 2016-04-14-002 (bug 29472) on MHK to propose detailed wording to implement comment #7.

Received on Monday, 18 April 2016 14:25:50 UTC