- From: Geert Bormans <geert@gbormans.telenet.be>
- Date: Tue, 19 Sep 2017 09:53:17 +0200 (CEST)
- To: Norman Walsh <ndw@nwalsh.com>
- Cc: XProc Dev <xproc-dev@w3.org>
Little has been discussed, other than a very basic general idea So, very sketchy this is.... It seems there are a two common scenarios used now - develop an XProc pipeline, then dynamically change that pipeline adding steps that do storage of output port results, do reporting, logging, messaging or validation of intermediate results - develop a bloated XProc pipeline with all of that storage, reporting... in it (and switching it off when not in debug "mode") Static or dynamic "bloating" of the pipeline so to speak, mostly serves a purpose of debugging or inspection Rather than trying to add all sorts of mechanisms such as debug levels to initiate port logging or storage, passing the database url or base uri for result storage... leaving us with a more complex core spec, yet not completely satisfying, for debug purposes.... The idea is to have an injection spec We agreed we need it, we committed to work on it soon, but we have not specified a single bit of it, so all of the following is pure speculation - address output ports (or input ports, steps maybe... nothing specified explicitly so far) by means of + XPath into the static version of the XProc + (likely preferred) using an @xml:id on a port that is unique across the pipeline - have a series of declared steps in the pipeline that have "instructions" (conditions and parameters such as uri and debug levels + a series of potential port addresses) in order to be injected in the "core" pipeline when required Not sure any of this makes sense (have not had enough coffee yet) So I encourage other attendees to take sufficient coffee first and then comment Best regards Geert ----- Oorspronkelijk bericht ----- Van: "Norman Walsh" <ndw@nwalsh.com> Aan: "XProc Dev" <xproc-dev@w3.org> Verzonden: Maandag 18 september 2017 17:01:13 Onderwerp: Re: Remove p:log? Geert Bormans <geert@gbormans.telenet.be> writes: > Thanks Vojtech for the correction, > we have discussed an "injection spec" not necessarily a framework Can someone sketch up a few paragraphs of what that might look like? I’m having a hard time getting my head around the idea. Be seeing you, norm -- Norman Walsh Lead Engineer MarkLogic Corporation Phone: +1 512 761 6676 www.marklogic.com
Received on Tuesday, 19 September 2017 07:53:43 UTC