W3C home > Mailing lists > Public > xproc-dev@w3.org > April 2014

Re: Initial impressions [was: Re: try/catch with messages]

From: Graham Seaman <graham@theseamans.net>
Date: Thu, 24 Apr 2014 08:29:13 +0100
Message-ID: <5358BD49.5030806@theseamans.net>
To: xproc-dev@w3.org
The messages work fine. Thanks again.

I saw the extension library included a schematron wrapper with
parameters. Should this work before your calabash patch
(https://github.com/ndw/xmlcalabash1/pull/146) goes through, or does
this do something different? I have provisionally left some schematron
variables hardcoded in the .sch file, assuming I wouldn't yet be able to
pass them from xproc.

Graham

On 23/04/14 12:49, Imsieke, Gerrit, le-tex wrote:
> For messages, you may use cx:message [1]. Just import this:
> <p:import href="http://xmlcalabash.com/extension/steps/library-1.0.xpl" />
> 
> Usage example (taken from [2]):
> <cx:message>
>   <p:with-option name="message" select="concat('COVER: ',
> /epub-config/cover/@href)">
>     <p:pipe port="meta" step="create-ops"/>
>   </p:with-option>
> </cx:message>
> 
> You’ll find plenty of (more or less) working XProc examples linked on
> our transpect page [3]. One of the less complex is the CSS→CSSa
> converter [4]. Please note that you’ll have to check it out with an SVN
> client to get all externals.
> 
> Gerrit
> 
> [1] http://xmlcalabash.com/docs/reference/cx-message.html
> [2]
> https://subversion.le-tex.de/common/epubtools/modules/create-ops/xpl/create-ops.xpl
> 
> [3] http://www.le-tex.de/en/transpect.html#transpect-modules
> [4]
> https://subversion.le-tex.de/common/sandbox/css_expand_standalone/trunk/
> 
> On 23.04.2014 12:39, Graham Seaman wrote:
>> On 17/04/14 11:33, James Fuller wrote:
>>> Graham,
>>>
>>> nice job sticking with it ... xproc has a difficult learning curve
>>> which we (as in the W3C XML Processing WG) are trying to address in
>>> vnext.
>>>
>>> Be interested in your fresh impressions, to see what you found
>>> confusing, useful , etc.
>>>
>>> cheers, Jim Fuller
>>>
>>
>> I'd say the thing which has probably slowed me down most is an inability
>> to do simple debugging easily: a p:message as an equivalent to a 'print'
>> statement in other languages, or a more flexible p:log.
>>
>> I've now more or less managed to create the pipeline that I wanted.
>> However, since this pipeline is meant to carry out a series of
>> validation steps the difficulty in getting back diagnostic information
>> when something unexpected goes wrong does limit its usefulness.
>>
>> Looking forward to seeing the new version!
>>
>> Cheers
>>
>> Graham
>>
>>
>>
>>
> 
Received on Thursday, 24 April 2014 08:24:00 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:03:12 UTC