29 Jun
2009
29 Jun
'09
10:33 p.m.
On Jun 29, 2009, at 10:28 PM, Marius Dumitru Florea wrote:
Vincent Massol wrote:
I like "syntax" even though we don't know what syntax we're talking
about. +1 from me for either outputSyntax (removes the ambiguity) or
simply syntax (keeps the ambiguity but maybe ok).
On Jun 29, 2009, at 4:05 PM, Thomas Mortagne
wrote:
Why not simply "syntax" instead of "outputType"? "type" is
too generic
IMO. "syntax" is more like "language" but for applications. I have a
resource (Space.Page) and I'd like to access a representation of this
resource in a specific syntax known by my client application:
Space/Page?syntax=pdf Hi devs,
We need to decide something for http://jira.xwiki.org/jira/browse/XWIKI-3413
I propose to take Vincent's proposal and have a parameter
"outputType"
in the URL which provide the syntax identifier of the renderer to
use.
Also since syntax identifier contains "/" which i never remember the
escaped version when i need it i propose to support short names by
looking at available renderers and take the last version of the
asked
syntax (which mean the only renderer of this syntax most of the time
anyway but I prefer having something in case of).
So for example you could ask for the page Space.Page printed as
plain
text and without UI with the URL:
http://host/xwiki/bin/view/Space/Page?outputType=plain&xpage=plain
or the full form
http://host/xwiki/bin/view/Space/Page?outputType=plain%2F1.0&xpage=plain
I don't like xpage=plain but it already exists and is doing exactly
this, printing the content of the document without UI. We can decide
latter if we want a new parameter for this.
Feel free to propose another name, I'm not sure of "outputType"
but i
could not find a better one and "renderer" seems too technical.
+1 for short name. We could also have an optional outputVersion
rather
than the full name with the "/". I'd prefer this it separates more
clearly the namespaces and there's no need for special parsing (it
would work with a syntax which has a "/" in its name for example). I agree with Vincent that a separate, optional,
parameter for the
syntax
version is better.
I'd then propose syntaxVersion (if we use syntax) or
outputSyntaxVersion (if we use outputSyntax) since version alone is
probably not specific enough and could be used for other things in
other contexts.
Thanks
-Vincent
+1 in general.
+1 in general too.
Thanks,
Marius
>
> Thanks
> -Vincent
>
>> On Mon, May 4, 2009 at 13:15, Vincent Massol<vincent(a)massol.net>
>> wrote:
>>> On May 4, 2009, at 12:46 PM, Jean-Vincent Drean wrote:
>>>
>>>> On Mon, May 4, 2009 at 11:29 AM, Thomas Mortagne
>>>> <thomas.mortagne(a)xwiki.com> wrote:
>>>>> Hi devs,
>>>>>
>>>>> In 2.0 architecture we have no way currently to print a strongly
>>>>> formatted text (like JSON) or simply a plain text without XHTML
>>>>> which
>>>>> we are doing generally using xpage=plain in XWiki 1.0
>>>> With the 1.0 syntax xpage=plain was mainly used to:
>>>> 1/ Output data in special format (JSON, xls, etc). In those cases
>>>> {pre} was used to avoid xhtml transformations.
>>>> 2/ Output xhtml content without the XWiki UI, I can think of one
>>>> use
>>>> case where this content was put in ajax tooltips.
>>>>
>>>> Is 2/ covered ? Do we need a xpage=xhtml for it ?
>>>> Note that xpage=xpart&vm=contentview.vm can be used as a
>>>> workaround.
>>> If you we want to have something clean for the future it seems to
>>> me
>>> that having a parameter called outputType (or simply output, or
>>> type,
>>> or contentType) which can take all renderer values would be best.
>>>
>>> For ex outputType=xhtml, xwiki, text, pdf, rtf, etc
>>>
>>> And when not specified it would default to outputType=xhtml.
>>>
>>> For removing the UI I'd use another parameter since it's
>>> orthogonal.
>>> Something like showUI=true|false
>>>
>>> Thanks
>>> -Vincent