Stations
========
If a station is downloadable, its content is his source
So, getSimplestDownloadUrl == getSimplestSourceUrl.
The *download* action is like the *source* action but only accept *GET* HTTP
command.
The *download* http output is like the *source* one.
DataHolders
===========
If a dataholder is downloadable, its content is his source, but we have specify
in the HTTP stream that the action is "download only".
So, the getSimplestDownloadUrl is the standard action URL, with action ==
*download*.
The outputHttpDownload output data 'as is' is the HTTP stream, with a special
HTTP header Content-Disposition that specify the name of the holded data file
name.
The outputHttpSource acts the same maner but do not touch to the HTTP headers.
So technicly, the outputHttpDownload is this "file previous version" way of
handling outputHttpSource and outputHttpSource is the old classic way of itself.
<element name="summary" namespace="http://www.0d.be/ns/build">
<mode name="atom">
<users commands="ACCESS GET"><everybody/></users>
<xsltFilePath mode="raw" mimeType="application/atom+xml">xsl/atom.xsl</xsltFilePath>
</mode>
</element>
and those documents (<build:summary>) will be callable with /atom that will
transform them with the given stylesheet. mode=raw means it is served as-is
while mode=anything else means the result document is passed to site xslt (that
part was not tested).
Alliance SP).
Now account -> user -> session -> account (before it was: account -> user ->
session -> user).
Accounts and users can now be located outside web site (still some work
needed for password accounts).
Some Liberty Alliance login improvements.
is still sent but with the real HTML page instead of the error message. So
the user sees the good page after he presses the cancel button.
Added a parameter "station" to outputData.
Directories now have a view mode, which displays its default view (used by
docs-manager).
Added element pythonPath. It allows to specify where to search for a
pythonClass.
Instead of using an heuristic to find them in the current directory or in
the modules or in the application or..., it now uses absolute URIs and
catalogs.
(see <element name="article" .../> in abracadabra/_config.xml). The <module>
definition in server config file will soon be removed for non Python modules.
The naming of new articles, now works again.
80 columns.
Added notion of command for administrators. The reason is: When a site
doesn't accept WebDAV at all, even the administrators should not be able to
do WebDAV commands.
Improved users search in config files, to be more simple to understand.
fpeters