Log of /CityKid/citykid/commands.py

Parent Directory

No default branch
Bookmark a link to HEAD: (view) (download) (as text)

Add dump command, to export raw CityDesk data w/.json sidecars
and/or front matter

Py 2.7 compat; add postmortem debug for CityBlog
(which essentially is dead meat now anyway since
Google dropped the old blogger API for good.)

Uploads via rsync, assuming you have Cygwin in c:\cygwin
and the target specified in the ``rsync_target`` CityDesk
variable.

Read the blogger blog ID, user id, and password from CityDesk
variables instead of the command line, so at least you don't
have to put them in the registry (and so you can upload to a
different blog from different files with the same Tools menu
entry).

Don't try to convert unicode to a string before encoding it to UTF8.
(D'oh!)

Don't crash on empty folders.

Added 'text' attribute to articles, to get processed version with PTMFOG 
magic names removed.  Added timestamp preservation: files whose contents 
haven't changed will keep the same timestamp, to better support the use 
of tools like "make" and rsync.  Orphan deletion: files that do not 
have a corresponding CityDesk object will be deleted when the 
containing folder is published, thus cleaning up debris in the staging 
area when objects are renamed or deleted.  Simplified save() process for 
greater efficiency.  Added 'byTemplate' mapping to find articles by 
their template type.  Added 'sorted()' implementation for Python 2.3.

Simplified template loading, by using a version of Kid that incorporates 
my namespace initialization patch.  Only launch generated .html files 
or directories, not other file types like .gif, .css, etc.

Support absolute links in teasers.  XML and XHTML output.  Refactor to 
allow easier calling of Kid code that's in an article body.  Indexes
for audience, article name, keywords, and path.  CityKid can now 
regenerate the entire dirtSimple.org site in about 26 seconds, including 
startup time, which is slightly faster than CityDesk.  Regenerating 
individual pages of course takes much less time.

Output files.  Save articles in XHTML Transitional.  Walk the site 
hierarchy and generate files.  Launch the folder or article after 
completing generation.  Allow templates access to the running command
as 'cmd'.  The 'citykid' command is now essentially sufficient to 
substitute for CityDesk's built-in scripting language, although there 
are many convenience features that will probably be needed or desirable, 
such as XML output for articles with an .xml extension (e.g. RSS/Atom 
feeds).

Write page(s) to files in a staging area adjacent to the .cty file.  
Support teaser [[self-links]].  Add a template-accessible variable for 
the 'all' collection, containing all articles available to publish 
during this run.  Preload all articles and folders for maximum 
processing speed.

Implement kid templating.  Declare setup() dependencies correctly.  
Compute relative links.  More efficient state preloads.  Processing 
for CityDesk's PTMFOG##### "magic name" links.

About all that's left to implement is writing out articles to files in 
a staging area, and some usable querying methods for common operations.
(And maybe options to launch a web browser on the file(s), or rsync them 
to a server.  And maybe docs, in case anybody besides me ever wants to 
use this...)

Another attempt to fix the screwed-up import