Garth's test of PyDS

Another Big Set of Changes to PyDS

Thu, 05 Aug 2004 21:45:36 GMT

Georg has committed another big patch of mine. If you run into any problems, yell at me. :)

IronPython is out!

Thu, 29 Jul 2004 09:27:28 GMT

Continuations

Fri, 16 Jul 2004 14:21:52 GMT

Ian Bicking's wee rant about continuations prompted a comment by Nathan Gray pointing to a paper (PostScript) he'd written about them. Y'know, I almost think I get it now.

About *$$%& Time!

Fri, 16 Jul 2004 13:49:52 GMT

Okay, got this PyDS instance back up and running, with all my latest patches in CVS thanks to Georg.

Diffed!

Tue, 01 Jun 2004 10:46:24 GMT

With respect to P120, it looks like Georg has incorporated my changes into CVS. If you do an update, you won't need no steenking diff.

New To Dos:

Mon, 31 May 2004 14:06:56 GMT

P124: Where'd my ambiguous links resolve to?
P123: What's this item's internal link?
P122: Show only category items
P121: Show first, not edit
P125: This text points to nothing at all

Diff!

Mon, 31 May 2004 08:59:44 GMT

Georg! Grab gtk-2004-05-31.diff for the changes:

MeshTool:

P110: use edit/view links where possible

P119: redirect explicit pyds:/ links
handle cached link target no longer existing

ShutdownTool:

Added startShutdown API call

StructuredText:

Add Psyco testing (if installed)

Fix testing (previously broken with _flet)

WeblogTool:

Fix renderItemByUniqueId

WikiTool:

P117: fill in empty Wiki titles

P118: prevent backslashes in Wiki node names

Remove Delete button when editing

DirectPydsUriTest

Sun, 30 May 2004 11:24:48 GMT

As DirectPydsUriTest points out, I should capture outbound pyds:/ links and redirect them appropriately. [Fixed.]

Backslashes in Wiki node titles

Sun, 30 May 2004 04:33:04 GMT

When I first made EmptyParentNodeTest, I accidentally used backslashes to separate the ID. That caused problems because the code I was using to separate the names worked (_fullSplit), but the subsequent name lookups I was performing after recomposing with '/' didn't. I need to automatically translate them, or at least issue an appropriate warning. [Fixed.]

Empty Title Test

Sun, 30 May 2004 03:26:56 GMT

I should automatically fill in the titles of new Wiki nodes. I've added EmptyTitleTest to track it. [Done.]

New bug!

Sun, 30 May 2004 03:22:40 GMT

I was trying to post a new Wiki node (EmptyParentNodeTest), and...

Exception exceptions.NameError: 
global name 'getPost' is not defined

PyDS\Tool.py 967 process_request 
u = getattr(tool,meth)(req) 

PyDS\WikiTool.py 569 edit_redir 
id = self.addNode(id, title, text, template=int(template)) 

PyDS\WikiTool.py 699 addNode 
meshTool.renderItemsWithUnresolvedLinks() 

PyDS\MeshTool.py 986 renderItemsWithUnresolvedLinks 
return [handles, self.renderItems(handles)] 

PyDS\MeshTool.py 1015 renderItems 
method(item.uniqueId) 

PyDS\WeblogTool.py 325 renderItemByUniqueId 
post = getPost(uniqueId)

Fix:

# Render an item by its unique ID
def renderItemByUniqueId(self, uniqueId):
        """Render all pages associated with the post identified by `uniqueId`."""
        post  = self.getPost(uniqueId)
        # ... and so on

I've got no idea why renderItemsWithUnresolvedLinks was triggering weblog renders. I should also look into that.

Psyco vs PyDS

Fri, 28 May 2004 14:00:32 GMT

Anyone tried using Psyco with PyDS? I'm thinking it might be especially handy for rendering reStructuredText...

[later] I've tried it out, and it speeds up reST on the example in StructuredText.py by a good 40%. Just for fun, I've changed pyds-start to load Psyco. I'm posting from it now. :)

Still trying to nut out this local rendering problem in WeblogTool...

Fri, 28 May 2004 13:13:36 GMT

I thought I was on to something, but WeblogTemplate.tmpl does indeed refer to the local rendering.

AHA! Fixed. New code at the bottom of ReferToWeblogPost.

That'll do me for the night, I reckon. :)

Aha!

Mon, 24 May 2004 13:13:36 GMT

My problem with Wiki uploads (see P111) turned out just to be a problem with the templates. Georg, you need to catch that Cheetah.NameMapper.NotFound exception and throw up something nicer...

Now, why can't I force a refresh? Or, assuming I can: why is my call to PyDS.Tool.debugOutput not working? [Cough. look in error.log, silly!]

I don't get the render problem

Mon, 24 May 2004 11:01:20 GMT

Following on from P111; if I hit the Status page I get a report on the size of the StoryTool render queue. Maybe it's there now?