Garth's test of PyDS 16.4.2003

2003-04-16

He's fast!

Trackback namespace in RSS, date thingy fixed (with %b rather than my suggested %m, not that I care). There were also some P changes, which I'm not sure about. Must look 'em up.

I love CVS.

posted at 20:30:24    #    comment []    trackback []
 

Tracking down rendering error

I'm glad I have a reflexive "select all and copy" behaviour (born of Radio instability); PyDS just tossed my changes when editing an earlier post. So, here's a refreshed version as a post of its own.
Background thread exception exceptions.ValueError: time data did not match format
D:\Python22\Lib\site-packages\PyDS\Tool.py[763] in thunk
D:\Python22\Lib\site-packages\PyDS\WeblogTool.py[165] in _initthread
D:\Python22\Lib\site-packages\PyDS\WeblogTool.py[864] in renderTimeframeOnline
D:\Python22\Lib\site-packages\PyDS\Tool.py[513] in renderPage [46] in respond
D:\Python22\Lib\site-packages\PyDS\Tool.py [1386] in getRSSNewestDate
D:\Python22\Lib\site-packages\PyDS\MacrosTool.py[509] in parseTodayFull
D:\Python22\Lib\site-packages\PyDS\strptime.py[448] in strptime

Further investigation reveals that PyDS spends a lot of time converting time tuples to strings and back. And, the reason I'm having trouble with some rendering is that the renderer is bombing out. Which might also explain why I have no RSS.

Hacking in some cheaty code

def parseTodayFull(self, date):
    parseFormat = '%a, %d %h %Y %H:%M:%S GMT'
    try: 
        return calendar.timegm(strptime(date, parseFormat))
    except ValueError: 
        raise ValueError, 'time data "%s" did not match format "%s"' % (
            date, parseFormat)

reveals:

Background thread exception exceptions.ValueError: time data "Wed, 16 2003 03:57:52 GMT" did not match format "%a, %d %h %Y %H:%M:%S GMT"

I'm bewildered at the design choice. Storing the timestamps as strings is very fragile. PyDS should store timestamps as time tuples or seconds since epoch.

If PyDS must, for whatever reason, store strings, I'd recommend a standard format. For the time being, though, this diff will put the month back and fix RSS as a side-effect.

posted at 13:57:52    #    comment []    trackback []
 

Where's the beef?

The beef is in \Documents and Settings\username\.PyDS.
posted at 13:35:28    #    comment []    trackback []
 

Getting ahead of PyDS' upstreaming

I've been editing and deleting posts, and editing and deleting stories, and for some reason I'm now well ahead of what's actually upstreamed to the cloud.

I've tried asking the Upstreaming tool to fix it, but nothing seems to be happening. There's no mention in Events of that request having been made, either.

Later: after restarting, asking for an upstream, compensating for an upstream cache, and taking the Draft setting off the story, it's getting there. I'm still not sure what stopped upstreaming, but I'm hoping that waiting another minute or two will get the story up there.

posted at 13:24:48    #    comment []    trackback []
 

RSS

Hmmm. Where's my RSS feed? My template points to it, but there's nothing there.

[Aah, it was caused by the exception...]

posted at 11:40:16    #    comment []    trackback []
 

FAST

One thing I'll say about PyDS: it's fast. Given that it's doing much the same as Radio is doing on my real blog, I'm newly bemused at Radio's CPU hunger.

posted at 11:31:44    #    comment []    trackback []
 

reStructuredText handling bug

After a few wrong turns, I've nailed a reStructuredText handling bug. Here's the diff.
posted at 11:11:28    #    comment []    trackback []
April 2003
MoTuWeThFrSaSu
  1 2 3 4 5 6
7 8 910111213
14151617181920
21222324252627
282930    
Mar
2003
 May
2004

Garth is testing PyDS rather more extensively than at first he thought he would.

XML-Image Letterimage

© 2003-2004, Garth T Kidd