[TxMt] Blogging Bundle / Serendipity / Empty Fields

[Philip Hofstetter]

Hi there,

after getting nagged at my blog (http://www.gnegg.ch/archives/354- 
External-blogging-tools.html) to look into why TextMate's Blogging  
Bundle behaves strangely with Serendipity blogs, I had a look into  
the problem.

Actually, there are two problems, but so far, I've only fixed one  
(it's getting late here):

1) When you have a Date:-Header in your Posting, s9y will interpret  
the date wrongly and will assume it to be the beginning of the epoch  
(1970-01-01). This is problably a flaw in s9y and I'll report it  
there as soon as I know what's going on.

2) The Pings- and Comments-Headers were ignored and Pings/Comments  
were turned off regardless of what the headers were set to. This is  
what I've actually looked into this evening (GMT+1 here):

Basically it all comes down to the fact that parse_post (blogging.rb)  
is unable to cope with empty headers of which s9y produces at least  
one: s9y always sends an empty mt_convert_breaks-member in the post  
record and when no tags are defined, it'll also send an empty  
mt_keywords member.

With an unpatched blogging.rb, this will create something like this  
(ignore the f*ed up date - that's caused by bug 1) above):

Type: Blog Post (Markdown)
Blog: s9y
Post: 1
Title: First entry
Date: 1970-01-01 02:00:00 +0100
Keywords:
Format:
Pings: On
Comments: On

This causes the parser to stop after the Keywords-header (lines  
206/207) which will lead to the following headers to be ignored, thus  
turning off comments (s9y assumes Comments=off if the information  
isn't present in the RPC-call).

Now I'm not sure if s9y's behaviour is actually legal here (please  
enlighten me), but I think the blogging bundle should be able to cope  
with empty fields.

There are two possible fixes: One is to fix the fields-parser to make  
it stop bailing out at empty fields. I've attached bb_fixparser.diff  
which does that. The other way is to make post_to_document stop  
adding empty format and keyword fields. bb_skipempty.diff does that.

I'm not sure which way is the correct way to go. fixparser has the  
tendency of working with even more crude blogging engines, while  
skipempty is what the bundle did so far (look at the handling of  
mt_tags for example.

Anyways. Please consider merging one of the patches to make the  
blogging bundle work better with serendipity.


Thanks

Philip


-------------- next part --------------
A non-text attachment was scrubbed...
Name: bb_fixparser.diff
Type: application/octet-stream
Size: 1231 bytes
Desc: not available
URL: <http://lists.macromates.com/textmate/attachments/20070323/7d5f1c1c/attachment.diff>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: bb_skipempty.diff
Type: application/octet-stream
Size: 1098 bytes
Desc: not available
URL: <http://lists.macromates.com/textmate/attachments/20070323/7d5f1c1c/attachment-0001.diff>
-------------- next part --------------