Re: pyblosxom 3000

On Thu, 13 Jan 2005, Bill Mill wrote:
>
> Well, now that pyblosxom 1.1 is out the door, we should be revving up to 
> change big stuff while we have a new branch. As far as I see it, and in 
> no particular order, here are things that are waiting to be done:
>
> 1) Caching - we should make infrastructure to assist with caching. My 
> latest patch (which I haven't looked at in about a month - may not still 
> be valid) is available at http://llimllib.f2o.org/files/cache.tar.gz , 
> and is described in a previous email to this list 
> (http://sourceforge.net/mailarchive/message.php?msg_id=10236127)

Just to be specific here, this is index caching.  We have entry caching 
already (though I'm not sure how useful it is) and the two are totally 
different.


> The big question for me is still this: how and when should the cache be 
> refreshed?  Anybody have any ideas?

I think we should allow the user to invalidate the cache whenever they 
want with a shell or python script.  I could see some users wanting to 
invalidate the cache every night via a cron job.  Or perhaps every time 
they add an entry.

Additionally, we should allow the user to set the amount of time their 
index cache should invalidate itself.  I could see some users wanting to 
invalidate the cache every hour or every 24 hours or never (and rely upon 
manual invalidation like the above).

Are there other use cases these two possibilities don't cover?


> Any other suggestions about this will be appreciated.  Ted, how have you 
> gotten along with the patch? are you still using it?

I haven't looked at the patch.  Either that, or I don't remember what was 
involved.

So, speaking from ignorance, I think that index caching should be a 
plugin.  The changes to allow for index caching will be in the core and 
probably be somewhat extensive and require changes to the interfaces we 
currently use for pulling a list of entries from the file system (i.e. 
tools.Walk).

While we're allowing for index caching, it'd be cool to also allow for 
storing entries in other places (database, flat file, ...).  I think this 
might require more effort, though.


> 4) Contrib stuff - Will's still asking for a contrib maintainer. I
> can't reliably concentrate on anything for more than 5 minutes, so I
> don't want to volunteer. Anyone?

This involves someone taking ownership of everything in /contrib and 
handling incoming patches, testing, documentation and releases of a 
contributed plugins pack.

I think it'd probably be a bit of a big job to get it started, but once 
the processes for handling contributed fixes, testing, and releases was in 
place it should be a couple of hours a week at most.  But I'm 
speculating--take my thoughts with a grain of salt.


> 5) Comment Administrator - I have one that works, more or less, except 
> it can't change the dates on comments. This is because I want to change 
> the comments module to use the date field inside the comments file 
> instead of the time embedded in the filename.

Anything is better than nothing.  I'm happy to take what someone else has 
and absorb it into what I need.


> 8) refactoring - There are some really ugly functions in pyblosxom.py 
> that could use refactoring. Especially blosxom_file_list_handler (which 
> sets some magic undocumented variables in the Request['data'] array to 
> work) and blosxom_process_path_info (which is just pretty awkward).

I'll take this task.  I've been making minor changes to these two for some 
time now.  Might as well continue to hone them.


> 9) documentation - This is getting better (excellent work will), but 
> always needs to be maintained. The new website is the hottness.

I'll continue to work on this as well.  If people have suggestions, I'm 
all ears.  Better yet, if people have actual content, that's the best kind 
of help.

So, from this, I'll have the following action items:

   1. I'll continue to work on, hone, and absorb documentation
   2. I'll continue to run the pyblosxom.sourceforge.net site
   3. I'll work on refactoring stuff in pyblosxom.py
   4. I'd like to advise on the storage/index-caching work,
      mod_python/wsgi work, static rendering work, andinternationalization
      work--but don't have the bandwidth or expertise to own those tasks
   5. Acting as maintainer and release manager for the pyblosxom core code

I'm still thinking about unit testing.  I may start re-writing bits of 
code to make it easier to unit test with doctest tools or something like 
that.  If other people have ideas, I'm all ears.

Getting there....

Also, I don't have an internet connection this weekend because RCN is 
flakey.  So I won't be getting much done and won't be in contact until 
Tuesday.

/will


-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt