[pLog-general] Notes, quesitions and comments

[Oscar Renalias]

Good comments, let's see...

> * Updating/Adding posts was broken at first, StringValidator.class.php 
> wasn't included. quick fix.

Sorry, unfortunately you'll run into many of those :) I'm now checking 
every after commit http://devel.plogworld.org/devel, which is also 
periodically updated to the latest version from the repository.

> * Admin interface is sexy and easy to use. I think thats the coolest 
> part.

Yeah, it's a great improvement compared with older versions :) However 
it's still incomplete (the "new xxx" and "edit xxx" screens are not 
ready yet) and there is still a lot of work to be done to adapt every 
single screen.

> * Still havn't uncovered how to use plugin objects in the templates. 
> $karma->Xxx is broken and returns error.

The karma plugin now works a bit differently than before, since it uses 
three custom fields to store the karma score of a post: one for 
positive votes, one for negative votes and a checkbox that tells 
whether the post is modded down or not. Try to checkout the plugins 
module and enable the karma plugin. In order to fetch the karma scores 
of a post, {$post->getFieldValue("positiveKarma")} and 
{$post->getFieldValue("negativeKarma")} will do. Also manually changing 
the karma scores has become now as easy as changing the values of the 
custom fields in the "edit post" page.

Give it a try, I hope it's not broken :)

> * Noticed a slight speed increase.

We have reduced the amount of queries needed to generate all the pages, 
so it's good to know that that's already noticeable. Have you also 
tried to enable smarty caching?

> * Monthly, aka archive, links is working. Much easier. I've removed 
> links to category views because of the huge amount of resources it 
> uses when trying to pull all 300+ articles from a catagory. Paging 
> plugin sounds sexy.

I don't understand this... Wasn't it working before? Was there any 
problem with it?

> * Searching is very handy, I've gotten some positive responses from 
> that.

Yeah, this probably was the most requested feature too :))))

> * Had a few random errors with pluginmanager. Haven't traced through 
> them but will post if i run accross them again.

They are really random and I haven't been able to nail down the reason 
yet... The thing is that now, plog does not scan the plugins/ folder 
every request but only when we visit the "plugin manager" section, so 
that we can save some time (my benchmarks really improved after 
changing this!) The only problem is that sometimes, when removing a 
plugin and revisiting the "plugin manager" page so that the list of 
plugins can be updated, it will generate an error... I still don't know 
why, any testing/feedback will be appreciated!

> * Havn't tried subdomains but its on the agenda.

I'm really interested in people testing this one!! Please let me know 
when/if available for testing this.

> * Register.php is missing, is that no longer in use?

It's been merged with summary.php (which has also been expanded and 
improved :))

> * What is the stub_post field when creating new posts? How will that 
> be used?

It is for generating nicer permalinks (by the way, have you already 
enabled the custom format links? You can enable/change them in the same 
place where you select "search engine friendly" or plain links)

The first implementation of nicer permalinks used a cleaned-up version 
of the post topic for generating links to posts (removed all accented 
vowels, punctuation, blank spaces replaced with '_', etc), and it 
worked fine for users of latin character sets.  The problem is that 
users of non-latin characters (such as chinese) were getting strange 
urls because of php's poor support out of the box for other characters 
sets than iso-8859-1 and because the chinese alphabet is really 
different from the traditional western alphabet so there wasn't any 
possibility to clean up the post topics by using the same 
transformations as before...

So now the "post slug" (I took the name directly from wordpress) is now 
used for generating permalinks. Chinese users are expected to type some 
short description of the contents of the post but in English so that 
permalinks can be correctly generated...

> * Feature request for function to only pull article text after the 
> extended text seperator. Didn't see it in there. That way we could 
> also use the intro/brief text area as a article summary/preview that 
> isn't shown when the full article is viewed.

Can you elaborate a bit more? Do you mean something like 
Article::getExtendedPostText()? You can always change your templates so 
that main.template shows the intro text while post.template shows the 
extended text only.

> * Resource manager seems much cleaner, good job.

There's still some work to do in there... Probably it will look even 
cleaner in a few weeks! :)

> * Havn't played with the bays. filtering, suggest a way to turn it off 
> indefinately and to not create database entries when comments are 
> made. Also function to truncate the table and re-create rules would be 
> nifty. Could be an admin cleanup feature.

The bayesian filtering hasn't changed at all. If it works well, don't 
touch it (as they say :)) But I will write down the idea of 
reinitializing the rules, sounds interesting.

> * Havn't had to block anyone, searching for aol users.

These features have now been moved to plugins ('hostblock' and 
'contentfilter') but once installed, they will register a few options 
in the admin menu and it will look like they are perfectly integrated.

> * Looking at postgres sql structure, any ideas on creating compatable 
> queries. Would be interesting to create a class that is called to 
> build queries and then the class, based on db type in use, create 
> queries on certain types. MySQL 3.x, 4.0.x, 4.1.x and Postgres all 
> have version specific functions. MySQL 4.1.x also has some very cool 
> speed enhancements as well as subqueries.
>   Example: select * from articles where (id, blog.name) = (select 
> max(id), blog from users group by id);

most of the sql code is pretty standard... except a few things that we 
should look at. The good thing is that we only need to look at 
class/dao, that's were all the sql queries for pulling data are. Please 
feel free to take a look, i've already been asked a couple of times 
about the possibility of making plog compatible with postgres but I do 
not have the time myself to do it... Ideas are welcome :)

> I think thats it for now. Have questions? Please ask! I'm liking it 
> alot so far and look forward to an official release.

I'm really interested in testing the subdomain features, to really see 
if they work. Also I've been thinking on carrying out some load and 
stress tests to see how the system will handle it... I think we can 
find some ready made scripts here: 
http://forum.mamboserver.com/showthread.php?t=11782

Thanks for the feedback, I really appreciate it :)

Oscar