[pLog-svn] 1.2.6 release?
plog at hugi.to
Fri Jan 18 14:30:23 EST 2008
On 01/17/2008 05:44 PM, Jon Daley wrote:
> On Thu, 17 Jan 2008, Mark Wu wrote:
>> About the bug(or say issue) about return 404 or 200 status code ... I think
>> we should fix it before we release 1.2.6
I agree, too. I actually thought that was fixed a while ago. that's
really important, I think.
>> 1. If we use raw url (native url), we should return the customized page(
>> with our current error page "article fetch error") with http response status
>> 200. Becasue in native url, the script name is the reasouce we want to
>> fetch, the other parts just only parameters.
> Is that necessary? I would expect either search engines to not be
> able to search a parameterized site at all, or be able to distinguish
> different parameters as different sites, and so not think that a 404 on
> one page would mean a 404 on all others. I was under the impression that
> search engines didn't like parameterized sites, and so that is why we
> created the "search engine friendly" URLs before we had custom URLs. But,
> I don't really know the answer to that, and if someone else says that
> isn't how it works, then we should return a 200.
Interesting thought. I'm not sure it either. Do we have any SEO research
paper that did some empirical test? Because, I'd intuitively implemented
a 404 on urls with wrong parameters. But I can't link to any resource
that is proving nor disprove it.
>> 2. If we use custom/pretty/mod_rewrite url, we should return customized
>> page( with our current error page "article fetch error" ) with http reponse
>> status 404. Becasue the "whole url" is the resource we want to fetch. These
>> three kind of url is more similar to RESTful url ...
> That sounds good.
fully agree here.
>> ** There will be one problem if we use this method. In ie6, if it get 404
>> status, it will redirect to it's own 404 page in local desktop. But, I think
>> it is okay.
> That is a setting in IE6 (and IE7 I think), and that is the
> person's choice, and we shouldn't design around that.
true, that's an IE setting, active by default. But I don't think we
should care. search engines are much more important. Users should browse
links and not type them in the url field (or click them on search
results, and that's why the status code and SEO is more important).
And why I was looking at the build script was because of updates to the
plugins and templates. we tend to forget to upload them to sf.net if we
fix bugs and update the plugin version. We're still not up to date on
all plugins right now, but I'll work on it (including bugfixes to some
It's still be good to have version 1.2.6 realeased this month. I think
that would be reasonable.
More information about the pLog-svn