I'm trying to get rid of the sputnik.cgi?p= in my URLs and I'm missing a
bit.  I thought I had seen detailed instructions on the sputnik wiki at
one point, but I cannot find them now.

Thanks,

Matt

On Mon, May 19, 2008 at 12:16 AM, Matthew M. Burke <mmburke@gwu.edu> wrote:
> I'm trying to get rid of the sputnik.cgi?p= in my URLs and I'm missing a
> bit.  I thought I had seen detailed instructions on the sputnik wiki at
> one point, but I cannot find them now.

You would want to change NICE_URL in the _config page, but you'll also
need to enable rewrite handlers for your web server.  From the Sputnik
perspective, that variable is used to craft all the URLs.

What web server are you using?

- Jim

> You would want to change NICE_URL in the _config page, but you'll also
> need to enable rewrite handlers for your web server.  From the Sputnik
> perspective, that variable is used to craft all the URLs.

I use the ?p= format by default to not assume too much about the web
server.  If you want nicer URLs, you should set NICE_URL to whatever
you want to come before the node name, and most URLs would be
constructed as NICE_URL..node_name.  It's up to you, however, to make
sure that node_name is then mapped back onto the "p" query parameter.
How you do this depends on the server.  I am using Apache for the
Sputnik wiki so I use mod_rewrite:

    RewriteRule ^en/(.*)$ /cgi/sputnik.cgi?p=$1

For the released version and the current version in git this would
affect only simple URLs that do not have any additional parameters.
Mostly because at the time I thought that URLs with & but without ?
would be confusing.  So, we would have

    /en/Sputnik

and

    /en/Sputnik.edit

but instead of

    /en/Sputnik&version=000009

we would have:

    /cgi/sputnik.cgi?p=Saci&version=000009

This decision has not been hugely popular, so I am now leaning on
using NICE_URL _always_ if it is defined.

Note that we could theoretically get rid of NICE_URL altogether and
just not add ?p= to BASE_URL automatically, telling the user that
BASE_URL should be set to something like
"http://localhost/sputnik.cgi?p=" if they are not using rewriting.  I
am afraid that this would be confusing at just the wrong time in the
installation process.  I want to make it as easy as possible to get
Sputnik running, leaving nice URLs as an extra improvement that you
could do later.  But I am open to suggestions on this.

- yuri

-- 
http://sputnik.freewisdom.org/

Jim, Yuri,

Thanks for the help.  As often is the case, I made a small mistake in 
configuring the rewrite rules.  Your examples helped me track the 
mistake down and correct it.  There are two remaining issues:

1) When you use the URL: scrawl.bluedino.net there is an error message 
about page index not being able to handle action .html

I've fixed that by using a DirectoryIndex directive and putting 
Home_Page as the first choice.  This feels slightly like a kludge.  Any 
better suggestions?

2) After editing and saving a page, the URL in the browser's address bar is

http://scrawl.bluedino.net/sputnik.cgi

Is there a way of having the sputnik.cgi not show?

Thanks,

Matt

> 1) When you use the URL: scrawl.bluedino.net there is an error message about
> page index not being able to handle action .html

If this is happening, it must be that your server is doing something
strange.  I am guessing that requests for "scrawl.bluedino.net" get
mapped onto "scrawl.bluedino.net/index.html" or something else that
ends in .html.  I wouldn't know why this happens.  Sputnik doesn't
like being asked for foo.bar when "bar" command is not mapped onto an
action.  You _could_, if you wanted to, tell Sputnik that "html"
command maps onto "wiki.show" action or, that it is the same as the
"show" command (which mapps onto "wiki.show") by putting "html=show"
in the "actions" field.

  - yuri

> 2) After editing and saving a page, the URL in the browser's address bar is
>
> http://scrawl.bluedino.net/sputnik.cgi
>
> Is there a way of having the sputnik.cgi not show?

This has been fixed in the code repository, I believe, but not in any
released versions.  The git version should work without problems when
using the standard filesystem storage method.  (Jim's mysql sqlite3
storage backends and S?rgio's SVN backend will have to be updated for
the latest changes that I checked in.)  It should definitely be at
least as stable as the released version in terms of not losing your
data.  (Which is to say, you can't easily lose data in either, since
there is no "delete" operation.)  And if you do find bugs, we'll be
able to fix them!

The version in source repository uses the same data format as the
version you are using, and if you decide for some reason to go back to
the last release, you'll have that option too.

- yuri

-- 
http://sputnik.freewisdom.org/

Yuri Takhteyev wrote:
>> 1) When you use the URL: scrawl.bluedino.net there is an error message about
>> page index not being able to handle action .html
>>     
>
> If this is happening, it must be that your server is doing something
> strange.  I am guessing that requests for "scrawl.bluedino.net" get
>   
I'm not sure I'd call it strange--it's the default behavior for 
Apache.   See http://httpd.apache.org/docs/2.2/mod/mod_dir.html.  But 
I'm happy enough with the fix I mentioned previously, so I won't worry 
about this.


> The version in source repository uses the same data format as the
> version you are using, and if you decide for some reason to go back to
> the last release, you'll have that option too.
>   

Ok, given this and your email about permissions, the real solution is I 
should just git the current code.  I tried that unsuccessfully once, but 
maybe after a good night's sleep :) I'll take another stab at it.

Thanks!

Matt