-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
I think I found (and fixed) the problem. See
http://twiki.org/cgi-bin/view/Support/UnwantedUserLayoutAuthentication
- - evidently user-specified styles are still half-baked.
I modified the template file, so it should no longer attempt
authentication unless explicitly requested. Of course, this removes
the user-specified style ability.
- --Bruce
Ben Scott wrote:
| On 10/11/05, Bill McGonigle <[EMAIL PROTECTED]> wrote:
|
|>> Unfortunately, that doesn't tell us very much. The "viewauth"
|>> URL you are requesting is *supposed* to request authentication.
|>>
|>
|> When you get the default page, it includes:
|>
|> @import url("%USERLAYOUTURL%"); @import url("%USERSTYLEURL%");
|
|
| Well, now *that* looks broken. And I see what you mean. Request
| the default page and view the HTML source. Look at the top, where
| the style sheets are imported. You will see something like this
| (edited for brevity):
|
| ----- Cut ----- /* Default TWiki layout */ @import
| url("http://wiki.gnhlug.org/twiki2/pub/TWiki/PatternSkin/layout.css";);
| /* Default TWiki style */ @import
| url("http://wiki.gnhlug.org/twiki2/pub/TWiki/PatternSkin/style.css";);
| /* Custom overriding layout per web or per topic */ @import
| url("%USERLAYOUTURL%"); /* Custom overriding style per web or per
| topic */ @import url("%USERSTYLEURL%"); ----- Cut -----
|
| The "@import" business is normal CSS syntax. It's just like an
| #include directive in C, for example. It just tells the browser to
| pull another URL in as a style sheet. Style sheets are
| cumulative, so having multiple imports is fine.
|
| What I think is broken are those last two imports. Those look like
| TWiki variable names. Normally, an author can put a %VARIABLE% in
| some TWiki text, and TWiki will substitute the value of said
| variable when it generates the HTML for that TWiki text (i.e., when
| the page is requested by the reader). You should never see those
| in actual HTML, since TWiki should have done the substitution by
| then.
|
| %USERLAYOUTURL% certainly isn't a valid URL, any which way you
| slice it!
|
| I suspect that different UAs (browsers) are handling the bogus URLs
| differently. Perhaps Mozilla is just ignoring them, while
| whatever you are using ends up requesting it from the server, which
| in turn gets confused and ends up sending you to an authentication
| page.
|
| I'll look into it more when I can, hopefully tonight.
|
| -- Ben "Would you believe I'm not the webmaster?" Scott
| _______________________________________________ gnhlug-org mailing
| list [email protected]
| http://mail.gnhlug.org/mailman/listinfo/gnhlug-org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org
iD8DBQFDTCFM/TBScWXa5IgRAmmIAJwLgwzuaEOMSw7nHobH96EiiHL8MACeIjGj
b4YkRe/L9Tb6t7hmQ9lMOhM=
=KhUe
-----END PGP SIGNATURE-----
_______________________________________________
gnhlug-org mailing list
[email protected]
http://mail.gnhlug.org/mailman/listinfo/gnhlug-org
