httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Brian Behlendorf <br...@organic.com>
Subject Re: Good deed for the day?
Date Tue, 28 May 1996 02:18:22 GMT
On Sat, 18 May 1996 dirkx@elect6.jrc.it wrote:
> A few suggestions:
> 
>  - Would it be an idea to put, by default *ON*, an 'Authorative'
>    directive in each auth module; which ensures that if the module 
>    cannot find the user, or if the password mismatches, etc for that
>    module, an AUTH_REQUIRED is returned, unless the admin has
>     specifically swithced off this authorative keyword.
>  - Would it be an idea to make any problem with the access checking or
>    inconsistencies, missing informaion,e tc, such as not being able to open
>     afile in the gdb, etc  lead to a SERVER_ERROR
>   - I.e. be really hesitant to give a DECLINE.

This all sounds right.  Also sounds like the time might be right for a 
"guide to module writers", since writing a good module is more than just 
coding to the API.

>   - As a minor legacy problem, we have in the 'require user x y z' spaces
>    as the separators, same for groups, whih prohibit spaces in the
>   userIDs. 

Hmm, does it work if you say "require valid-user"?

> Furthermore two of those 'require user' lines give unexpected
>   results. How about allowing multiple lines ? or require user= multiple
> lines.

I'd support that.

>  - Furthermore, in the 'err' msql in the command_rc block, should that
>    be a human readble string for the API-self documenting page, or should that
>    really bedesigned for an error message; it seems that some modules
>    have an explanotory string, others have a typical error strong:
> 
> command_rec msql_auth_cmds[] = {
> { "Auth_MSQLhost", msql_set_string_slot,
>     (void*)XtOffsetOf(msql_auth_config_rec, auth_msql_host),
>     OR_AUTHCFG, TAKE1, "Auth_MSQLhost: The mSQL Host must be set to 
> value; Syntax:  Auth_MSQLhost < IP address | FQHN | localhost >" },
>     
> { "Auth_MSQLhost", msql_set_string_slot,
>     (void*)XtOffsetOf(msql_auth_config_rec, auth_msql_host),
>     OR_AUTHCFG, TAKE1, "The Host with the mSQL database (or localhost)" },
> 
> what is best ?

Hmm, given the action in mod_info I'd say the latter is preferred, but I 
think either can be derived from the other, so I'm not sure it matters a 
whole lot.

	Brian

--=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=--
brian@organic.com  |  We're hiring!  http://www.organic.com/Home/Info/Jobs/


Mime
View raw message