If a browser gives you a pop-up box for the username & password, then the
curl equivalent would be:

-u username:password

You'd need that on *every* curl command line. The browser automatically
remembers and resends the authentication for you. Curl doesn't remember
anything between command line invocations, other than cookies, which are
stored in a files and only then if you specifically tell it to save and
reuse them.

Ralph Mitchell

On 11/15/05, mgerics <mis_at_albar.com> wrote:
>
> Thanks for the reply - sort of helpful, but...
> We don't use a proxy here, so I'm not sure that advice applies - do I need
> to provide our name server addresses to curl?
> I'm also fairly certain cookies aren't involved; data requests are a long
> string the script puts into the address bar;
>
> _https://messaging.covisint.com/invoke/HTTPConnector.Mailbox/get?action=msg&
> id=M2067114{ENTER}
> the resulting page(s) that is(are) loaded is(are) then saved as a text
> file
> for processing.
> The first data request requires a pause for a pop-up username password
> input
> box; subsequent requests don't require anything special.
> Hmm! Did finally get sonmething! See attached text file!
> The connection is https style...do I need something extra for this type of
> call?
>
> Thanks again
>
> -----Original Message-----
> From: Daniel Stenberg [mailto:daniel_at_haxx.se]
> Sent: Monday, November 14, 2005 5:24 PM
> To: the curl tool
> Subject: Re: Newbie requests help
>
>
> On Mon, 14 Nov 2005, mgerics wrote:
>
> (disclaimer: I'm so snowed-in on curl so I'm probably one of the worst
> guys
> to
> offer advice to newbies...)
>
> > Have an operation that gets data via HTTP.
> > 1st data access requires entry of username and password.
>
> It is very important you details exactly how this "requires entry of
> username
> and password". It can be done in many ways and the particular way your URL
> uses of course decide how it is done with curl.
>
> > Subsequent accessses simply require the URL.
>
> It sounds as if you'd need to use cookies received from step 1 to
> successfully
> manage step 2. Things like this are usually managed with cookies.
>
> > I am fairly certain curl will do this for me. dl'ed some archives and
> tried
> > curl <my url>, where <my url> is any of several urls that work fine with
> > ping. Apparently mising something, as I always get back 'curl: (6) Could
> not
> > resolve host: <my url>; Host not found'
>
> That makes it sound as if you should be using a proxy or something. curl
> doesn't detect or use that automatically, but you need to tell it to.
>
> > Could someone please repond to this email with exactly what files are
> needed
> > and where and how to install them - I am not familiar with 'make', etc,
> so
>
> > prefer a pre-compiled binary.
>
> You already have curl, that should be enough. The rest is about using curl
> correctly. Possibly from a scripting language or similar, but that is your
> pick.
>
> > Any helpful hints and suggestions that do not refer to lake jumping or
> > acroabatic kite control will be appreciated.
>
> I hope I managed at least partly! ;-)
>
> --
> Commercial curl and libcurl Technical Support: http://haxx.se/curl.html
>
>
>
>
Received on 2005-11-15