Re: [htdig-dev] Re: [htdig] Using cookies

[Adam B. <ad...@fr...>]

Hi All,

Thanks for the feedback. I've installed the 3.2 Beta4 snapshot that was=20
available. The cookie support is working fine using:
disable_cookies: false=20
I was able to log in to our WebGUI (highly recommended:=20
http://www.plainblack.com/webgui) development site by using:=20
start_url:http://www.mysite.org.au/?op=3Dlogin&username=3Dusername&identi=
fier=3Dpassword=20
(this approach also works for the Phprojekt component of the site). The=20
cookies must have worked because Htdig was subsequently able to index=20
protected pages.

As a first timer with Htdig my first impressions have been very positive =
with=20
snappy performance and 90% of the functionality we require out of the box=
=2E

Our site has multiple user and groups who can see diferent pages. They al=
so=20
need to be able to choose the type of information source: organisations,=20
publications and/or discussions. I am currently intending to implement th=
is=20
by creating numerous different htdig databases created by searching selec=
t=20
parts of the site with appropriate permissions. A wrapper script will com=
bine=20
several searches and order the results by category.

The site is for an NGO in Melbourne, Australia that provides women's=20
information services. Our approach is to make the search engine the prima=
ry=20
source of public and internal information. Results with Htdig are=20
encouraging.

regards,

Adam

On Saturday 07 September 2002 07:09, Gabriele Bartolini wrote:
> Ciao Gilles,
>
>     thanks for pointing to the cookies patch. Yes, it is a backporting
> patch from 3.2.x and works pretty well on my Linux system. I expect oth=
er
> people on different platforms to test it and give me some feedback
> regarding portability.
>
>     As far as patch posting is concerned, I don't know whether Joe rece=
ived
> my e-mail 2 days ago. If you need more stuff please tell me, ok?
>
> >How you'd use that for authorisation may be another matter, though.
> >I don't know if Gabriele's cookie support has a way of pre-loading coo=
kies
> >from another browser into htdig's cookie jar.  Care to comment, Gabrie=
le?
> >(Or anyone else for that matter?)
>
>     Well, to be sincere it is not a difficult thing. The code was desig=
ned
> to support any kind of Cookies storage, it is pretty flexible indeed. F=
or
> now only the memory jar is supported. This regards though the persisten=
cy
> of cookies among different crawls, which is different to the pre-loadin=
g of
> cookies. This should not be a big work to do as I just said.
>
>     We only need a format for storing them. We could decide this, if yo=
u
> want. My vote would be to store them in a text file using the HTTP synt=
ax.
> By doing this, we would avoid any kind of parsing and the insertion int=
o
> the memory Jar would be straightforward.
>
>      I don't know though if this is a flexible solution and the right o=
ne,
> because it seems to go against the usual way of configuration and leads=
 in
> some cases to confusion (a user should use the specific and exact synta=
x in
> order to issue one or more cookies). Any ideas?
>
> Ciao ciao and thanks for rising up the topic
> -Gabriele