To recap, it worked fine when I used (url) instead of (redirect). However the generated URL incur two problems:
1.) It looks horrible. 2.) Subsequent AJAX calls are using a relative path and I don't want to hardcode absolute paths in the JS, hence they won't work and the application fails since it relies on them. I was forgot to save the IRC buffer today so I didn't catch the patch but I remembered the zeroing out of Http1, anyway I tried simply doing this is in the entrypoint: (de start () (off *JS) (zero *Http1) (if (getUsr) (desktop) (signin))) Note the (zero *Http1). That approach seems to be working, I can submit without any problems but I don't know if it will hold up, tomorrow will tell. One very strange thing though, the HTML output looks like this when the form first loads: *start* 408 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd";> <html xmlns=3D"http://www.w3.org/1999/xhtml"; xml:lang=3D"en" lang=3D"en"> .. </html> 0 *end* Note the initial 408 and the final 0. Check it out live here: http://vizreader.com/ at the moment I'm getting 438 instead of 408. /Henrik On Tue, Nov 24, 2009 at 8:55 PM, Henrik Sarvell <[email protected]> wrote: > This is odd, the problem seems to be happening when I switch machines, > I logged out and in several times at work today to try and trigger the > problem. When I came home I still had the cookie so everything worked > ok, I logged out, looks like this now: > > (de logout () > =A0 (setq *Cookies NIL) > =A0 (cookie 'uid NIL) > =A0 (signin)) > > No redirects anymore, and when I tried to log back in I got exactly > the same problem, despite removing the redirect and having (start) > look like this: > > (de start () (off *JS) (if (getUsr) (desktop) (signin))) > > So no JS either. If you're out of ideas I'll try to reproduce the > problem with a simple app that you can easily run. > > /Henrik > > > On Tue, Nov 24, 2009 at 10:57 AM, Henrik Sarvell <[email protected]> wro= te: >> Yes (redir) is just a shortcut to: >> >> (de redir (Func) >> =A0 (redirect (pack *Domain (or Func "@desktop")))) >> >>>>redirecting causes you to lose part of the session's context<< >> Good to know. >> >> I'm sorry, it seems like I missed to list (start) earlier: >> >> (de start () (if (getUsr) (desktop) (signin))) >> >> This is the entry point, it will try and get a user object with the >> help of the user id in the cookie, if it can't it will simply call >> signin which when submitted routes to the dead end (sometimes). >> >> In fact (usrQuit) and (logout) actually doesn't have anything to do >> with the problem at hand since they are not called at all. I'm sorry >> that I've been unnecessarily confusing. >> >> I suppose we're back to the beginning here if manipulating *SesId >> won't help me and since I'm not doing unnecessary redirects in this >> case? >> >> Obviously the form is rendered correctly and when I submit the >> following should be executed since I'm pretty sure I enter the right >> password: >> >> (prog >> =A0(cookie 'uid Uid) >> =A0(setq *Cookies (cons 'uid Uid)) >> =A0(redir)) >> >> And (redir) should render the desktop. However somehow something else >> cuts in between, but only sometimes when the server has been running >> for some time, never when I just restarted it. >> >> Could this thing that captures the button click before I have time to >> redirect be something in the new JavaScript stuff you've been doing >> lately? >> >> >> >> >> On Tue, Nov 24, 2009 at 10:01 AM, Alexander Burger <[email protected]>= wrote: >>> On Tue, Nov 24, 2009 at 09:27:16AM +0100, Henrik Sarvell wrote: >>>> I think I understand, how would I properly log someone out of the >>>> normal GUI, ie when using (app) et al, setting *SesId to NIL maybe? I >>>> just browsed http.l and noticed that variable. >>> >>> I would not manipulate '*SesId'. This would just give errors due to >>> illegal accesses to the session. >>> >>> I use "lib/adm.l", which has functions for login and logout. I'm not >>> sure in your case, but for terminating a session just (bye) might also >>> suffice. >>> >>> >>>> (de usrQuit () >>>> =A0 (let Usr (getUsr) >>>> =A0 =A0 =A0(or Usr (redir "@logout")))) >>> >>> What does 'redir' do? Something like 'redirect' in "lib/http.l"? Then i= t >>> would be better to call (logout) directly instead of redirecting to >>> "@logout", wouldn't it? Also, this would avoid an additional HTTP >>> transaction. >>> >>> >>>> So the following redirects to logout if the user id is not in a cookie= : >>>> >>>> (de logout () >>>> =A0 =A0(setq *Cookies NIL) >>>> =A0 =A0(cookie 'uid NIL) >>>> =A0 =A0(redir "@start")) >>> >>> Same here. I would directly call 'start' or any other functions needed. >>> >>> >>>> (de signin () >>>> =A0 =A0(app) >>>> =A0 =A0(action >>>> =A0 =A0 =A0 (html 0 "RSS Reader" *Css NIL >>>> =A0 =A0 =A0 =A0 =A0(form NIL >>>> =A0 =A0 =A0 =A0 =A0 =A0 (<table> NIL NIL NIL >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0(<row> NIL "Username" =A0(gui 'uname '(= +TextField) 10)) >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0(<row> NIL "Password" =A0(gui 'pwd =A0 = '(+PwField) =A0 10)) >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0(<row> NIL >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 (gui '(+Button) "Login" >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0'(let Uid (chkLogin (val> (= : home uname)) (val> >>>> (: home pwd))) >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0(if Uid >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 (prog >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0(cookie= 'uid Uid) >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0(setq *= Cookies (cons 'uid Uid)) >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0(redir)= ) >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 (err "Could no= t login."))))) >>>> =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0(<row> NIL (<href> "Sign Up" "@register= "))))))) >>>> >>>> It was after pressing the Login button here I got the problem, >>>> redirecting to, and rendering the login form was not a problem. >>> >>> Thre redirecting causes you to lose part of the session's context. >>> Perhaps it all works if you just do the processig directly, as I >>> suggested? The code will probably also become much simpler. >>> >>> You could try to just use the 'login' function from "lib/adm.l", as in >>> "app/gui.l". I think that it has all necessary mechanisms. >>> >>> Cheers, >>> - Alex >>> -- >>> UNSUBSCRIBE: mailto:[email protected]?subject=3dunsubscribe >>> >> > -- UNSUBSCRIBE: mailto:[email protected]?subject=unsubscribe
