[web2py] Problems in web2py 1.83.2

2010-12-29 Thread Maurice Ling
Hi

I have a problem since version 1.83.2. It gives me this error:

Traceback (most recent call last):
  File D:\web2py\gluon\restricted.py, line 186, in restricted
exec ccode in environment
  File D:/web2py/applications/init/controllers/default.py, line 91,
in module
  File D:\web2py\gluon\globals.py, line 96, in lambda
self._caller = lambda f: f()
  File D:/web2py/applications/init/controllers/default.py, line 55,
in index
name = check_login()
  File D:/web2py/applications/init/controllers/default.py, line 38,
in check_login
redirect(URL(r=request, f='../account/log_in'))
  File D:\web2py\gluon\html.py, line 226, in _URL
return URL(*args, **kwargs)
  File D:\web2py\gluon\html.py, line 186, in URL
raise SyntaxError, 'not enough information to build the url'
SyntaxError: not enough information to build the url

My codes worked up to version 1.82.1.

I am using my own authorization/login codes which is in controllers/
account.py. The offending code (check_login function) is as follows:

def check_login(session=session):
if session.login_time == None:
session.login_time = 0
if session.username == None or \
session.login_time + login_expiry * 3600  int(time.time()):
redirect(URL(r=request, f='../account/log_in'))
else:
return session.username

In version 1.82.1, check_login (called by http://localhost:8000/init)
redirects to http://localhost:8000/init/account/login. But from
version 1.83.2, it seems to redirect to 
http://localhost:8000/init/default/init/account/login
which is non-existent.

Any ideas what might be happening?

Thanks
Maurice Ling


Re: [web2py] Problems in web2py 1.83.2

2010-12-29 Thread Kenneth Lundström

If I remember correctly there was something about this on the list.

I think you use the URL the wrong way and until version 1.83.2 the 
helper worked wrongly. In your case it worked but it has been 
corrected and it doesn´t work anymore.


Try something like URL(c='account', f='log_in')


Kenneth



Hi

I have a problem since version 1.83.2. It gives me this error:

Traceback (most recent call last):
   File D:\web2py\gluon\restricted.py, line 186, in restricted
 exec ccode in environment
   File D:/web2py/applications/init/controllers/default.py, line 91,
inmodule
   File D:\web2py\gluon\globals.py, line 96, inlambda
 self._caller = lambda f: f()
   File D:/web2py/applications/init/controllers/default.py, line 55,
in index
 name = check_login()
   File D:/web2py/applications/init/controllers/default.py, line 38,
in check_login
 redirect(URL(r=request, f='../account/log_in'))
   File D:\web2py\gluon\html.py, line 226, in _URL
 return URL(*args, **kwargs)
   File D:\web2py\gluon\html.py, line 186, in URL
 raise SyntaxError, 'not enough information to build the url'
SyntaxError: not enough information to build the url

My codes worked up to version 1.82.1.

I am using my own authorization/login codes which is in controllers/
account.py. The offending code (check_login function) is as follows:

def check_login(session=session):
 if session.login_time == None:
 session.login_time = 0
 if session.username == None or \
 session.login_time + login_expiry * 3600  int(time.time()):
 redirect(URL(r=request, f='../account/log_in'))
 else:
 return session.username

In version 1.82.1, check_login (called by http://localhost:8000/init)
redirects to http://localhost:8000/init/account/login. But from
version 1.83.2, it seems to redirect to 
http://localhost:8000/init/default/init/account/login
which is non-existent.

Any ideas what might be happening?

Thanks
Maurice Ling




Re: [web2py] Problems in web2py 1.83.2

2010-12-29 Thread Branko Vukelić
2010/12/29 Kenneth Lundström kenneth.t.lundst...@gmail.com:
 If I remember correctly there was something about this on the list.

 I think you use the URL the wrong way and until version 1.83.2 the helper
 worked wrongly. In your case it worked but it has been corrected and it
 doesn´t work anymore.

Hm. Isn't that technically a case of breaking backwards compatibility?

-- 
Branko Vukelic

stu...@brankovukelic.com
http://www.brankovukelic.com/


Re: [web2py] Problems in web2py 1.83.2

2010-12-29 Thread Kenneth Lundström
I might be completly wrong but I think the way URL was used by Maurice 
was wrong, it should not have worked, but thanks to a bug in URL it 
worked the way Maurice wanted.


So I guess it depends on how you look at it. Did it break backwards 
compatibility.


If your are using a bug and somebody repairs that bug your application 
stops working.



Kenneth


2010/12/29 Kenneth Lundström kenneth.t.lundst...@gmail.com:

If I remember correctly there was something about this on the list.

I think you use the URL the wrong way and until version 1.83.2 the helper
worked wrongly. In your case it worked but it has been corrected and it
doesn´t work anymore.

Hm. Isn't that technically a case of breaking backwards compatibility?





Re: [web2py] Problems in web2py 1.83.2

2010-12-29 Thread Branko Vukelić
2010/12/29 Kenneth Lundström kenneth.t.lundst...@gmail.com:
 If your are using a bug and somebody repairs that bug your application stops
 working.

I agree completely. However, it's a tricky situation if you insist on
ever-backwards-compatible. You only insist on that because you don't
want to break people's apps. Yet a bug fix may sometimes do just that.

-- 
Branko Vukelic

stu...@brankovukelic.com
http://www.brankovukelic.com/


Re: [web2py] Problems in web2py 1.83.2

2010-12-29 Thread Kenneth Lundström

I guess not even web2py is perfect. Maybe some day.


Kenneth


2010/12/29 Kenneth Lundströmkenneth.t.lundst...@gmail.com:

If your are using a bug and somebody repairs that bug your application stops
working.

I agree completely. However, it's a tricky situation if you insist on
ever-backwards-compatible. You only insist on that because you don't
want to break people's apps. Yet a bug fix may sometimes do just that.





Re: [web2py] Problems in web2py 1.83.2

2010-12-29 Thread Branko Vukelić
2010/12/29 Kenneth Lundström kenneth.t.lundst...@gmail.com:
 I guess not even web2py is perfect. Maybe some day.

Sure. I don't mind the bugs getting fixed. I just wanted to say that
in a perfect world, you cannot expect backwards-compatibility-forever.
:)

-- 
Branko Vukelic

stu...@brankovukelic.com
http://www.brankovukelic.com/