the code is not in trunk, it's only on my machine: current web2py has no 
knowledge of saving sessions in redis.

On Monday, October 8, 2012 6:30:49 PM UTC+2, rif wrote:
>
> Thank you and take your time. I will report any issues I find with it.
>
> Right now after Massimo's fix I get this:
>
> Traceback (most recent call last):
>   File "/home/rif/Documents/webframeworks/web2py/gluon/main.py", line 541, 
> in wsgibase
>     session._try_store_in_db(request, response)
>   File "/home/rif/Documents/webframeworks/web2py/gluon/globals.py", line 
> 668, in _try_store_in_db
>     record_id = table.insert(**dd)
>   File "/home/rif/Documents/webframeworks/web2py/gluon/contrib/memdb.py", 
> line 256, in insert
>     id = self._create_id()
>   File "/home/rif/Documents/webframeworks/web2py/gluon/contrib/memdb.py", 
> line 291, in _create_id
>     id = self._tableobj.incr(shard_id)
> AttributeError: 'RedisClient' object has no attribute 'incr'
>
> -rif
>
> luni, 8 octombrie 2012, 19:21:08 UTC+3, Niphlod a scris:
>>
>> give me some time, I need to polish it with some improvements also on the 
>> redis cache side.
>> The current implementation is quite dumb (but redis is fast so no-one 
>> noticed problems). 
>> If you're up with testing it at least I can be sure it will work on 3 
>> machines (my test rig, my production env, your test rig).
>>
>> Il giorno lunedì 8 ottobre 2012 18:07:50 UTC+2, rif ha scritto:
>>>
>>> I find it better than memcache because the sessions would survive a 
>>> restart. So while I can settle on using memcache but I feel like I am 
>>> loosing something.
>>>
>>> I really see no reason to store them on the disk or in db (except 
>>> limited memory).
>>>
>>> -rif
>>>
>>> luni, 8 octombrie 2012, 18:44:09 UTC+3, Niphlod a scris:
>>>>
>>>> ATM, storing session in redis is a bunch of lines put together a month 
>>>> ago by me, but it needs to be tested carefully before releasing it. 
>>>> Moreover, it's possible that there are 3 peoples using redis with 
>>>> web2py (me, you, bruno), so I prefer spending time on developing other 
>>>> things.
>>>>
>>>> Il giorno lunedì 8 ottobre 2012 17:29:32 UTC+2, rif ha scritto:
>>>>>
>>>>> ## connect to Google BigTable (optional 'google:datastore://namespace')
>>>>>     db = DAL('google:datastore')
>>>>>     ## store sessions and tickets there
>>>>>     session.connect(request, response, db = db)
>>>>>     ## or store session in Memcache, Redis, etc.
>>>>>
>>>>> The last line above misled me. so it is possible to store them in 
>>>>> memcache but not in redis...
>>>>>
>>>>

-- 



Reply via email to