2010-10-12 07:11:34 utc |
jmettraux |
tosch_le: Hello Torsten, how are you feeling today ? |
2010-10-12 07:12:03 utc |
tosch_le |
fine, thanks. how do you do? |
2010-10-12 07:14:23 utc |
jmettraux |
doing fine |
2010-10-12 08:00:42 utc |
tosch_le |
jmettraux: i've seen the "[ ] If-Match etag" entry in rk's TODO and am keen to help implementing this |
2010-10-12 08:01:02 utc |
jmettraux |
how do you understand it ? |
2010-10-12 08:01:26 utc |
jmettraux |
the spec is cryptic ;-) |
2010-10-12 08:02:22 utc |
tosch_le |
where's the spec? |
2010-10-12 08:02:51 utc |
tosch_le |
i've learned that etags are used to track changes in entities |
2010-10-12 08:05:18 utc |
jmettraux |
spec == todo ;-) |
2010-10-12 08:05:36 utc |
jmettraux |
how would you do it then ? |
2010-10-12 08:05:37 utc |
tosch_le |
argl |
2010-10-12 08:05:48 utc |
tosch_le |
that was what i was wondering about |
2010-10-12 08:06:05 utc |
tosch_le |
i asked myself if there would be much to gain in rk. |
2010-10-12 08:06:48 utc |
jmettraux |
in the case of an expression, we could use the expression's revision (_rev) as the etag |
2010-10-12 08:07:20 utc |
tosch_le |
for instance the workitems resource: the storage has to be queried in any case and will return whole workitems if i understand ruote's rdoc correctly |
2010-10-12 08:07:38 utc |
jmettraux |
yes |
2010-10-12 08:08:11 utc |
tosch_le |
so the only thing to save is bandwith |
2010-10-12 08:08:23 utc |
tosch_le |
s/with/width/ |
2010-10-12 08:08:35 utc |
jmettraux |
it's a bit more "RESTful" |
2010-10-12 08:09:09 utc |
jmettraux |
HTTP has specific response codes for "you're trying to update a resource whose representation you have of is stale" |
2010-10-12 08:09:17 utc |
tosch_le |
and could later be complemented by cheaper query methods in ruote itself? |
2010-10-12 08:09:46 utc |
tosch_le |
ok, just thought of caching stuff. |
2010-10-12 08:10:22 utc |
jmettraux |
If-None-Match != If-Match |
2010-10-12 08:10:41 utc |
jmettraux |
GET If-None-Match vs PUT/DELETE If-Match |
2010-10-12 08:10:54 utc |
tosch_le |
stale checking with puts, posts or deletes is a good idea, i suppose |
2010-10-12 08:11:14 utc |
tosch_le |
sinatra seems to support etags out of the box: http://sinatra.rubyforge.org/api/classes/Sinatra/ResponseHelpers.html#M000037 |
2010-10-12 08:11:28 utc |
jmettraux |
ruote already does it internally |
2010-10-12 08:11:43 utc |
jmettraux |
I want to have the right HTTP response codes |
2010-10-12 08:12:43 utc |
tosch_le |
+1 |
2010-10-12 08:13:15 utc |
tosch_le |
but it's a nice example for implementation ;-) |
2010-10-12 08:13:22 utc |
jmettraux |
indeed |
2010-10-12 08:22:56 utc |
jmettraux |
have to escape for now |
2010-10-12 08:22:58 utc |
tosch_le |
btw.: 412 should be the correct response error code if the etag doesn't match |
2010-10-12 08:23:04 utc |
jmettraux |
enjoy ! |
2010-10-12 08:23:04 utc |
tosch_le |
http://www.w3.org/Protocols/rfc2616/rfc2616-sec14.html#sec14.19 |
2010-10-12 08:23:07 utc |
tosch_le |
bye! |
2010-10-12 08:23:11 utc |
jmettraux |
cool, we need that ! |
2010-10-12 13:21:51 utc |
tosch_le |
jmettraux: as a start: http://github.com/tosch/ruote-kit/commit/148e57ba56c014989056f4ea3259ec0f7a541f4f |
2010-10-12 13:31:09 utc |
jmettraux |
tosch_le: hello, it's already merged in |
2010-10-12 13:31:14 utc |
jmettraux |
thanks ! |
2010-10-12 13:31:43 utc |
tosch_le |
:) |
2010-10-12 14:17:12 utc |
anb_ |
Hello |
2010-10-12 14:17:24 utc |
tosch_le |
hello anb_! |
2010-10-12 14:17:58 utc |
jmettraux |
hello |
2010-10-12 14:18:29 utc |
jmettraux |
anb_: maybe for your friend Didier : http://tokyo.craigslist.jp/sof/1993147850.html |
2010-10-12 14:19:08 utc |
anb_ |
I don't know if he knows ruby but i'll show him, thanks |
2010-10-12 14:19:45 utc |
jmettraux |
:-) |
2010-10-12 14:25:34 utc |
jmettraux |
have a good night ! |