In my opinion we should go for it as the current behavior is problematic
enough. Regarding the testing I think it is probably hard to write meaningful
JUnit tests as the potential problems will most likely depend on the
environments: firewall, proxies,
So what I think we should do is:
-
OK I created http://jira.magnolia-cms.com/browse/MAGNOLIA-3390 and
referred to it from the Wiki.
I think the actual code change will be little, unless you'd like to
accompany it with a JUnit-test that spawns a little local HTTP server
in order to verify whether
Actually since the Greg's concerns about previous attempt to enable the same I
think we should get a test for the feature. You might be able to build test
upon existing integration tests.
On Nov 16, 2010, at 12:38 PM, Jörg von Frantzius wrote:
OK I created
What do you think of this improvement?
+1 from me
There is already page collecting features and changes that need to be made to
the activation to make it more reliable and maintainable. Could you please add
your suggestions to the page?
And there I thought you got Magnolia running on some super sized memory cluster
:D
On Nov 15, 2010, at 6:22 PM, Jörg von Frantzius wrote:
Whoops, that's 260MB, not GB, of course!
On 15.11.2010 17:57, Jörg von Frantzius wrote:
Hi,
we received a heap dump from our client, where there
Jan, didn't we look into chunking once in the past, and had problems with it ?
Or was it not in relation with activation ?
Otherwise, I guess the improvement's fine, as long as it can be turned off, in
case someone runs activation through a proxy or some odd http server ?
-g
On Nov 15, 2010,
afaik chunked encoding was a problem for cache not for activation.
http://jira.magnolia-cms.com/browse/MAGNOLIA-1996
On Nov 15, 2010, at 8:30 PM, Grégory Joseph wrote:
Jan, didn't we look into chunking once in the past, and had problems with it
? Or was it not in relation with activation