Scott Ferguson wrote:
It's probably already caught by our regression tests, but I've added
a bug report at http://bugs.caucho.com/view.php?id=2374 just in case.
The current snapshots do have a large number of regression failures.
This is fixed in resin-pro-3.1.s080208, thanks.
By the way, we should probably add your suggested configuration
somewhere like the wiki. I'm not sure what title we should give it,
though.
It strikes me that its the reliance on the physical directory name in
the host-deploy construct which is causing the problem. It mixes the
found dir name on disk with the contents of its particuar host.xml.
Using my resin:import way everything is defined in the found xml
files, no matter where they exist in the subdir tree, because I put the
resin:import tag in the def for http://localhost webapps in
resin.conf, i.e. /. So what would be perfect would be the
host-deploy functionality but honouring everything in the located
file, as in this example (as a useful side effect is the files can have
any name too, which helps manage many hosts easier as I usually stick
the hostname in the xml filename):
conf/hostconf/pligg/www.foo.com.xml:
host xmlns=http://caucho.com/ns/resin;
xmlns:resin=http://caucho.com/ns/resin/core;
web-app id='www.foo.com' document-directory='/www/pligg/www.foo.com'
/web-app
/host
Lets me use http://localhost:8080/www.foo.com and
conf/hostconf/pligg/www.bar.com.xml:
host xmlns=http://caucho.com/ns/resin;
xmlns:resin=http://caucho.com/ns/resin/core;
web-app id='www.bar.com' document-directory='/www/pligg/www.bar.com'
/web-app
/host
Does the same for http://localhost:8080/www.bar.com etc
-- Scott
On Jan 26, 2008, at 10:31 AM, Stargazer wrote:
Note - this worked for previous snapshots, it just been broken now.
We find it useful to develop sites locally, stage them, then move them
to production when ready. For local
development we fire up resin and configure them such that their url is
conveniently available after
http://localhost:8080, for example http://localhost:8080/www.foo.com,
http://localhost:8080/www.bar.com etc.
Then when moved to staging they become
http://www.staginghost.com:8080/www.foo.com and finally production
is as
expected.
The advantage with this is it is always clear where you are from the
url: you can have all 3 open at once. The
configuration for this has always been easy too, using this syntax:
host id= root-directory=.
web-app id=/ root-directory=webapps/ROOT/
resin:import
fileset dir=${resinHome}/conf/hostconf
include name=**/*.xml/
/fileset
/resin:import
...
This means that all we needed was the relevant *.xml file for each
host
anywhere under our conf/hostconf dir.
When the host-deploy directive was introduced I was sure this
meant we
could do the same thing but without
the restarts, but this isn't exactly the case. Sure, I can create a
localhost folder and stick host.xml in it for my
foo.com example as above, and access it with
http://localhost:8080/www.foo.com, but the definition of foo.com is
within
the host.xml file, so I have no way to do the same for bar.com as I
have
using the resin:import style.
I could edit my local hosts file and add aliases but that means
there is
extra (unnecessary) config and you can't tell
from the url where you are if at any time www.foo.com could really be
pointing to 127.0.0.1.
Unless there is a way of doing this I'd really appreciate the
resin:import functionality not to be dropped please.
The pligg verbose logging I reported is still present in this
snapshot.
One other small point I'm sure you'd like to address before
3.1.15 : its
now 2008 but the snapshot still announces itself
as copyright 1998-2007 ;-)
___
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest
___
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest
___
resin-interest mailing list
resin-interest@caucho.com
http://maillist.caucho.com/mailman/listinfo/resin-interest
