You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@jackrabbit.apache.org by Angela Schreiber <an...@day.com> on 2005/10/03 15:36:19 UTC

litmus results for simple webdav server (rev. 293331)

hi

as suggested by stefano some time ago, i let litmus run
for the simple webdav servlet.

please find attached the commented test results, one is
using jcr locks the other one is using locks that are
not reflected in the repository.

regards
angela


ps: in order to allow litmus to create properties, its
necessary to reset the default nodetype in the fileimport
command.




Re: litmus results for simple webdav server (rev. 293331)

Posted by Angela Schreiber <an...@day.com>.
hi stefano

> [...] 
> I'm sure the authors of litmus would be very happy to hear from you 
> about that (unfortunately, there are not so many webdav servers out 
> there and even less that care about testing enough to pass it).

that's stated in the TODO that is shipped with litmus 0.10.3:
litmus should respect errors embedded in the multistatus response.

i would not dare to say, there is a bug in litmus :))

in other words: litmus is happy with a slightly modified code,
that returns an error instead of a multistatus with the error
embedded. and if this turns out to be an issue with any
client, we can easily change this.

> Also, what would be cool, IMO, is to automate the tests (say a cron job 
> somewhere, we could even ask for a jackrabbit zone here in apache) and 
> then send email here on the list if the diff between the previous run 
> and the current run is different than zero, so that we can follow the 
> progress on the resolution of the locking issues (which are critical in 
> using a JCR repository as a MacOSX dav disk).

there are two things left (as far as i remember)

- shared locks not implemented: i didn't consider this
   a critical issue so far.
- refresh of jcr locks fails. this is reported in JCR-226.
   i quickly discussed it with dominique last friday and got
   the impression, that this is not a huge thing to fix.

nevertheless i would be glad if i get informed as soon
as i break something :)

kind regards
angela




Re: litmus results for simple webdav server (rev. 293331)

Posted by Stefano Mazzocchi <st...@apache.org>.
Angela Schreiber wrote:
> hi
> 
> as suggested by stefano some time ago, i let litmus run
> for the simple webdav servlet.
> 
> please find attached the commented test results, one is
> using jcr locks the other one is using locks that are
> not reflected in the repository.
> 
> regards
> angela
> 
> 
> ps: in order to allow litmus to create properties, its
> necessary to reset the default nodetype in the fileimport
> command.

Angela,

outstanding work!

Known bugs aside, I think we are in very good shape, the warnings don't 
really concern me, but the 'bug in litmus' is very interesting and could 
create some concerns since litmus uses (AFAIK) lib_neon which is pretty 
much the de-facto webdav client library out there. So if this is a bug 
in neon, it might create a lot more incompatibilities that we would want.

I'm sure the authors of litmus would be very happy to hear from you 
about that (unfortunately, there are not so many webdav servers out 
there and even less that care about testing enough to pass it).

Also, what would be cool, IMO, is to automate the tests (say a cron job 
somewhere, we could even ask for a jackrabbit zone here in apache) and 
then send email here on the list if the diff between the previous run 
and the current run is different than zero, so that we can follow the 
progress on the resolution of the locking issues (which are critical in 
using a JCR repository as a MacOSX dav disk).

Other than that, let me repeat, outstanding job.

-- 
Stefano.