On Wed, Apr 29, 2009 at 3:40 PM, Asiri Rathnayake
<
asiri.rathnayake(a)gmail.com> wrote:
On Wed, Apr 29, 2009 at 3:32 PM, Asiri Rathnayake <
asiri.rathnayake(a)gmail.com> wrote:
Hi,
On Wed, Apr 29, 2009 at 3:01 PM, Sergiu Dumitriu <sergiu(a)xwiki.com>wrote;wrote:
asiri (SVN) wrote:
> Author: asiri
> Date: 2009-04-29 11:14:28 +0200 (Wed, 29 Apr 2009)
> New Revision: 19186
>
> Modified:
>
platform/core/branches/xwiki-core-1.8/xwiki-officeimporter/src/main/java/org/xwiki/officeimporter/internal/openoffice/DefaultOpenOfficeServerManager.java
> Log:
> XWIKI-3721: Starting an already started openoffice server should be
prohibited
Shouldn't the UNKNOWN state also be considered, like:
if (ServerState.UNKNOWN != currentState && ServerState.RUNNING !=
currentState)
The server state may go into the UNKNOWN state if something goes wrong
while trying to start the openoffice server. The reason I added this UNKNOWN
state was because it is possible that the OOo server process indeed started
but jodconverter was unable to connect to it. I suggested mirko to kill the
OOo server process if this happens and he agreed:
http://groups.google.com/group/jodconverter/browse_thread/thread/3c33c6183f…
But still this issue has not been fixed as it seems. So yes, we should be
considered about UNKNOWN state too. I will fix this.
Ok, now I'm having second thoughts. I think we should get rid of the
UNKNOWN state. Because If the server goes into this unknown state there is
no comming back... the admin will have to restart the XE.
For an example, the admin might want to kill the straying OOo server
process and try to start it again from the UI. But this is not possible with
the UNKNOWN state and the above check. If it goes to UNKNOWN state there is
no comming back :(
Nope, no need to remove the UNKNOWN state because it is true, we don't know
the state. But not have the check you mentioned. That is not do:
if (ServerState.UNKNOWN != currentState && ServerState.RUNNING !=
currentState)
would be ok. So the condition is simply:
if (ServerState.RUNNING !=
currentState)
I think this is more correct.