Issue 461 - Deadlock in dipatch() API call
Summary: Deadlock in dipatch() API call
Status: CLOSED FIXED
Alias: None
Product: General
Classification: Code
Component: code (show other issues)
Version: 619
Hardware: PC Linux, all
: P1 (highest) Trivial (vote)
Target Milestone: OOo 1.1
Assignee: andreas.schluens
QA Contact: issues@www
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2001-02-22 12:39 UTC by Unknown
Modified: 2004-10-11 15:03 UTC (History)
1 user (show)

See Also:
Issue Type: DEFECT
Latest Confirmation in: ---
Developer Difficulty: ---

Attachments
call stack (5.40 KB, text/plain)
2001-02-22 12:40 UTC, Unknown 		no flags Details
call stack (4.76 KB, text/plain)
2001-02-22 13:02 UTC, Unknown 		no flags Details
patch (1.02 KB, patch)
2001-02-22 16:07 UTC, Unknown 		no flags Details | Diff
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this issue.
Description Unknown 2001-02-22 12:39:11 UTC 
Dispatching a URL (file: od private:) from a external Java Application (over
remote UNO) causes a deadlock. See
http://www.openoffice.org/source/browse/whiteboard/OOBean/src/org/openoffice/OpenOffice.java
(Method  loadComponent(String url, com.sun.star.io.XInputStream xInput)) for the
example code.

It seems, that the framework holds its own mutex owned while calling VCL (bad
idea).The deadlock should be obvious from the attached callstacks of the locking
threads.
Comment 1 Unknown 2001-02-22 12:40:47 UTC 
Created attachment 77 [details]
call stack
Comment 2 Unknown 2001-02-22 13:02:22 UTC 
Created attachment 78 [details]
call stack
Comment 3 Unknown 2001-02-22 13:03:08 UTC 
sorry, first attachment is useless - please use second attachment.
Comment 4 Unknown 2001-02-22 16:07:22 UTC 
Created attachment 79 [details]
patch
Comment 5 Unknown 2001-02-22 16:09:49 UTC 
attached patch solves this problem. But I think the whole source should be
reviewed for those kind of situations. It is not clear to me which of these
hundreds of UNO calls are 'save' in the sense, that the mutex lock of the
calling object can be hold.
Comment 6 stx123 2001-02-22 18:14:48 UTC 
Andreas might be able to help...
Comment 7 andreas.schluens 2001-03-02 09:18:47 UTC 
There exist another (different) deadlocks in our framework project too.
We will work on it ... but it's not easy to fix all of them.
I think we must redesign our whole project to solve these problems realy ...
Look for further comments on this board - please.
Comment 8 andreas.schluens 2001-07-31 08:35:35 UTC 
fixed in internal version 638
Comment 9 thorsten.ziehm 2004-10-11 15:03:56 UTC 
This prio1-Task can be closed. It isn't interessting anymore!