LH_Weather makes LCDHost crash on Windows startup

[GoogleCodeExporter]

What steps will reproduce the problem?
1. Having LCDHost on autorun, it may need an SSD if the problem is caused by 
"LCDHost starting too fast"
2. Having LH_Weather in the layout (currently using Eos)
3. Maybe, having the connection down while LH_Weather starts

What is the expected output? What do you see instead?
LCDHost should start as expected, loading the layout correctly and displaying 
LH_Weather info. Instead, the background stays white (it's LH_Weather 
dependant), LH_Weather doesn't show up, LCDHost freezes and crashes.

What version of the product are you using? On what operating system?
A14 on Windows 7 x64, but this has been happening on A12 and A13 too.

Please provide any additional information below.
As said above, it may be that LCDHost starts too fast for LH_Weather to find a 
stable connection. This may require an SSD to be reproduced.
The log shows as follow when the error pops up:

2011-01-30 13:28:03 | D.. | loading layout 
"C:/Users/Alessandro/Documents/LCDHost/layouts/g19-eos/Eos.xml"
2011-01-30 13:28:05 | D.. | LH_Lua: 
&nbsp;<tt>Cairo&nbsp;version&nbsp;1.10.0</tt>
2011-01-30 13:28:05 | D.. | LH_Lua: loaded SamplesLuaCairoTest from 
"lua/cairotest.lua"
2011-01-30 13:28:09 | .W. | "LH_Text" unresponsive, restarting
2011-01-30 13:28:09 | .W. | "LH_Weather" unresponsive, restarting
2011-01-30 13:28:10 | D.. | Exception in plugin child thread 6092

Or, in place of the last line, it may point to thread 6108, or 5748, or maybe 
other threads. I don't know much about programming so I don't know if knowing 
the exact thread may be crucial or not.

Original issue reported on code.google.com by ale.i...@gmail.com on 30 Jan 2011 at 2:07

[GoogleCodeExporter]

Does LCDHost actually crash (does Windows show a crash log for it) or just 
becomes unresponsive? Does it recover from that state eventually?

Original comment by linkdata...@gmail.com on 30 Jan 2011 at 2:21

• Changed state: Accepted

[GoogleCodeExporter]

Yes, LCDHost crashes and Windows pops up the crash log, then it just closes. I 
have to manually start it again each time.

Original comment by ale.i...@gmail.com on 30 Jan 2011 at 2:28

[GoogleCodeExporter]

Currently unable to replicate.

Can I ask you try disabling LH_Weather, restarting and, once everything seems 
to be complete, then load the plugin manually. This would just confirm it is 
the timing (i.e. it loads too fast).

LH_Weather should handle a missing 'net connection. I'll double check that 
though.

For now I'm working under the assumption that the issue is one within LCDHost 
(as LCDHost is crsahing) albeit one triggered by LH_Weather.

~Triscopic

Original comment by a...@bridgesuk.com on 30 Jan 2011 at 6:31

[GoogleCodeExporter]

It does not happen every time anyways, but I'll try restarting a few times to 
make some stats with and without LH_Weather loading.

Original comment by ale.i...@gmail.com on 30 Jan 2011 at 6:35

[GoogleCodeExporter]

Ok, here are the results from the test, since LH_Text was giving some problems 
I also tested the RSS Feed and the Text plugin itself:
On 10 reboots with Eos layout, LH_Weather and LH_Text loaded, 9 times LCDHost 
just crashed, the last time it loaded but both LH_Text and LH_Weather appeared 
as crashed.
Then I tried deactivating LH_Weather: 4 crashes on 6 reboots.
I then removed the RSS part from the layout: 0 crashes on 6 reboots.
Thinking that was the problem, I retried with LH_Weather and LH_Text loaded but 
without the RSS Reader: this was the weirdest try, 7 crashes off 10 tries.
Then for science, I tried deactivating both LH_Weather and LH_Text, and it went 
5/5 fine.
So, the only times it was acting fine was when LH_Weather and LH_Text were both 
off, or with LH_Text on but removing the RSS reader AND LH_Weather off.

On every single attempt, reloading LCDHost or a single plugin AFTER the first 
crash worked fine.

Original comment by ale.i...@gmail.com on 30 Jan 2011 at 9:01

[GoogleCodeExporter]

This is probably a race condition in the network initialization. Until I can 
reproduce it, it'll have to wait.

Original comment by linkdata...@gmail.com on 15 Feb 2011 at 3:30

• Added labels: Priority-Low
• Removed labels: Priority-Medium

[GoogleCodeExporter]

I have exactly the same behaviour.
OS is Windows7 64bit.
CPU is a Intel Q6600 with 8GB RAM.
LCD Host A14, EOS layout.
I updated to the new Logitech Gaming Software but the problem stays the same. 
Sometimes it works, most time LH_Text and LH_Weather crashing and LCDHost stop 
working.

Here is the Application Log from the Event Viewer.

- <Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
- <System>
  <Provider Name="Application Error" /> 
  <EventID Qualifiers="0">1000</EventID> 
  <Level>2</Level> 
  <Task>100</Task> 
  <Keywords>0x80000000000000</Keywords> 
  <TimeCreated SystemTime="2011-02-25T08:12:25.000000000Z" /> 
  <EventRecordID>177716</EventRecordID> 
  <Channel>Application</Channel> 
  <Computer>QUAD</Computer> 
  <Security /> 
  </System>
- <EventData>
  <Data>LCDHost.exe</Data> 
  <Data>0.0.0.0</Data> 
  <Data>4d3f5411</Data> 
  <Data>ntdll.dll</Data> 
  <Data>6.1.7601.17514</Data> 
  <Data>4ce7ba58</Data> 
  <Data>c000071c</Data> 
  <Data>0009ad42</Data> 
  <Data>1780</Data> 
  <Data>01cbd4c3ab8045d9</Data> 
  <Data>C:\Tools\LCDhost\LCDHost14\LCDHost.exe</Data> 
  <Data>C:\Windows\SysWOW64\ntdll.dll</Data> 
  <Data>f9fd2b58-40b6-11e0-985e-001d7d05edc7</Data> 
  </EventData>
  </Event>

Original comment by swaydiz...@bossmail.de on 25 Feb 2011 at 8:24

[GoogleCodeExporter]

Issue 31 has been merged into this issue.

Original comment by linkdata...@gmail.com on 5 Mar 2011 at 3:10

[GoogleCodeExporter]

There are several changes in A15 that affects both how networking is used and 
how plugins are initialized. Once A15 is released, please test again and report 
results here.

Original comment by linkdata...@gmail.com on 5 Mar 2011 at 3:15

[GoogleCodeExporter]

The crash bug seems fixed with A15, LCDHost starts as expected and doesn't 
crash, but I haven't tested deeply. Unless it happens again, it should be good 
now.

Original comment by ale.i...@gmail.com on 10 Mar 2011 at 7:49

[GoogleCodeExporter]

I confirm bug gone on my end with 15.

Original comment by jkosso...@gmail.com on 10 Mar 2011 at 7:50

[GoogleCodeExporter]

Good. Closing this.

Original comment by linkdata...@gmail.com on 12 Mar 2011 at 12:35

• Changed state: Verified