[ggo-discussion] Re: possible bug on first time installation

  • From: Taegil Noh <nayas@xxxxxxxxxxxxxxx>
  • To: ggo-discussion@xxxxxxxxxxxxx
  • Date: Sat, 11 Jan 2003 09:59:15 +0900

Peter Strempel wrote:

>Now, this is strange. It _should_ not hang up but start with the just 
>created defaults. There is a file $HOME/.ggo.log where stderr gets redirected 
>to
>(however not yet when the configuration is loaded, so you got the above 
>message in your xterm). If something strange happens, please have a look in 
>.ggo.log and
>in case email me the file. Please note, the log is overwritten everytime you
>start gGo, so copy it away in time.
>I cannot say what exactly is going wrong here without more details, but I
>suppose it is not a problem with the config file.
>  
>
Now I understand that 'hanging' has nothing to do with the configuration 
rc file...
I've tried to capture some log file, But The program hangs 'Before' log 
file get created. Actually, I'm suspecting that logging is something 
that hangs gGo in My environment.
Log file was not yet created when gGo hangs on my system... So Natually, 
I've tried to follow where the program stops...

Too bad that I barely know about Java... That JDB is somewhat (largely) 
different from Gdb, the only debuger I've known... Anyhow I've unjared 
gGo and run jdb like this.

----

nayas@nlps4:/opt/home/nayas/gGo/lib$ jdb ggo.gGo
Initializing jdb ...
 > run
run ggo.gGo
Set uncaught java.lang.Throwable
Set deferred uncaught java.lang.Throwable            
 >
VM Started: Could not load file '/home/nayas/.ggorc'
java.io.FileNotFoundException: /home/nayas/.ggorc (No such file or 
directory)

(<---- the haning point. nothing happens after this. so I press enter... 
which just gives me jdb prompt. (I donno why. multithreaded debugging? 
hmm. so I've tried to see what threads are there. (took some mins to 
find that 'threads' command :( )

 > threads
Group system:
  (java.lang.ref.Reference$ReferenceHandler)0x10b Reference 
Handler           cond. waiting
  (java.lang.ref.Finalizer$FinalizerThread)0x10a  
Finalizer                   cond. waiting
  (java.lang.Thread)0x109                         Signal 
Dispatcher           running
Group main:
  (java.lang.Thread)0x1                           
main                        running
  (java.util.logging.LogManager$Cleaner)0x1e9     
Thread-0                    unknown
  (java.lang.Thread)0x1b5                         Java Sound event 
dispatcher cond. waiting
 >

---
I donno a thing about gGo (yet!)... but logging thread looks suspicious...

In My system, when there is any file named $(HOME)/.ggorc exist, gGo 
goes on and does not hangs. when there is no such file, it hangs. 
nothing happens after '.ggorc not found'. As Peter pointed out, Logging 
behavior is different when there is a valid .ggorc exist or not. maybe 
the redirectioning of stderr has some problem on my system. Possible jre 
installation problem, or maybe just my account has some problem, I don't 
know. But I'm trying to locate the problem.

>Actually, giving gGo an empty file can wreck things seriously. Maybe I 
>should add some check here, but so far noone has tried that yet and told me. 
>:*)
>But I rest in good confidence, when a threaded user manually edits the config
>file and things go wrong, I am not to be blamed. I admit the idea with "touch"
>was good and I had probably done that, too, but as my code works right now, 
>this cannot fix it. Theoretically the missing config file should not cause any
>problem.
>
>  
>
 I see. Users should not touch that file... :) Lets add first line of 
.ggorc as "== do not edit this file by hand ==" as they does.

>As I said above, the first time you alter some options, the .ggorc file is
>saved. However, in your attempt with the empty config file this should most
>probably not work. To create a new default config file, the best way is to
>delete the old and start gGo again. I guess the hang you had was some other
>problem, which exactly is difficult to tell now.
>
>  
>
now I see that the .ggorc file creation is done when first option change 
happenes, not at the first running time...

>So I suppose you somehow got some configuration working. However, it 
>shouldnt be a problem normally, and _you_ seem to know how to fix things, but 
>not 
>everyone does and would have given up already.
>
I had a valid copy of .ggorc file. (from the solaris server at our lab, 
eh, me a graduate student majoring CS.) Just overwrote it. gGo does not 
have any problem on Solaris 9 out-of-box environment...

---

I'm gonna check some more things. Well, trying different version of JRE, 
or will try to learn JDB. I'm gonna write again when I find something.

Thank you for your reply, and thanksagain,
for this cool go client,

cheers Peter!


nayas







Other related posts: