[Psi-devel] [BUG] Unusable Psi

[Michał Jazłowiecki]

textshell wrote:

>> My question is: Should gdb return to its (gdb) prompt now, or it will 
>> wait for something to break into debugger? Ctrl-C doesn't seem to work 
>> (no response) and Ctrl-Break terminates gdb.
> 
> it should return to the prompt on Ctrl-C. Maybe you did something unusual
> with the terminals on windows. Windows is quite flakey with those
> commandline/console based apps, esp. if some kind of unix programms
> are involved (any cygwin <-> windows is really messed up)

Right, Windows cmd.exe is not so advanced as bash (however it's far more 
Unix-like than command.com was ;)). Anyway, I'm using GNU gdb 5.2.1 for 
MinGW on standard cmd.exe. The only thing that differs from standard 
cmd.exe settings is that I use quick copy&paste operations (just like 
the ones supported by gpm (I hope I remember the name for console mousse 
deamon right) on Unices).

>> Anyway, something has changed over the time as Psi does not freeze 
>> anymore while connecting to chrome.pl. And I didn't change anything in 
>> Psi options nor did I recompile Psi. I will retry the procedure once Psi 
>> starts freezing again.
> 
> i still think this is the gnupg bug. It seems to depend on who is online 
> while testing.

I don't even have GnuPG on my current system (that experiences problems 
with chrome.pl).

> can you please apply the attached information gathering patch and if psi
> freezes again post the last page of console output to the list?

I'll do that in a moment.

> And please try to find out if you have a gdb process around if psi freezes.

It's still out there once Psi freezes.

> I can't reproduce the bug here, so need a bit more of information...

I was able to reproduce the Psi behaviour at work and again at home in 
the evening.


-- 
Michał Jazłowiecki (michalj)
Psi Forum & Wiki Moderator :: Psi-Daisy Author