Re: Bug hunt: SIGINT and timeout update (fwd)

[Eskil Heyn Olsen <deity dbc dk>]

On Wed, 5 Jan 2000, Eskil Heyn Olsen wrote:

> I need to know, how many are experiencing The Timeout Problem, and how
> many are having The Sigint Problem.

Oh this is sooo creepy...

The two bugs are related to a certain point. If you start the pilot_applet
from the panel menu, press sync, and the gpilotd received a SIGINT and
terminates, you will see it as an timeout. I think this accounts for the
numerous timeout bug reports.

As for the SIGINT, thats just a load of bloody fun ; wednesday evening, at
work, on my laptop, I suceeded in reproducing the sigint using Manuels
guide. I peek and poke for at while and is simply baffled at the obscurity
(apparently, somewhere in the poll/select mechanism, a SIGINT was sent for
serial ports, the bug could not be triggered with eg. XCopilot).

Anyways, it was turd o'clock and I shutdown the laptop, packed up and went
home. About 50 mins later, at home, I unpacked, plugged in the pilot (same
pilot, other cradle), turned on the laptop (same machine), started
pilot_applet and gpilotd as described (same binaries), pressed sync,
awaited the SIGINT...

No. Synched as it should.

Only difference (except for the power plug in the wall, elevation above
sea level and air moisture as someone pointed out), is that the machine
was rebooted and has another IP address now (dhcp). Binaries, laptop and
pilot are the same.

Now I don't want to sound like MS tech support, but if you suddenly again
get sigints all the time, could you try drastic things, like booting (yeah
I know, sared uptime and stuff) and see if it disappears. I think I speak
for many, when I say this bug really needs to be nailed.

/dev/eskil
---