[haiku-development] Re: Busted!
- From: Rob Judd <haiqu@xxxxxxxxxx>
- To: haiku-development@xxxxxxxxxxxxx
- Date: Mon, 25 May 2009 08:51:08 +1000
Michael Lotz wrote:
I AM developing in Haiku, since months!
And Stephan's certainly not the only one. I'm developing on Haiku since almost
a year now. It's my main and indeed only platform. I very rarely revert to
BeOS for one thing or another (like once every few months) and I have one
Windows installation for a very specific task (I'm working on getting it done
on Haiku as well so I can remove that installation as well).
Nice to hear it!
Besides, running on Linux isn't viable; I'm developing kernel drivers.
The way you could do this the easiest would be to embed your driver into your
local Haiku tree and add it to your image. Then you could just build the
network boot image and boot right off it. This provides a great turn around
time.
I usually make a BeIDE project and install/deinstall with shell scripts.
Works for me, I haven't had a crash due to my own code lately.
The crash I reported earlier isn't related to r30825 at all, I just went
right back to r30782 ans tried a clean build (totally deleting
generated/objects) and I'm still getting problems with `copyattr`
throwing an exception and shutting down the build. This raises two issues:
1. Why can no-one else verify this yet?
Most probably because you have something very specific in your setup that
noone else has. This might be a certain piece of hardware (with a faulty Haiku
driver) or a certain configuration of the system.
The most helpful would of course be if you could track down why this happens
in your case. It's very hard for someone else to debug the issues you mention
without being able to reproduce them.
Yep. I'm realizing just how much detail is needed for someone else to
reproduce one of these bugs.
Since you seem to run into BFS corruption so reliably, could you provide some
info on how you set it up? BFS block size and the partition size you are
using, how full your filesystem usually is. Maybe it is more likely to happen
with that specific combination you are using.
Default 2k block size, nothing special. Partition size has been anything
from 7.5GB to 20GB and it doesn't seem relevant. Filesystem is nowhere
near full, I usually leave at least 20% margin. Corruption usually
happens after an unhandled kernel exception and crash, which is a
well-known mechanism. checkfs hasn't been able to remedy most errors,
especially the `phantom files` mentioned yesterday. I usually end up
with a rubbish bin full of apparently empty directories that can't be
deleted, because the OS thinks they contain something.
Rob
Other related posts:
