RE: [ros-dev] More RBuild issues...

[Steven Edwards]

--- Casper Hornstrup [e-mail] wrote:
> > Coudn't they written to hd ? On reboot reporterror would ask if he may
> > send it.
>
> Exactly.

You can't trust the system to write to the disk on a BSOD. If the crash is bad enough to BSOD then
it is bad enough to hose a fat16/32/ext2 fs. A crash that cannot be recovered from that results in
a BSOD should ONLY BSOD. If you can figure out a way to fix /DEBUGPORT=FILE to flush its buffers
and safely write to disk on BSOD I have a bounty for you. This is way doing a memory dump or
minidump of a crash is inheritly risky which is why it should be off by default.

Thanks
Steven

__________________________________ 
Do you Yahoo!? 
Yahoo! Small Business - Try our new Resources site
http://smallbusiness.yahoo.com/resources/

[WaxDragon]

That is why us real testers use /DEBUGPORT=COM1. ;0)

On 5/30/05, Steven Edwards [e-mail] wrote:
> --- Casper Hornstrup [e-mail] wrote:
> > > Coudn't they written to hd ? On reboot reporterror would ask if he may
> > > send it.
> >
> > Exactly.
>
> You can't trust the system to write to the disk on a BSOD. If the crash is bad enough to BSOD then
> it is bad enough to hose a fat16/32/ext2 fs. A crash that cannot be recovered from that results in
> a BSOD should ONLY BSOD. If you can figure out a way to fix /DEBUGPORT=FILE to flush its buffers
> and safely write to disk on BSOD I have a bounty for you. This is way doing a memory dump or
> minidump of a crash is inheritly risky which is why it should be off by default.
>
> Thanks
> Steven
>
> Do you Yahoo!?
> Yahoo! Small Business - Try our new Resources site
> http://smallbusiness.yahoo.com/resources/
> _______________________________________________
> Ros-dev mailing list
> [e-mail]
> http://reactos.com:8080/mailman/listinfo/ros-dev
-- 
<kjk_hyperion> I have a penchant for mischief, property damage and stalking
<kjk_hyperion> and cheesecake, of course

[Maarten Bosma]

Steven Edwards wrote:
> You can't trust the system to write to the disk on a BSOD. If the crash is bad enough to BSOD then
> it is bad enough to hose a fat16/32/ext2 fs.

On Linux you can use a LVM snapshot volume for stuff like that. Of curse 
blue.sys needed some kind of independent file system and hard disk driver.

Maarten Bosma