Hi Dominik,

Domink wrote:
> as I've updated my branch, I discovered someone changed the way devices 
"are located."
No I didn't. Devices as well as services are still located the same way. 
However, because I converted ata_bd to a DDF driver, the name of the service
exported by the driver has changed from bd/ataNdiskM to devices/\hw\... I 
sent a heads up message for that, I believe and it was discussed.


 
"I didn't like the way it was done before and now I consider 
it even worse."
Sure the new service name is less human-friendly, but that's because it's 
something different - it corresponds to the physical device path.



"Yes, it has clear structure, but why do I have to write 
"devices/\hw\pci0\00:01.0\ata-c1\d0"? Why can't it simply be the full 
path?""Right now I just write the full path where I can enjoy 
path-completion from bdsh and then erase the beginning or cd into that 
directory, as JS previously suggested. But it's just annoying."



As I suggested you can also cd /loc and then just tab-complete...

 
"How I miss 
"/dev/sda"..."We'll surely need a mechanism for assigning stable (and 
possibly more concise) names, too. Solaris, for example, assigns net0, net1,
etc. to network devices based on some policy (e.g. if you plug in the same 
USB wifi adapter into a different slot, it will be recognized as the same 
network device). This remains to be implemented.

"

Also, I couldn't find anything documenting these changes, including making 
ata_bd into a DDF driver. If there is any documentation, could anyone 
provide a link? Otherwise, how should I know that from now on the usual 
"bd/ata1disk0" changes into that string mentioned before?"Okay, my fault, I 
thought I sent up a heads up. The conversion is mentioned in the meeting 
minutes, though.

Cheers,
Jiri
_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/listinfo/helenos-devel

Reply via email to