On Fri, 05 Sep 2014 08:17:24 -0700, Sean Healy <jalopeura@xxxxxxxxxxx> wrote:
On Fri, 05 Sep 2014 06:49:59 -0700, Axel Dörfler <axeld@xxxxxxxxxxxxxxxx> wrote:Without seeing any code, the only conclusion that can be drawn from the above would be some kind of user error :-)Well, it's not my code, it's stippi's. But he makes mistakes, too, I
Actually, I checked the initial commit message, and the original code was Ingo's; stippi was just the one who committed it.
guess. And I have the impression that previous testing was done using the fallback settings file (driver-settings format) instead of the normal settings file (flattened BMessage), so maybe the code that loads the normal settings file was never tested.I'll keep looking.
I shouldn't have assumed that the code in question worked before package management.
There was a bad logic test. Changed a '&&' to a '||', popped in a '!', and now it works.
I'm kind of annoyed that my eye kept skipping over that logic test, but seeing as that bug has been in place since the code was first committed over 4.5 years ago, and 4 different core devs have made changes to the code since then, I probably shouldn't feel too bad.
I guess everybody was just using the fallback file settings for configuration and never bothered generating a regular settings file.
Also, I suspect NetFS won't get much use outside of events like BeGeistert, since you need multiple Haiku machines on the same network to use it.
