On 01/09/2011 01:04 AM, Ivan X wrote:
On 1/8/11 9:16 AM, in article f_udnT750Ms87LXQnZ2dnUVZ_tGdnZ2d@giganews.com, "Steven Hirsch"<snhirsch@gmail.com> wrote:On 01/07/2011 07:59 AM, Ivan X wrote:Netatalk 2.1.5 (compiled from source, with randnum authentication and netboot) runs very nicely in Ubuntu Server 10.10 under VirtualBox 4.0.0.And, although it appears to violate some of the more nit-picky aspects of the protocol definition, it's a good idea to apply my one-line patch to preserve file modification stamps. Without it, file copied to the server from a IIe client receive the current time. I greatly prefer having the modification time preserved.Yes. In my setup, and instructions for it, I apply that.There were a handful of small fixes required to the Ubuntu package. I sent them to pitz since he appeared to be working on a release to the community.By "fixes" what do you mean? Apart from having to install the various library packages you originally outlined, I didn't have to make any changes to Ubuntu itself; as for netatalk, there were a small handful, such as your date mod, and editing of its init.d script (and its config scripts, of course).
That's what I'm talking about. I moved all that stuff into a modified debian package and sent the diffs to pitz. I was of the impression he was working on such a package and didn't want to duplicate efforts.
Every LocalTalk-to-Ethernet transceiver I've tried is still flawed in one way or another (and I've not yet obtained the fabled GatorBox CS). Here's the rundown:Yea, verily... Geoff Body has been plugging away at network traces to evaluate why the Farallon devices drop sessions on some boxes and work on others. I owe him a couple more captures (and a large debt of gratitude for taking this on!).I'd love to hear the results from that too. I'm even more mystified as to why the Asante and Dayna devices cause GS/OS to turn the pointer into a blotch and otherwise hang when the AppleShare CDEV is opened.
Well, you're doing better than I ever did with the Asante. I never had the slightest bit of success with mine. We did discover yesterday that the system speed appears to have something to do with the IIgs disconnects over the Farallon devices.
I tried disabling the ZipGSX during boot and network init, then had the utility poke it into full speed later. At this point, things began working properly and the machine would stay connected over the EtherPrint LT. Geoff asked me to try at normal (2.8Mhz.) speed - whereupon the drops re-appeared. The net is that I must have the machine boot at normal speed and switch to accelerated later on.
Steve