jSyncManager OS X Transport now in CVS
Brad has been busy weekends trying to code
his own jSyncManager USB transport plug-in for Mac OS X using JNI and Apple's
IOKit. Here is a quick status report.
I'm happy to announce that I've checked the first
revision of my new Mac OS X USB transport for the jSyncManager
into CVS. I've been working on this transport on
weekends mostly, trying to wrap my head around Apple's
IOKit.
The code now in CVS works -- sort of. After fixing a USB configuration problem that caused me to miss the first packet (apparently, IOKit configures the device for me, and if I configure it myself I miss reading the first packet which contains the header for the NetSync packet being transmitted), I can read the first packet, and send the response.
Things after that, however, aren't so good. I'm not sure if the response to the wake-up packet isn't being received and/or processed on the handheld end, if IOKit isn't sending the packet, or if there is some timing issue that is causing me to miss the next packet, but after sending the wake-up response, I get...nothing. The new transport code eventually times out as the handheld doesn't respond.
I seriously doubt this is a problem with the initialization packet, as it works for pure NetSync via TCP/IP, and it works on Linux when using the javax.usb transport, so either the handheld isn't getting the initialization packet, or I'm not getting its response. Either way, the transport current exchanges two NetSync packets (one in, one out), and than soon aborts as no new data comes from the handheld.
This is somewhat irritating, as it's not really possible to diagnose the problem, and Palm lacks documentation. So I'm hoping someone can point me in the right direction.
Want to try your hand at a fix? Take a look in the contrib/osxusb_transport directory in our CVS repository to get the source code, and try building it yourself (I'm using Xcode v2.1, but these sources should compile with previous Xcode versions as well), and if you discover anything, let me know .
The code now in CVS works -- sort of. After fixing a USB configuration problem that caused me to miss the first packet (apparently, IOKit configures the device for me, and if I configure it myself I miss reading the first packet which contains the header for the NetSync packet being transmitted), I can read the first packet, and send the response.
Things after that, however, aren't so good. I'm not sure if the response to the wake-up packet isn't being received and/or processed on the handheld end, if IOKit isn't sending the packet, or if there is some timing issue that is causing me to miss the next packet, but after sending the wake-up response, I get...nothing. The new transport code eventually times out as the handheld doesn't respond.
I seriously doubt this is a problem with the initialization packet, as it works for pure NetSync via TCP/IP, and it works on Linux when using the javax.usb transport, so either the handheld isn't getting the initialization packet, or I'm not getting its response. Either way, the transport current exchanges two NetSync packets (one in, one out), and than soon aborts as no new data comes from the handheld.
This is somewhat irritating, as it's not really possible to diagnose the problem, and Palm lacks documentation. So I'm hoping someone can point me in the right direction.
Want to try your hand at a fix? Take a look in the contrib/osxusb_transport directory in our CVS repository to get the source code, and try building it yourself (I'm using Xcode v2.1, but these sources should compile with previous Xcode versions as well), and if you discover anything, let me know .
Posted: Monday - June 27, 2005 at 11:28 PM
Quick Links
Statistics
Total entries in this blog:
Total entries in this category:
Published On: Aug 30, 2007 08:14 PM
Total entries in this category:
Published On: Aug 30, 2007 08:14 PM