For Windows users, *please* continue to link to the D2XX import library, as you always have done. Switching to an alternative would only be acceptable if you:
1) Don't break other tools that do use D2XX for the same JTAG hardware 2) Allow channel B to be used as a standard COM port 3) Ensure performance must be at least as good If you can't do all of the above, you'll kill OpenOCD for use on Windows. Item (1) is important for those of us that use JTAG adapters with more than one tool. For example, I use Amontec JTAGKey-Tiny with Rowley Crosssworks for ARM and OpenOCD. Would your proposed replacement for D2XX allow the same device VID and PID to be used by two different drivers? A common configuration for FT2232 is JTAG + UART. On Windows, we get JTAG plus a virtual COM port, which we can use as a console for the target. Will your proposed replacement for D2XX still provide this? If not, you've made the project half as useful as it is now. Personally, I would pursue options for keeping things the way they are. A couple of observations lead me to think that FTDI intend their D2XX driver to be used by anyone using products with their devices in: 1) The driver is publicly available *and* advertised... with example projects! 2) A statement: "The driver may be distributed in any form as long as our license information is not modified." I've found FTDI very helpful in the past so I've e-mailed them to ask for their stance on this. As a British company, we should expect sanity to prevail. In the mean time, how about a 'make' option for those of a nervous disposition, to avoid linking with FTDI's import library: make --enable-paranoid-mode Regards Tim Clacy _______________________________________________ Openocd-development mailing list [email protected] https://lists.berlios.de/mailman/listinfo/openocd-development
