Zach Welch wrote: > Hi all, > > Unless I accidentally missed something, the attached patches provide the > last _functional_ changes in Jeff Williams' J-Link patch; my final task > will be to extract and apply the debugging improvements that he made, as > it looks like they will then need to be immediately put into use. > > These patches do the following: > > * Add the new tap_get_tms_path_len function which -- when combined with > a new state table -- allows drivers to execute correct TAP state paths. > > * Updates the transition table in tap_get_tms_patch. This will require > drivers use the tap_get_tms_path_len function, rather than assume there > are always 7 steps. > > * Switches the J-Link driver over to using the tap_get_tms_path_len > function, rather than assuming 7 steps. > > These patches are _not_ ready to be committed, for reasons so numerous > that they need not be mentioned here. Let's just cut to the thick: > what's right, what's wrong, and what do we want to do about this? > >
I nak the arrays. I hate them, they are a crash waiting to happen. Put the same logic into a series of switches please, nested if necessary. Switches reads better, and are less subject to breakage. We just changes the values of the #defines in the last 10 minutes, this kind of change breaks these tables in the future. Dick _______________________________________________ Openocd-development mailing list Openocd-development@lists.berlios.de https://lists.berlios.de/mailman/listinfo/openocd-development
