applet.internal.selftest: permit pin mapping to be direct instead of swapped #207
Conversation
|
You can use a ribbon cable crimped with an offset. |
... that's a very fair point. I'm not sure how common that'll be as an "off the shelf" part though, and I don't have any parts to hand (or a cable that I'm willing to rework). |
|
I've never used off-the-shelf ribbon cables with Glasgow; it's really easy to crimp IDC cables so there's no point. |
|
I second this, I have some off-the-shelf straight through cables and would be great to be able to use them here. Currently I am just using jumpers. I think everybody playing with a glasgow should have jumpers available, so I don't think this matters that much, but IMO it would definitely be a good addition. |
I have to admit that for some reason it did not occur to me (while writing the selftest applet) that you could simply twist the cable. That's clearly much easier for everyone (myself included!) so I think we should just switch the selftest applet to always assume that. |
attie
force-pushed
the
selftest-pins-straight
branch
from
6e57af3
to
b7187b2
Compare
Ah, yes - I should have been more explicit, sorry.
Done! |
I was suggesting to make the new behavior the only available. Even I don't actually have a selftest cable crimped according to the "old" behavior because... I accidentally crimped it for the "new" behavior but also made it too short to be usable. |
attie
force-pushed
the
selftest-pins-straight
branch
from
b7187b2
to
16f5305
Compare
Ah, okay. How's this? |
Currently, the
glasgow run selftest pins-looptest requires that pins are mapped as opposites / swapped (i.e: A0:A7 to B7:B0).I don't think it's possible to achieve the required mapping with ribbon cables, as each port's data pins would end up on the opposite port's ground pins (in a cable where pin 1 maps to pin 20 on the other end).
As I have direct cables to hand (where pin 1 maps to pin 1 on both ends), I'd like to use this to perform the self test, instead of 8x jumpers.
This patch adds a
--pinmap-straightflag that permits the "straight-through" A0:A7 -> B0:B7 mapping, instead of the "straight across the board" mapping that currently stands.It's not likely to be beneficial for production runs where a test jig is available, but should help those who don't have access to one, and are bringing up / testing a non-trivial number of boards.