Comment 26 for bug 218671

> Bryce, can you please review this and upload to hardy/intrepid-proposed, and make sure it gets into jaunty? It seems to be fixed upstream in 1.2.2. Thanks!

Martin, are you certain this should be uploaded? This bug has a number of irregularities...

1. Several patches are mentioned in this bug report (comments #7, #18, #19, #23) each of which is different. I'm assuming you mean patch #7 should go in. However, in reviewing this patch I don't think it's a proper fix. Presumably the current axis settings in -elographics at least works for some hardware, else why would they have coded it that way? But this patch reverses the y-axis. I can believe that there may be some hardware that has backwards y-axes, which this patch would fix, however I would think this would thence cause regressions for all existing hardware with y-axes that go the other direction.

2. In addition, patch #7 is not upstream (neither is patch #19) so this conflicts with a few people who indicate the problem is fixed with the 1.2.2 release. Are multiple bugs being reported in this report?

3. The only code change from 1.2.1 to 1.2.2 simply adds a check for XINPUT ABI version 3. This would fix an ABI breakage issue, which would have very different symptoms than this bug. So while that looks like a safe and easy fix to backport, I really don't think it's what is needed to fix this issue.

Due to these irregularities, I'd be much more comfortable if this patch were taken upstream before considering whether to accept it for SRUing. I think there's just too much risk of regression.