Re: [MiNT] Issue with Mint 1.19 and Litchi

[Alan Hourihane <alanh@fairlite.co.uk>]

On 08/02/15 11:16, Miro Kropáček wrote:

On Sun, Feb 8, 2015 at 11:17 AM, Peter Persson <pep.fishmoose@gmail.com> wrote:

I agree 100% with Markus in this case. It's not a good idea to change existing system calls in a way that affect any unknown number of existing programs, and compatibility for these APIs are handled through mintlib. I'd much rather see a nice implementation of Fselect2() :)

In my opinion it absolutely doesn't matter -- with the old Fselect you've got broken vim (and perhaps some other sparemint software), with the new Fselect you've got broken other software (MyAES, Litchi, ...).  And since the changes has already been done and all the authors are aware of that change, I tend to agree with Alan, let's not confuse things even more.

I'd thought about it more too, and there's possibly other software that just hasn't come forward yet.

Also, changing an existing system call isn't a good idea.

Regardless, the initial problem was vim, and given that relies on the MiNTlib, a recompilation of that program should fix it to use Fpoll() instead now and there shouldn't be a problem.

Programs that call Fselect() directly should continue to work without change.

So I've reverted the change to Fselect().

We should be all good now and no issues.

Thanks for everyones insights.

Alan.

NOTE: people with old vim binaries, upgrade.....