[syslinux] where to swap

[Ady Ady]

> On Sun, Mar 26, 2017 at 6:43 AM, Ady Ady via Syslinux
> <syslinux at zytor.com> wrote:
> >
> >> On Sun, Mar 26, 2017 at 09:50:17AM +0000, Ady Ady via Syslinux wrote:
> >> > > i wrote:
> >> > > > > I think it still has good reason to exist.
> >> > >
> >> > > Ady wrote:
> >> > > > Would you please clarify what do you mean with "it"?
> >> > >
> >> > > "It" was the code part about
> >> > >
> >> > >      xor     %cx, %cx        /* Clear EBIOS flag. */
> >> > >
> >> > > which Martin mentioned as "can be ignored" in the the text which i
> >> > > quoted before my statement.
> >> ...
> >> > _ clarification from Martin (since I asked a similar question regarding
> >> > his own email/wording); and
> >>
> >> Yes, it's the xor %cx, %cx we're talking about. Which might seem
> >> overkill now that we know that David's BIOS _didn't_ corrupt CX.
> >> And given that space in isohdpfx is very limited (432 bytes) perhaps
> >> the right thing might be to not have that xor in there.
> >>
> >> But then better safe than sorry, and given that there is space _now_,
> >> why not?
> >>
> >>
> >> > _ after reverting the (wrong) commit 48e94f4fa7b3c3..., whether the 2
> >> > sections of Martin's patch should rather be applied as one commit, or 2
> >> > separate commits, or just the first part of the patch should be
> >> > applied.
> >>
> >> The patch in <http://www.syslinux.org/archives/2017-March/025706.html>
> >> should be applied after reverting
> >> 48e94f4fa7b3c32cbd43b6e57c64bc933f76d059.
> >>
> >>
> >> Except possibly the xor %cx, %cx part as above depending on how
> >> causcious or spacesaving you want to be.
> >>
> >> It seems both Thomas and I think it should remain.
> >>
> >>
> >> --
> >> MartinS
> >
> >
> > Then I would suggest:
> > 1_ revert the wrong commit;
> > 2_ apply the first part of the "right" patch in one commit, as it
> > corrects one known bug;
> > 3_ apply the second part of the "right" patch in a separated commit,
> > which can potentially deal with a different situation than the first
> > part of the patch, and, if for some reason additional space is needed
> > in the future, it could be considered on its own for future edition.
> >
> 
> Essentially the approach I took on the isohdpfx branch on my personal repos.
> -- 
> -Gene

 
Gene,

IMHO, not really. It should rather be in 3 commits: One reverts, second 
applies the first part of the right one from Martin, and third commit 
applies the second part of the patch.

Reasoning:

When/if a distro wants to cherrypick the commits, the first revert 
should not be needed. Having only 2 commits as you did in your repo, it 
makes the cherrypick unnecessary complex, having to take 3 commits 
instead of of 2.

Additionally, having 3 commits would make it easier to apply the same 
patch (by using the last 2 commits I am sugggesting) to the older 
branches (syslinux-3.xx and syslinux-4.xx), which are even more 
relevant in this situation than the need for 6.xx.

Regards,
Ady.