Re: [lmi] Inversion of control

[Greg Chicares]

On 2018-09-18 12:04, Vadim Zeitlin wrote:
> On Tue, 18 Sep 2018 02:31:29 +0000 Greg Chicares <address@hidden> wrote:
> 
> GC> On 2018-09-18 01:07, Vadim Zeitlin wrote:
> GC> > 
> GC> > [...] But I'm fine with this change except that I'd really, really
> GC> > like to ask you to postpone it until you merge my pagination changes
> GC> > because there are going to be tons of conflicts between them.
> GC> 
> GC> I figured it wouldn't matter exactly when I apply the wxPdfDoc makefile
> GC> changes here:
> GC>   https://github.com/vadz/lmi/pull/91
> 
>  Yes, this won't conflict with anything else.

I can go ahead and apply it now. But the rest is more important:

> GC> But do you have other changes that are ready to merge now? 
> 
>  Yes, but these changes depend on the PR 91 above and also the yet
> unsubmitted PR changing the wxWidgets commit used by lmi, so I hadn't made
> a PR for them yet as you wouldn't be able to apply it yet anyhow.
> 
>  I.e. my plan was (and still is, unless you tell me differently) to submit
> both the wx PR and the rest of my changes on September 24, with the
> understanding that you will need to apply the PRs in order.

I'm afraid that the work I have in progress will make it difficult
to apply these changes that you already have. Therefore, let me
propose an alternative plan:

* PHASE ONE
- I'll apply PR 91 now.
- You tell me what wxWidgets SHA1 to update to. (Or, equivalently,
  let me have a patch that does that; but the patch would be trivial,
  and I'm trying to save you the trouble of writing it.)
- Tell me how to integrate your latest stable set of changes. A new PR
  rebased to current HEAD would probably be easiest, because I know
  exactly how to handle a PR.

This brings your latest stable changes into savannah HEAD, so nothing
I do subsequently will make merging those changes any harder.

* PHASE TWO
- On 24 September, we'll upgrade wxWidgets again if necessary.
- Meanwhile, you can continue to make changes, and they'll probably
  be difficult to merge; but the difficulty is less because it arises
  only with phase-two changes.

Assumptions:
- This alternative plan will actually produce worthwhile savings in
  total effort required on your part.
- You can complete your part of phase one very quickly. I've made
  commitments and don't have much time to meet them, so I can't
  delay my work for long. Is, say, 16:00Z realistic?