Re: [PATCH] {maint} remake: behave better with non-GNU make in subdirectories

[Stefano Lattarini]

On Sunday 29 May 2011, Peter Rosin wrote:
> Den 2011-05-29 19:32 skrev Stefano Lattarini:
> > Hi Peter, thanks for the review.
> 
> I wouldn't call it review, it was more of a knee-jerk reaction
>
Well, I still appreciate it anyway.

> (and I haven't even read the actual patch).
>
There's still time to remedy that, right?  (Just kidding ;-)

> > On Sunday 29 May 2011, Peter Rosin wrote:
> >> Den 2011-05-29 16:26 skrev Stefano Lattarini:
> >>> Currently, every decent non-GNU make program makes it possible
> >>> to remake out-of-date autotools-generated files with a simple
> >>> "make Makefile" issued from the top-level directory.  And while
> >>> having to run "make Makefile" by hand isn't as convenient as
> >>> the GNU make feature of "automatic remake *and reloading* of
> >>> out-of-date makefiles", it is usually good enough and definitely
> >>> worth having.
> >>>
> >>> Unfortunately, a silly limitation in the current implementation
> >>> of remake rules prevents this idiom from working when it's used
> >>> outside the top-level directory.  Luckily, this limitation is
> >>> easy to   remove, and that's what this patch does.
> >>
> >> Nitpick: The tense is wrong for a ChangeLog entry, and there's a tab
> >> in there as well.  Suggestion:
> >>
> >>    Remove silly limitation that prevents this idiom from
> >>    working outside the top-level directory.
> >>
> >> I'm not sure if the essay style of the first paragraph should stay
> >> or go though?
> >>
> > I mostly agree with your objections.  What do you think about this
> > ChangeLog entry instead?
> 
> Hmmm, it just feels like you have identified some problem, and then
> you proceed to writing a patch.  Then you write some words to
> tell the world about the problem and the solution and put it all in
> the ChangeLog.
>
Actually, I did the other way around (mostly: identify problem, write
ChangeLog entry, fix problem, tweak ChangeLog entry if needed).  While
this might sometimes lead to overly longish ChangeLog entries, IMVHO it
also usually leads to better patches, as it forces me to express clearly
and unambiguously what I'm trying to do, and more importantly, why.

> It feels backwards to have the *original* problem description in
> the ChangeLog, but I'm no longer sure if it's really all that bad
> or if it just feels backwards _to_me_.
>
Do you have explicit, specifical objections to the ChangeLog change,
or suggestions about how to improve it?  I'm prepared to listen to
them, and tweak the ChangeLog entry accordingly.

> And regarding the tense, now that I think of it that's probably only
> the rule for the actual changes in the asterisk lines, not the "header",
> so that was probably off base too.  Not sure though.
>
In effect, using the present tense to describe a just fixed, and thus
past, problem, is suboptimal and might be confusing.  So what about
this updated ChangeLog entry?

    remake: behave better with non-GNU make in subdirectories
    Currently, with every decent make program, it is possible to
    rebuild out-of-date autotools-generated files with a simple
    "make Makefile".  But for this to work reliably with non-GNU
    make implementations, that command had be issued from the
    top-level directory.  This patch has removed such limitation.
    * lib/am/configure.am (am--refresh): Depend on `%MAKEFILE%'.
    * tests/defs.in (using_gmake): New function, backported from the
    `master' branch (and simplified).
    * tests/remake-subdir.test: New test.
    * tests/remake-subdir2.test: Likewise.
    * tests/remake-subdir-gnu.test: Likewise.
    * tests/remake-subdir-from-subdir.test: Likewise.
    * tests/Makefile.am (TESTS): Update.

> Me feels bad, sorry for the interruption.  I lost the confidence
> to voice my opinion...
>
Why?  I think that the modified ChangeLog entries prompted by your
answer are already an improvement over my original one; so I wouldn't
qualify your intervention as useless, or as an "interruption".

Regards,
  Stefano