old patches that might have been missed

old patches that might have been missed

[Oliver Kiddle]

Checking over mailing list messages that arrived at times when I was
busy with other things, I see there are some patches that didn't get
applied. Without having gone into details, some of the following
seem reasonable enough:

42101, dana: Fix value of $? when used in array assignment
42113, dana: Millisecond and microsecond specifiers for TIMEFMT
42180, dana: Completion segmentation fault
42234/42226, Stephane: kill %jobspec tries to kill dead processes
42297, dana: Fix (e) subscript parameter used in scalar context
42363, Eitan Adler: REPORTMEMORY units

Any thoughts on any of those?

There's also the Inf/NaN changes which I was leaving until after 5.5 was
settled, which I guess it is now.

Finally, there's 42533 that was rejected because it was suggesting
disabling typeset family reserved words in ksh emulation which I
wouldn't think improves compatibility with real ksh. But some subset of
that patch may be reasonable, for sh emulation only.

Oliver

Re: old patches that might have been missed

[dana]

On 9 May 2018, at 10:24, Oliver Kiddle <okiddle@yahoo.co.uk> wrote:
>42113, dana: Millisecond and microsecond specifiers for TIMEFMT

That one was merged, cb04ae40af

dana

Re: old patches that might have been missed

[Peter Stephenson]

On Wed, 09 May 2018 17:24:25 +0200
Oliver Kiddle <okiddle@yahoo.co.uk> wrote:
> Checking over mailing list messages that arrived at times when I was
> busy with other things, I see there are some patches that didn't get
> applied. Without having gone into details, some of the following
> seem reasonable enough:

Thanks, this certainly needed doing.  I've started looking.

> 42101, dana: Fix value of $? when used in array assignment

I slightly tweaked this to avoid a compiler warning.  Manipulating the
status during code execution is never elegant (as you may have noiced
with the patch a couple of days ago).

> 42180, dana: Completion segmentation fault

There may be more to this problem, but without a research project making
the code a little safer as here seems a definite improvement.

I've submitted these two.

I'll look at the others.

> There's also the Inf/NaN changes which I was leaving until after 5.5
> was settled, which I guess it is now.

Sounds OK to me, though I don't think anything I do is going to be
particularly affected.

> Finally, there's 42533 that was rejected because it was suggesting
> disabling typeset family reserved words in ksh emulation which I
> wouldn't think improves compatibility with real ksh. But some subset
> of that patch may be reasonable, for sh emulation only.

Yes, I think so.

pws