Hi Collin, Dmitry, and all,
The first patch removes a K&R declaration which are removed by C23, and
likely to be removed by compilers in the future. I assume all compilers
support ANSI C declarations nowadays.
...
mips:*:*:UMIPS | mips:*:*:RISCos)
But the question is not
Forwarding this config.sub patch for your consideration ... --thanks, karl.
Date: Wed, 27 Dec 2023 15:41:27 +0100
From: Aleksandar Rikalo
Resent-From: Aleksandar Rikalo
Resent-CC: automake-patc...@gnu.org
Resent-Date: Wed, 27 Dec 2023 18:04:02 +
Resent-Message-ID:
Resent-Sender: help-debb..
When combining variable assignments with a shell command, some older
shells (notably heirloom-sh and presumably also Solaris 10 /bin/sh)
Just to confirm: Solaris 10 /bin/sh does indeed also have this bug.
sparcsol$ /bin/sh -c 'x=good; x=bad :; echo $x'
bad
$ uname -a
SunOS sparcsol 5.10 G
Hi Jacob,
On Solaris 10, /bin/sh is a non-POSIX
shell and fails to execute config.guess because POSIX "$(...)" command
Yeah. It's been reported wrt config.{guess,sub} several times before,
but Dmitry has not replied, so I can only surmise that he does not want
to keep portability with S
1. $(...) breaks Solaris 10 /bin/sh.
2. Solaris 10 is still supported by the vendor, and people still use it
with GNU tools.
3. There is no technical benefit to $(...) in config.*.
What's the harm in using `...` a few more years in config.*?
Answer, as far as I can see: none. -k
At some point, failing to support $(...) is in the same ballpark
FWIW, I fully disagree. (Along with, it seems, everyone else except
you and Ben.)
1) There is no actual benefit to using $(...) over `...`.
It is purely cosmetic. In other scripts, fine. In config.*, no.
2) Using $(...) causes
Hi Dmitry,
I tried to reach the author of that change [2], but, unfortunately,
received no response.
Ben's lack of response is no reason not to revert this
unportability. After all, one steps down from maintainership in order
not to spend time thinking about it any more.
Clearly Ben had
Dmitry - I've received another bug report from a Solaris 10 user about
config.guess breaking because of using $(...).
I implore you to revert to good old `...`. As far as I can see, config.*
should be maximally portable, since after all their purpose is to
identify systems, and that means using `.
Forwarding ... so I guess $cpu is thumbv7a, which explains why the
previous case didn't catch it.
Date: Fri, 04 Dec 2020 00:30:01 +0300
From: KOLANICH
To: Karl Berry
Subject: Re: [kola...@mail.ru: Re: thumbv7[arm] support]
1. The apps were built with thumbv7a-none-linux-androideab
OK, I've just added thumbv7*,
Thanks.
although I still have some doubts because
"thumb" is translated to "arm" since 2011-03-23
(commit 45a6a4d02d7ee81332328b412441ac982e2c8a81).
I didn't know. Maybe $cpu is not actually "thumb" (instead "thumbv7"?)
on this system.
I have no pe
Solaris 10
Well, it is true that Solaris 10 /bin/sh does not understand $(...).
"Configure scripts will automatically locate a shell that supports
Yes, I saw that too. Which is fine for autoconf's configure scripts (and
I'm very glad they're doing it), but what about when config.{guess,s
Hi Dmitry,
> + | arm | arm[bl]e | arme[lb] | armv[2-8] | armv[3-8][lb] | armv7[arm] |
thumbv7[arm] \
thumbv7[arm] means thumbv7a | thumbv7r | thumbv7m - did you mean that?
Yes, evidently that is what is needed, from his response that I
forwarded on Nov 21. I'll include it below again.
Hi Dmitry,
It's time to join the 1990s and replace the backtick substitutions in
config.guess and config.sub with the $(..) command substitution
I admit I don't understand the reason behind this last change that Ben did.
What's wrong with `...`?
No question $(...) is nicer in principle,
Dmitry - passing on ...
Date: Sat, 21 Nov 2020 13:59:59 +0300
From: KOLANICH
To: Karl Berry
Subject: Re: thumbv7[arm] support
As I understand, thumbv7 is just armv7, but using thumb-2
instructions. I have googled the variants, thumbv7r and thumbv7m are
also found in the Net, and I have
Hi Dmitry - thanks for getting back so quickly.
I hope Mr. Kolanich can answer the questions. I know nothing.
> + | arm | arm[bl]e | arme[lb] | armv[2-8] | armv[3-8][lb] | armv7[arm] |
thumbv7[arm] \
thumbv7[arm] means thumbv7a | thumbv7r | thumbv7m - did you mean that?
My understanding
We received this report on automake from Kolanich (thanks):
(https://debbugs.gnu.org/cgi/bugreport.cgi?bug=43945):
Hello. When I crosscompiled some app I had to patch some file
generated by autoreconf -i, where I have to add thumbv7[arm] (using
the analogy to armv7[arm] already present
Hi Karl,
As you likely already know, Ben Elliston is the maintainer of
config.{guess,sub}. (He's on this list.) So he's the one who will have
to make any changes. I never use Solaris myself and have no insight into
the issue. Everything you write sounds reasonable to me :). --best, karl.
___
karl.
Date: Tue, 17 Jul 2018 17:00:22 +0300
From: Apostolos Syropoulos
To: Karl Berry
Subject: Re: [tlbuild] x86_64 or i386 on Solaris
karl> We run config.guess, which should do the right thing. [...]
No this is not working properly! It reports that my system is
i386-pc-solaris2.11
while i
Hi Ben - in the new config.sub (2017-09-12) there is a space before tab,
which gnulib git will not accept (not my decision). Please fix?
remote: build-aux/config.sub:1368: space before tab in indent.
remote: + # -solaris* is a basic system type, with this one exception.
Tha
Hi Ben,
For your consideration ...
grep -q is not supported on the sadly-still-widely-used Solaris 5.10. I
realize that at least some, maybe all, of these tests only execute on
other systems, but it somehow seems simpler and more readable to just
always redirect to /dev/null instead of worrying
Ben and all - courtesy of Adam and Dick (cc'd, to whom all questions
should be directed :), here is our prospective patch to eliminate
invoking the compiler on new-enough Macs. (diff -c output seemed the
most readable.)
The idea was not to perturb the results, but just avoid the compiler
invocati
Ben, Jan, et al.,
Sorry; this Mac thing has once again devolved into meandering messages
(as Mac things always seem to do) with no immediate relevance to
config.guess. I did my best to avoid this outcome when I asked for the
submissions, but obviously did not succeed.
Anyway, can you please cons
22 matches
Mail list logo
