>shopt -p extglob >/dev/null; s_extglob=$?
>shopt -s extglob
>( do processing )
>((s_extglob==1)) && shopt -u extglob
>I must think before posting. This still doesn't get over the fact
>that you've 'set -e'
Right - that's exactly the point. Since OP is running "set -e", he
should use a con
>Could fltexpr maybe link to some other libm that is correctly rounded,
>like GNU MPFR, or LLVM's libm, to mention just a few?
I doubt they'd want to make bash dependent on any external library
like that (that may or may not be installed on a given system).
But you are certainly welcome to take a
> Otherwise, you're *always* going to run into issues of this type.
I'm not. OP might - unless/until he reads and understands the "What every ..."
document.
=
Please do not send me replies to my posts on the list.
I
>Bourne Again SHell wrote:
>> $ fltexpr -p "0.1 + 0.2"
>> 0.30004
>> https://0.30004.com/
Of course there are no bugs in the OP and, of course, just about
everything there is covered by the usual "What every computer
programmer should know about floating point math" clic
>>
>> +The old format \fB$[\fP\fIexpression\fP\fB]\fP is deprecated and will
>> +be removed in upcoming versions of bash.
>>
>That's not true though, both bash and zsh still support it and neither
>party plan to remove it. It's better left undocumented.
I disagree that it should be left undocumen
>>Sure, except this isn't a script. It's in a file named Bash_aliases
>>contained in a directory named examples/startup-files. It was clearly
>>intended to be copied into one's ~/.bashrc file (or to be dotted in
>>from there).
>>Including it in your ~/.bashrc would effectively mask the real se
>May be the 'repeat' function in the 'examples' is a bit lengthy, and use
>subshells, the example I got is ...
I think all of the files in the examples/startup-files directory
should be treated as museum pieces. You wouldn't go into a museum and
start insisting that all the displays be changed
> On my machines (Slack 15 / LMDE 6) 'seq 100' prints a line 100 times.
> seq is an elf binary, not a function repeat doesn't exist as an
> executable or function.
To fix that, you need to "dot" (aka, "source") the file Bash_aliases,
which you will find in the "examples/startup-files" subdirecto
Very entertaining thread! Blast from the past!
Regarding the fundamental question of "Is this a bug or not?", note that
in computer stuff, yeah, most people's intuitive idea of what "repeat n
times" means is "Do it n times". And most implementations I've seen of
a "repeat" keyword in programming h
>What, no takers? Come on, guys, that can't be normal for the read
>command to get stuck like that. After all, I've given it the timeout
>option, "read -t 1", which means, "In one second, read-or-not, be out
>of there!", does it not? There must be something I could do to avoid
>getting stuck like t
For the benefit of anyone trying to follow this thread, code in question seems
to be:
>if awk '{exit $1 < $2 ? 0 : 1;}' <<< '1.1 1.2' && awk '{exit $1 < $2 ? 0 :
>1;}' <<< '1.3 1.4'
>then
> echo ok
>fi
Actually, that is correct code - that actually works, although no one
would ever do it like
tl;dr: This really has no place being discussed in public. Whatever differences
Individual 1 and Individual 2 have, they should be hashed out in private email.
Or they should step outside and settle it like the manly men they are.
I do have a strong opinion as to who is right and who is wrong, b
>> On 9/1/25 10:24 AM, pourko--- via Bug reports for the GNU Bourne
>> Again SHell wrote:
>> > Is there any possible (easy) way of transforming a bash loadable
>> > into a bash builtin?
>>
>> Sure. You just have to convert the information in the `xxx_doc' array
>> and the `xxx_struct' structure
>>Maybe it changed between 5.3 rc2 and 5.3 release. 5.3 rc2 was compiled as
>>is; no
>>local changes.
>It didn't.
I could disprove this, but as I said before, it is not worth arguing about.
I'm surprised people are still posting about it.
===
>>The crux of the confusion is the difficulty in identifying when
>>subprocesses are created.
>>- Brackets always ensure a subshell.
>Parentheses. Brackets are [ and ] in bash and POSIX nomenclature.
I think Europeans (**) often call them (*) "brackets".
I've noticed this a lot over the ye
Maybe it changed between 5.3 rc2 and 5.3 release. 5.3 rc2 was compiled as is;
no
local changes.
Anyway, it is what it is. Doesn't matter. Not worth arguing about.
=
Please do not send me replies to my posts on the
(>> == pourko, > == Greg)
>> Sorry, Greg, but this is no "compound" command, these are separate commands:
>>
>> $ shopt -s extglob; echo !(this)
>That is another case where it fails, correct.
>However, in your original email, you did indeed have the shopt inside
>a compound command. That's why
>> If the feature is really desired by the community as an essential feature, it
>> should be implemented as a new syntax.
>I looked at adding alternate syntax for doing command substitution without
>stripping trailing newlines, and it turned out to be easier than I expected.
>This patch adds t
>> This has never been documented.
>>
>If that's true, how are there scripts using it, and people asking about it?
>It's not just documentation that you've written that needs to be
>considered, because clearly this feature must be mentioned SOMEWHERE.
Yeah, that occurred to me as well, but I j
Connor wrote:
> I'm of the opinion that if there is a supported feature with no plans for
> deprecation
> or removal, it should be documented. Even if it's not widely used, it's still
> supported with (as you said) no plans for removal. reply via email to
I'm of the opinion that the feature in q
Just for kicks, here's an implementation of "if-then-else" called "ite", as a
function. Note that:
1) It is pretty much a drop in for the missing functionality (in the usual
${var:X} syntax), when used inside $().
2) All of the existing ${var:X} syntaxes are special cases of ite(
>I think they meant to use IFS=$'\n' rather than IFS='\n'
In which sense of the word "they" (*) ?
In which sense of the word "meant" ? (**)
(*) I'm pretty sure there's just one poster, not a bunch of them.
(**) I'm pretty sure there was no typo or mechanical error when he posted to
the list.
==
Martin wrote:
>On Tue, 5 Aug 2025 at 02:17, Stan Marsh wrote:
>> (I am talking about things like ${var:-default value})
>>
>> I have long wished there was a substitution for "substitute a string if var
>> is
>> unset else substitute nothing".
>I&#x
(I am talking about things like ${var:-default value})
I have long wished there was a substitution for "substitute a string if var is
unset
else substitute nothing". This would basically be the inverse of :+, which
means
substitute a string if the var is set (ignoring the actual value of the
v
What's with the spate of meaningless noise responses on this thread?
Really, folks, you need to get a grip!
As to the substance of this thread, yeah, if you carefully read the INSTALL
file (as I
did this afternoon), it is all there, but it is not a particularly easy read.
I get where OP is comi
>It is not a bug.
>> The SUS[1] says "If directory is an empty string,
>> the results are unspecified.".
>>
>> [...]
>>
>> [1]:
>> https://pubs.opengroup.org/onlinepubs/9699919799/utilities/cd.html
>This was changed in POSIX.1-2024:
>If _directory_ is a
>From: Carl Johnson
>Subject: Re: Segfault when running alias if PROMPT_COMMAND is not empty
>Date: Mon, 7 Jul 2025 22:18:30 +0300
>> Thanks for the report. In the general case, evaluating the value of a shell
>> variable as a command should save and restore the parser state. In this
>> edg
(I wrote):
>> It seems to me that there is never any need for $ in arithmetic contexts, so
>> the
>> obvious fix would be to just remove any $ found between (( and )). But from
>> what I
>> understand that is easier said than done.
(Causing Greg to write):
>Unfortunately, that's not true. Yo
(I wrote):
>> https://pubs.opengroup.org/onlinepubs/9799919799/utilities/V3_chap02.html#tag_19_06_02
>> I don't read POSIX specs. I have people who read POSIX specs.
(And Chet wrote):
> Consider asking them first. They're your people, after all.
Good one! Well done.
On 6/30/25 5:06 AM, Isabella Bosia wrote:
>>this causes a parser error since commit 35bc7025
>>a=(0 1 2)
>>b=(0 1 2)
>>x=0
>>((i=$x,a[b[i]]))
>>using $ anywhere in (( )) makes subscripts not expandable
(And then Chet wrote):
>This was one of the results of this 2021 discussion:
>https://lis
(I wrote):
>> I always thought the "colon-free" versions of the P.E. was a bash-ism, but
>> experimentation shows that it works in "dash" as well. And "man dash"
>> contains
>> the following text:
>> In the parameter expansions shown previously, use of the colon in the
>> format
>> results
And here is another interesting thing inspired by this thread.
I always thought the "colon-free" versions of the P.E. was a bash-ism, but
experimentation shows that it works in "dash" as well. And "man dash" contains
the following text:
In the parameter expansions shown previously, use of the
>"Note that a negative offset must be separated from the [non-following]
>colon by at least one space to avoid being confused with the ':-'
>expansion,' since a negative offset can exist immediately beside a
>subsequent colon.
>Wiley
Note, incidentally, that it is not strictly true that a negat
This is a "Doctor, it hurts when I do this" type thing.
The short answer is that alias substitution occurs very early, so when you type:
alias x='this and that'
x() { ... }
you are for all practical purposes, typing:
'this and that'() { ... }
At which point, anything can happen. I
(Followup to my previous post)
To answer my own question, I guess it is pretty obvious that the answer to "How
to
fix?" is "Install texinfo and re-run the make install", but it raises two
important
questions:
1) Why is the error ignored, making it almost impossible to determine that
I found the problem (I think). See below:
--- Cut Here ---
make[1]: Entering directory '/home/username/Build/bash-5.3-rc2/doc'
rm -f bashref.info
makeinfo --no-split -I../lib/readline/doc ./bashref.texi
make[1]: makeinfo: Command not found
Makefile:181: recipe for target 'bashref.info' failed
mak
I did: ./configure --prefix=$HOME/local/bash5
Then make and then make install (all of this is done as a non-root user).
It created the directory listed above and all the necessary subdirs, including
share/man/man1, but left that directory empty.
Shouldn't (at least) a copy of "bash.1" have been p
On Fri, Jun 13, 2025 at 07:08:21 -0600, Stan Marsh wrote:
> BTW, and only tangentially related, "man bash" says that "let" and "(( ))" are
> exactly the same, but "shellcheck" thinks otherwise. "shellcheck" says you
> should
> use &qu
>All of this is intentional, and not a bug.
It is possible to be both.
But, yes, it reflects a fundamental inconsistency in the C/Unix ecosystem.
The fact that in most programming languages (e.g., C, AWK), 0 means false and
non-zero
means true, but in the shell, it is the opposite.
E.g., in AWK
>Description:
>The shorthands "++", "--", "+=1" and "-=1" in bash arithmetic
>erroneously return exit code 1 when a variable is either incremented by
>"++" or decremented by "--" from 0 or incremented by "+=1" or decremented
>by "-=1" from -1 and 1 respectively, although the variable valu
On Wed, Jun 11, 2025 at 10:38:17AM -0400, Chet Ramey wrote:
> The business of installing the loadables and
> headers on `make install' came later as the result of feature requests.
>
And Duncan, correctly, replied:
>Fine for `make install' to *install* them. But it flies in the face of
>convention
Somebody flitted:
> Nah
This is not an answer. This is just someone blowing off steam.
=
Please do not send me replies to my posts on the list.
I always read the replies via the web archive, so CC'ing to me is unnec
See below. As far as I can tell, it built OK.
Should I be concerned about the warning?
--- Cut Here ---
rm -f bash
gcc -std=gnu11 -L./builtins -L./lib/readline -L./lib/readline -L./lib/glob
-L./lib/tilde -L./lib/malloc -L./lib/sh -rdynamic -g -O2 -o bash shell.o
eval.o y.tab.o general.o make
>On Thu, Jun 5, 2025, at 8:37 AM, Stan Marsh wrote:
>> Actually, I am not too fond of the habit of having builtins (particularly
>> those supplied as part of the distribution) with the same name as well-known
>> Unix commands.
>
>It allows for potential drop-in replaceme
On Thu, Jun 5, 2025, 12:56â¯PM Greg Wooledge wrote:
> On Thu, Jun 05, 2025 at 12:34:44 +0200, microsuxx wrote:
> > i installed bash dev , but there is no `help cut`
>
> It's a "loadable builtin". You have to build those, install them,
> and then load the "cut" builtin explicitly.
It was far fro
I think most of what needs to be said about the technical aspects of this user's
question has been said. But I want to make a few suggestions for future
practice,
both for this mailing list and for the OP.
In the gawk-bug list (which is structured very much like this list), they have
a setup
wh
(This is all after doing "set -o vi")
In previous versions of bash, I could do the following keys: / ^ foo
and it would find the most recent command that starts with foo. This is all as
expected.
However, in bash 5.0 beta (self-compiled), this doesn't work. It always leaves
the
cursor line e
47 matches
Mail list logo
