Talk:Almquist shell

The ash included in *BSD

The ash included in *BSD as /bin/sh does include job control. There is also command line editing though it is not enabled by default.

---

There was another shell named "ash" (or "adsh") which was the "adventure shell", like a text adventure game (Zork or whatever). AnonMoos 18:30, 15 September 2006 (UTC)

---

OpenBSD does not use Almquist shell code. Not for nearly 20 years.

The given reference pointed to code in the OpenBSD CVS that hasn't been touched in that long. I've removed the mention. Placing this note here in case somebody unknowingly tries to put it back in the future.

for further verification: compare http://cvsweb.openbsd.org/cgi-bin/cvsweb/%7echeckout%7e/src/bin/sh/ to http://cvsweb.openbsd.org/cgi-bin/cvsweb/%7echeckout%7e/src/bin/ksh/ (the currently-maintained shell, derived from PDKSH, Public Domain Korn Shell) — Preceding unsigned comment added by 108.192.13.216 (talk) 02:04, 20 December 2016 (UTC)

Comment on Merge Suggestion

The article is pretty clear that it is not describing a Linux-specific program, so merging it would damage this source of information Tedickey 14:24, 8 September 2007 (UTC)

Red Hat nash?

Red Hat/Fedora Linux uses a program called nash [1] to interpret scripts during start-up. However, despite the name it does not appear to be related to ash. I've checked the versions of nash back to summer 2001 without finding any traces of ash — but perhaps they used ash originally, and at some point decided to write their own program instead. Rune Kock (talk) 20:13, 25 March 2008 (UTC)

primary-sources tag

Reading the topic, it appears that the editor has the notion that anyone using the shell, whether they're modifying it or not is automatically a primary source. Some clarification is needed to retain the tag. TEDickey (talk) 18:04, 4 September 2011 (UTC)

Typically people who are (re)distributing a product are not counted as independent third party sources for the purposes of wikipedia notability. Stuartyeates (talk) 20:26, 4 September 2011 (UTC)

Oh. Do you have a pointer to one of the guidelines supporting your opinion? TEDickey (talk) 20:57, 4 September 2011 (UTC)

License?

There is no mention how ash is licensed... --Arny (talk) 19:42, 16 December 2012 (UTC)

Debian's copy (dash) lists a BSD 3-clause license. That's likely the same license in the others, barring some possible modification to use the 2-clause license. TEDickey (talk) 21:33, 16 December 2012 (UTC)

Proposed merge with Debian Almquist shell

The page on dash has major notability problems: the only reference that can be counted as third-party is the one to ubuntu.com. Also, all the significant material on the other page is in fact already present on this one: dash is a minor modification to ash and it's used in Ubuntu and Debian as the boot shell. QVVERTYVS (hm?) 16:39, 27 February 2014 (UTC)

I performed the merge. QVVERTYVS (hm?) 10:26, 5 March 2014 (UTC)

Bash as default shell in Ubuntu

Tedickey wants a better source for the fact that Bash is the default shell on Ubuntu. What kind of source would be appropriate? I referenced the adduser package's default setup file, which very clearly states

# The DSHELL variable specifies the default login shell on your
# system.
DSHELL=/bin/bash

It's a primary source, and it requires a bit of interpretation, but it's the only non-self-published (Ubuntu wiki, Debian wiki, blog, StackExchange, etc.) source that I could find so far. QVVERTYVS (hm?) 07:49, 21 March 2014 (UTC)

That "bit of interpretation" is a problem. A better source for instance would point to one of the Debian policy guidelines, or corresponding mailing-list discussion of the distinctive uses of dash/bash in Debian. (Ubuntu is perhaps harder to since their value-added part is hard to find) TEDickey (talk) 08:14, 21 March 2014 (UTC)

AFAICT, debian-policy only prescribes the behavior of /bin/sh. It mentions bash a few times but not explicitly as the default login shell. QVVERTYVS (hm?) 08:36, 21 March 2014 (UTC)

Well, as it stands, the topic relies upon an inference from a single line of code, providing more information than could be in the given source (looks like WP:SYNTH) TEDickey (talk) 23:01, 21 March 2014 (UTC)

Respectfully, I would accept this. Yes, it's a primary source and maybe there could be better sources, but so what? It's not offered in support of notability nor is there any reason to believe it might be an unreliable source. To the contrary, it's authoritative. I suppose an argument could be made that reading the source and reporting what it says is WP:OR based on the need for the special expertise to interpret the source. But I'm not persuaded. There are millions of people with the necessary programming skill to read this; it's just not that special. Msnicki (talk) 23:18, 21 March 2014 (UTC)

See my comment: the Wikipedia topic provides more information than the source given. It implies a policy; my suggestion of how to find a usable source was glossed as too difficult. TEDickey (talk) 23:58, 21 March 2014 (UTC)

Again, respectively, I disagree. I do not read the claim as suggesting a policy. I read it only as reporting what is. And as I understand it, the statement is correct, that's what it is right now, and the claim can be verified in the source given. Were all the rest of Wikipedia so well supported. Msnicki (talk) 00:58, 22 March 2014 (UTC)

What the article says is "Bash is the default login shell for interactive use" which coincides exactly with what the comment in /usr/share/adduser/adduser.conf says. The necessary inference is that /bin/bash is Bash and that adduser is the command to add users.

The alternative is to cite the Debian and Ubuntu operating systems as such, like we do with e.g. movies for plot summaries. Both are free software (free as in freedom and beer) and thus easier to obtain than many a book, while Ubuntu does not require greater technical skills to install than it takes to read the scientific papers we happily cite in other articles. The procedure for verification boils down to firing up a terminal and checking that the shell is Bash. I don't really like this solution, since it's much more work than checking a single file. QVVERTYVS (hm?) 10:38, 22 March 2014 (UTC)

Sure: it is an inference, as noted here and here, "adduser" is a particular script used to invoke "useradd". As a reminder, nothing in your comments responds to WP:SYNTH. TEDickey (talk) 00:29, 24 March 2014 (UTC)

For the record, I don't think WP:SYNTH applies here: it forbids drawing "a conclusion not explicitly stated by any of the sources", but the desired conclusion that Bash is the default login shell is explicitly stated in the file referenced. I have the "B and C" parts described in WP:SYNTH, just not "A", the proposition that adduser is the default command to add users. But I found a better source, so I'll add that; feel free to remove the old source if you think it's no longer necessary. QVVERTYVS (hm?) 12:49, 24 March 2014 (UTC)

It is only explictly stated by the source which you provided the other day if you gloss over the interpretation. The second source, likewise (does not mention login shell), and I don't see that in the third source. Of course, the third alludes to the change which I pointed out should be sourced (certainly there are better sources than this). There's no point in continuing this discussion. TEDickey (talk) 23:19, 24 March 2014 (UTC)

Portability

At least dash seems to be non-portable as it depends on many GNUisms that are not guaranteed to exist on a typical UNIX platform. Schily (talk) 14:12, 17 August 2015 (UTC)

Bash-specific featuers may not be so

Many features came from ksh88, and so scripts generally do run under ksh derivatives/clones with no modifications. So perhaps the article should say "ksh-specific" features --Ismael Luceno (talk) —Preceding undated comment added 06:55, 18 April 2018 (UTC)

You really have to be careful because there are so many different versions with different lineages. The final true Almquist version was a fairly barebones Bourne shell clone, for a number of years the primary development was as the NetBSD shell which is where most of the other variants seem to have forked from. The NetBSD team made a large number of enhancements to the orginal code but for POSIX compliance rather than ksh compatibility (in fact I tend to informally consider the NetBSD /bin/sh as the reference POSIX shell, it's a lot closer than anything else I've seen). The original POSIX shell specification looked at lot like the traditional Bourne shell, but the revised version (1997?) looked a lot more like ksh88. While many features may have originated in ksh, once they have made it into POSIX they can hardly be regarded as ksh-specific. 3142 (talk) 12:38, 9 January 2019 (UTC)

topicality of test

A recent edit removed a see-also for test (Unix).

Perhaps someone had added that see-also after reading this page TEDickey (talk) 21:49, 26 February 2020 (UTC)