[EMAIL PROTECTED] (Mark Baker) wrote on 21.11.97 in <[EMAIL PROTECTED]>:
> I don't know whether Ian does, but I do. I find
>
> cp /very/long/path/foo /very/long/path/bar baz
>
> a lot harder to read than
>
> cp /very/long/bath/{foo,bar} baz
>
> And your suggestion of
>
> (cd /ve
> Do you still think that "the use of { } makes the rules files clearer and
I'd just like to request that people stop making the same mistake that
deb-check does... Please abbreviate this construct to {,} but *no less
than that*. It's legitimate, even *necessary* in some contexts, to
use { } in
-BEGIN PGP SIGNED MESSAGE-
Mark Baker wrote:
> [EMAIL PROTECTED] (Santiago Vila Doncel) writes:
>
> > Yes, bash is essential because we always *need* a POSIX shell. But GNU
> > bash provides *two* of them: /bin/sh and /bin/bash. Only /bin/sh should
> > be essential.
>
> However, dangling
[EMAIL PROTECTED] (Mark Baker) writes:
> And your suggestion of
>
> (cd /very/long/path; cp foo bar baz)
To be safe, this should be
(cd /very/long/path && cp foo bar baz)
--
Rob Browning <[EMAIL PROTECTED]>
PGP fingerprint = E8 0E 0D 04 F5 21 A0 94 53 2B 97 F5 D6 4E 39 30
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Santiago Vila Doncel) writes:
> Yes, bash is essential because we always *need* a POSIX shell. But GNU
> bash provides *two* of them: /bin/sh and /bin/bash. Only /bin/sh should
> be essential.
However, dangling symlinks are not terribly
Ian Jackson wrote:
> Santiago Vila Doncel:
> > bash is currently essential because there is no other POSIX
> > shell. Point.
>
> No, that is not the (only) reason why bash is essential. bash is also
> essential because it provides a fixed useful set of facilities for
> people to write init.d scri
Santiago Vila Doncel:
> bash is currently essential because there is no other POSIX
> shell. Point.
No, that is not the (only) reason why bash is essential. bash is also
essential because it provides a fixed useful set of facilities for
people to write init.d scripts, preinst scripts, &c., - ie,
-BEGIN PGP SIGNED MESSAGE-
Ian Jackson wrote:
> What the people who want to make bash nonessential are asking is to
> further restrict the facilities available to such programs, in a way
> that may make life very difficult for people.
"make life very difficult" is an exaggeration.
bash i
Santiago Vila Doncel:
> So if we have to admit bashisms in debian/rules, we are in fact saying
> "Debian packages will always be for Debian/Linux distributions".
and later in another message
> No, we are saying "If you want to port a single Debian package to some
> system, you have to Debianize th
It is clearly necessary for there to be some basic set of
functionality which can be used by packages before others are
available.
We decided very early on to make this set be Perl (without most of the
modules) and bash.
What the people who want to make bash nonessential are asking is to
further
[EMAIL PROTECTED] (Scott Ellis) wrote on 13.11.97 in <[EMAIL PROTECTED]>:
> "Because it is that way now" is NOT necessarily a valid argument for
> keeping things the same way. Slavery used to be common, East Germany used
> to exist. That is not a valid arguement for the continuance of East
> Ge
Hi,
I agree with the statement that any script that uses bashisms
should in fact start with the string #! /bin/bash.
Do we know of any other shell that is POSIX compliant? Bash,
invoked as /bin/sh, has a stated goal of being a POSIX compliant
shell, and I think is quite good at
Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
> No, we are saying "If you want to port a single Debian package to some
> system, you have to Debianize the system first". This is nonsense.
I know this is a small sample, but every system here in the CS dept:
SUNs, SGIs, AIX machines, and HP mach
Hi,
>>"Santiago" == Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
Santiago> -BEGIN PGP SIGNED MESSAGE- Rob Browning wrote:
>> Santiago Vila Doncel <[EMAIL PROTECTED]> writes: So if we have to
>> admit bashisms in debian/rules, we are in fact saying "Debian
>> packages will always be for
Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
> > > So if we have to admit bashisms in debian/rules, we are in fact
> > > saying "Debian packages will always be for Debian/Linux
> > > distributions".
> >
> > No, we are saying that you need bash on any Debian system.
>
> No, we are saying "If
Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
> > How does one satisfy the need for an essential POSIX bourne shell
> > with multiple possibilities?
>
> Having a virtual package tagged as essential.
Say what? How do you tag a virtual package as essential?
> > You, surely, *must* have at le
-BEGIN PGP SIGNED MESSAGE-
Rob Browning wrote:
> Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
> > So if we have to admit bashisms in debian/rules, we are in fact saying
> > "Debian packages will always be for Debian/Linux distributions".
>
> No, we are saying that you need bash on any
-BEGIN PGP SIGNED MESSAGE-
James Troup:
> How does one satisfy the need for an essential POSIX bourne shell with
> multiple possibilities?
Having a virtual package tagged as essential.
> You, surely, *must* have at least one POSIX bourne shell marked as
> essential, otherwise people can
Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
> So if we have to admit bashisms in debian/rules, we are in fact saying
> "Debian packages will always be for Debian/Linux distributions".
No, we are saying that you need bash on any Debian system. Which
systems will bash not build on?
And anoth
Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
[ stuff snipped to get down to the core issue ]
What I want to see answered is this:
How does one satisfy the need for an essential POSIX bourne shell with
multiple possibilities?
You, surely, *must* have at least one POSIX bourne shell marked as
-BEGIN PGP SIGNED MESSAGE-
On 13 Nov 1997, James Troup wrote:
> Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
>
> > > I suggest you file bugs on those which needlessly do.
> >
> > Well, will they be "legitimate" as `wishlist' bugs? Or they will be
> > refused by saying "I don't think
Scott Ellis <[EMAIL PROTECTED]> writes:
> > The point is, bash is already essential, that is Debian policy.
> > The onus is on *you* to demonstrate why that should be changed.
>
> "Because it is that way now" is NOT necessarily a valid argument for
> keeping things the same way.
I never said it
Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
> > I suggest you file bugs on those which needlessly do.
>
> Well, will they be "legitimate" as `wishlist' bugs? Or they will be
> refused by saying "I don't think it is important, since bash is
> essential" and closed immediately?
Herbert has be
On 13 Nov 1997, James Troup wrote:
> Scott Ellis <[EMAIL PROTECTED]> writes:
>
> > Nice post, but very little in the way of rational arguments why bash
> > should be essential when it doesn't have to be.
>
> Cheap shot, not well made. The point is, bash is already essential,
> that is Debian po
Scott Ellis <[EMAIL PROTECTED]> writes:
> Nice post, but very little in the way of rational arguments why bash
> should be essential when it doesn't have to be.
Cheap shot, not well made. The point is, bash is already essential,
that is Debian policy. The onus is on *you* to demonstrate why tha
-BEGIN PGP SIGNED MESSAGE-
On 13 Nov 1997, James Troup wrote:
> Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
>
> > In fact, I am not worried by the fact that bash is essential or not.
> > I am worried by the fact that so many packages depend on it.
>
> I suggest you file bugs on tho
On 13 Nov 1997, James Troup wrote:
[lots snipped]
> What on earth for? That's completely redundant if bash is essential.
Nice post, but very little in the way of rational arguments why bash
should be essential when it doesn't have to be. It's not like there is
significatly more power there than
Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
> > > I want to change the policy. I think bash should not be
> > > essential.
> >
> > Why?
>
> Because this seems to be the only way of encouraging #!/bin/sh in
> favour of #!/bin/bash...
That seems
-BEGIN PGP SIGNED MESSAGE-
> Because this seems to be the only way of encouraging #!/bin/sh
> in favour of #!/bin/bash...
Sorry, "against #!/bin/bash", I meant, of course
-BEGIN PGP SIGNATURE-
Version: 2.6.3ia
Charset: latin1
iQCVAgUBNGsEZiqK7IlOjMLFAQFp9wP5AWq4qDFY+RAb8cKDCmF9U
-BEGIN PGP SIGNED MESSAGE-
> Santiago> I want to change the policy. I think bash should not be
> Santiago> essential.
> Manoj> Why?
Because this seems to be the only way of encouraging #!/bin/sh
in favour of #!/bin/bash...
> Manoj> Do we have any alternatives?
Currently, maybe not.
In
Hi,
>>"Santiago" == Santiago Vila Doncel <[EMAIL PROTECTED]> writes:
Santiago> I want to change the policy. I think bash should not be
Santiago> essential.
Why? Mind you, I'm not being argumentative, I just want to
see technical reasons for and against having a standard, free Bourne
she
-BEGIN PGP SIGNED MESSAGE-
Ian Jackson wrote:
> If you want to change the policy and say that bash shouldn't be
> essential then please come to debian-policy and we'll talk about it.
I want to change the policy. I think bash should not be essential.
> It might be wo
32 matches
Mail list logo
