Re: [PATCH] Use "-script" postfix for scripts

Johannes Schindelin [off-list ref] writes:

For consistency reasons, the names of all scripts should end in "-script".

This may be a bit controversial (people might find it unnecessary). 
Subject to discussion.

I have never liked the original -script name convention.  It
only meant that they are implemented as scripts (as opposed to
those on the $(PROG) Makefile variable), but the end users who
end up typing their names from the command line, and to a lesser
degree the people who use them in their scripts, should not care
how they are implemented to begin with.

And to cope with long names and make things look a bit more
familiar to CVS migrants, "git" wrapper was invented to supply
the -script suffix to grok "git whatever", but just in case if
something was _not_ implemented as a script, it ends up needing
to try "git-whatever" in addition to "git-whatever-script".  The
patch alleviates the "git" problem for "git-whatchanged" and
friends whose names currently do not end with -script, but it
still does not help "git apply", for example.  I like the
general direction of making things consistent, but I wonder if
we can become consistent by losing -script suffix, not adding to
the ones that lack it.

And as you imply, this kind of change inevitable breaks people's
scripts.  But I do agree with you that we should do something
about it, so it may be better to break them sooner rather than
later, as long as we make sure we break them just once.

hI,

On Fri, 12 Aug 2005, Junio C Hamano wrote:

Johannes Schindelin [off-list ref] writes:

quoted

For consistency reasons, the names of all scripts should end in "-script".

This may be a bit controversial (people might find it unnecessary). 
Subject to discussion.

I have never liked the original -script name convention.

Okay. I'm all for dropping the "-script" postfix.

Ciao,
Dscho

On Fri, 12 Aug 2005, Junio C Hamano wrote:

I have never liked the original -script name convention.  It
only meant that they are implemented as scripts (as opposed to
those on the $(PROG) Makefile variable), but the end users who
end up typing their names from the command line, and to a lesser
degree the people who use them in their scripts, should not care
how they are implemented to begin with.

Well, end users _don't_ care, since they are supposed to use just a simple
"git xxx".

The advantage with "git-xxx-script" is for git developers: at least yours 
truly does "grep xyz *.c" all the time, and the "grep abc *-script" is 
entirely analogous to that. That's where the "-script" ending comes from: 
it really helps pick out the stuff you can grep from (as opposed to the 
stuff that got compiled and isn't greppable).

Sure, I could have called it ".sh" instead to make it look even more like 
a shell script thing, but I actually think "-script" describes any 
scripting language - shell, perl, you name it..

			Linus

On Fri, Aug 12, 2005 at 02:55:01PM -0700, Linus Torvalds wrote:

Sure, I could have called it ".sh" instead to make it look even more like 
a shell script thing, but I actually think "-script" describes any 
scripting language - shell, perl, you name it..

See, for example, the history on git-rename-script for why this is good.

-- 

Ryan Anderson
  sometimes Pug Majere