usage: do not insist that standard input must come from a filejc/usage-stdin

The synopsys text and the usage string of subcommands that read list
of things from the standard input are often shown like this:

	git gostak [--distim] < <list-of-doshes>

This is problematic in a number of ways:

 * The way to use these commands is more often to feed them the
   output from another command, not feed them from a file.

 * Manual pages outside Git, commands that operate on the data read
   from the standard input, e.g "sort", "grep", "sed", etc., are not
   described with such a "< redirection-from-file" in their synopsys
   text.  Our doing so introduces inconsistency.

 * We do not insist on where the output should go, by saying

	git gostak [--distim] < <list-of-doshes> > <output>

 * As it is our convention to enclose placeholders inside <braket>,
   the redirection operator followed by a placeholder filename
   becomes very hard to read, both in the documentation and in the
   help text.

Let's clean them all up, after making sure that the documentation
clearly describes the modes that take information from the standard
input and what kind of things are expected on the input.

[jc: stole example for fmt-merge-msg from Jonathan]

Helped-by: Jonathan Nieder <jrnieder@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>

1 files changed, 7 insertions, 5 deletions

diff --git a/Documentation/git-get-tar-commit-id.txt b/Documentation/git-get-tar-commit-id.txt
index 1e2a20dd26..ac44d85b0b 100644
--- a/Documentation/git-get-tar-commit-id.txt
+++ b/Documentation/git-get-tar-commit-id.txt
@@ -9,17 +9,19 @@ git-get-tar-commit-id - Extract commit ID from an archive created using git-arch
 SYNOPSIS
 --------
 [verse]
-'git get-tar-commit-id' < <tarfile>
+'git get-tar-commit-id'
 
 
 DESCRIPTION
 -----------
-Acts as a filter, extracting the commit ID stored in archives created by
-'git archive'.  It reads only the first 1024 bytes of input, thus its
-runtime is not influenced by the size of <tarfile> very much.
+
+Read a tar archive created by 'git archive' from the standard input
+and extract the commit ID stored in it.  It reads only the first
+1024 bytes of input, thus its runtime is not influenced by the size
+of the tar archive very much.
 
 If no commit ID is found, 'git get-tar-commit-id' quietly exists with a
-return code of 1.  This can happen if <tarfile> had not been created
+return code of 1.  This can happen if the archive had not been created
 using 'git archive' or if the first parameter of 'git archive' had been
 a tree ID instead of a commit ID or tag.