On 07/21/13 11:05, Jason McIntyre wrote:
On Sun, Jul 21, 2013 at 09:15:00AM +0200, Alexander Hall wrote:
On 07/21/13 08:11, patrick keshishian wrote:
If so, maybe a better wording would be:
Successful confirmation is signaled by a zero exit status,
and the first line of the program's output SHOULD be either
empty or the string "yes."
Suggesting the following:
Successful confirmation is signaled by a zero exit status,
and the first line of the program's output being either
empty or the string "yes".
Must really the full stop should be within the quotes, even though it's
not part of the quote? If so, and we may not ignore this for the
percieved accuracy of the man pages, I'd suggest we rephrase it so that
it is not the last word.
/Alexander
the reference patrick sent you is, i presume, a US resource. here's
oxford style manual (a uk resource):
Except where the matter is quoted for semantic or bibliographic
scrutiny, the relationship in British practice between
quotation marks and other marks of punctuation is according
to the sense. While the rules are somewhat lengthy to state
in full, the common-sense approach is to do nothing that
changes the meaning of the quotation or renders it confusing
to read.
In US practice, commas and full points are set inside the
closing quotation mark regardless of whether they are part
of the quoted material. The resulting ambiguity can cause
editorial problems when using material from US sources in
British works.
regarding: ... "yes."
for me, it's definitely a case of "renders it confusing to read". is the
string requested "yes." or "yes"? er, "yes?"
your idea to reword it might be best!
note also that your diff introduces a second instance of "signaled".
there's a few in the ssh code too. and one "signalled". that's another
uk/us tussle. it's hard to keep things consistent, and not annoy
others.
regarding your diff... i don't know this stuff well enough to be
able to say whether your moving stuff around makes sense, and whether
you're moving it to the right place. note, for example, that
ssh-agent(1) now documents SSH_ASKPASS, though it does not have an
ENVIRONMENT section. and ssh-add(1), which does, and which documents
SSH_ASKPASS, does not now directly refer to that environment variable.
I noticed that about ssh-agent did not have an ENVIRONMENT section, but
decided to keep style rather than rewriting the lot. The man pages could
probably use a rather thorough overhaul.
are you sure, for example, that just changing -c from "indicates"
to "indicates that ssh-agent(1)" wouldn;t be enough? not saying it
would, just asking. but your diff feels incomplete.
I don't think so. It would still say "Confirmation is performed by the
SSH_ASKPASS program mentioned below." which to me implies that it's the
environment of ssh-add that is being referred to.
/Alexander
jmc
Index: ssh-add.1
===================================================================
RCS file: /cvs/src/usr.bin/ssh/ssh-add.1,v
retrieving revision 1.58
diff -u -p -r1.58 ssh-add.1
--- ssh-add.1 3 Dec 2012 08:33:02 -0000 1.58
+++ ssh-add.1 21 Jul 2013 07:14:13 -0000
@@ -84,14 +84,10 @@ to work.
The options are as follows:
.Bl -tag -width Ds
.It Fl c
-Indicates that added identities should be subject to confirmation before
-being used for authentication.
-Confirmation is performed by the
-.Ev SSH_ASKPASS
-program mentioned below.
-Successful confirmation is signaled by a zero exit status from the
-.Ev SSH_ASKPASS
-program, rather than text entered into the requester.
+Indicates that
+.Xr ssh-agent 1
+should ask for confirmation before added identities are used for
+authentication.
.It Fl D
Deletes all identities from the agent.
.It Fl d
Index: ssh-agent.1
===================================================================
RCS file: /cvs/src/usr.bin/ssh/ssh-agent.1,v
retrieving revision 1.53
diff -u -p -r1.53 ssh-agent.1
--- ssh-agent.1 21 Nov 2010 01:01:13 -0000 1.53
+++ ssh-agent.1 21 Jul 2013 07:14:13 -0000
@@ -161,6 +161,18 @@ Later
.Xr ssh 1
looks at these variables and uses them to establish a connection to the agent.
.Pp
+If confirmation before using a key is requested by
+.Xr ssh-add 1 ,
+it is performed by the program specified in the
+.Ev SSH_ASKPASS
+environment variable, or
+.Xr ssh-askpass 1
+if
+.Ev SSH_ASKPASS
+is not set.
+Successful confirmation is signaled by a zero exit status, and the first line
+of the program's output being either empty or the string "yes".
+.Pp
The agent will never send a private key over its request channel.
Instead, operations that require a private key will be performed
by the agent, and the result will be returned to the requester.
Index: ssh_config.5
===================================================================
RCS file: /cvs/src/usr.bin/ssh/ssh_config.5,v
retrieving revision 1.166
diff -u -p -r1.166 ssh_config.5
--- ssh_config.5 27 Jun 2013 14:05:37 -0000 1.166
+++ ssh_config.5 21 Jul 2013 07:14:13 -0000
@@ -286,7 +286,7 @@ will cause ssh
to listen for control connections, but require confirmation using the
.Ev SSH_ASKPASS
program before they are accepted (see
-.Xr ssh-add 1
+.Xr ssh-agent 1
for details).
If the
.Cm ControlPath
