[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Nmh-workers] rcvdist with non-default port
From: |
Ralph Corderoy |
Subject: |
Re: [Nmh-workers] rcvdist with non-default port |
Date: |
Sat, 27 Jan 2018 13:31:12 +0000 |
Hi David,
> I just committed that fix to master, please review.
Looking at 28ebf6bb..ff7b97ca, I was puzzled by the macros to make
switches disappear. This doesn't affect your patch; it's a general
nmh-wide issue.
#ifndef OAUTH_SUPPORT
# define OAUTHminc(a) (a)
#else /* OAUTH_SUPPORT */
# define OAUTHminc(a) 0
#endif /* OAUTH_SUPPORT */
X("oauthclientid client-id", OAUTHminc(-12), OAUTHCLIDSW) \
X("oauthclientsecret client-secret", OAUTHminc(-12), OAUTHCLSECSW) \
1---to----12
If we *do* have OAUTH then 0 is set for both options. This is a
reasonable default that means enough characters for each option have to
be specified so there's no ambiguity, and that the option appears in
-help's output. It works:
$ post -help | grep oauthclient
-oauthclientid client-id
-oauthclientsecret client-secret
$ post -oauthclient
post: -oauthclient ambiguous. It matches
-oauthclientid client-id
-oauthclientsecret client-secret
$ post -oauthclientid x -foobar
post: -foobar unknown
$
If we *don't* have OAUTH then -12 is set for both options, meaning at
least twelve characters must be given for each. Given that's how many
are needed to resolve ambiguity then it also distinguishes between the
two, just like 0, but that it's -12 rather than 12 tells -help to skip
printing it. Kind of makes sense; no support so don't show the option.
(Alternatively, show them and that they're disabled at build time.)
$ post -help | grep oauthclient
$ post -oauthclient
post: -oauthclient unknown
$ post -oauthclientid x -foobar
post: -foobar unknown
$
It starts off OK; an ambigious, unsupported, option is unknown. That's
because smatch() doesn't see the twelve characters demanded. But
provide them and the option is happily matched by smatch(), and post.c
does the work of stashing the value for the `unsupported' option. I
think this is a bug.
One last `Huh?'.
$ post '-sasl -user' -foobar
seen -sasl
post: -foobar unknown
$
There's no `seen: -user' because smatch() was only called once with
"-sasl -user" and said it matched the -sasl switch. I think this is
another bug. :-)
--
Cheers, Ralph.
https://plus.google.com/+RalphCorderoy
- Re: [Nmh-workers] rcvdist with non-default port, (continued)
- Re: [Nmh-workers] rcvdist with non-default port, David Levine, 2018/01/15
- Re: [Nmh-workers] rcvdist with non-default port, Ken Hornstein, 2018/01/15
- Re: [Nmh-workers] rcvdist with non-default port, David Levine, 2018/01/15
- Re: [Nmh-workers] rcvdist with non-default port, Ken Hornstein, 2018/01/15
- Re: [Nmh-workers] rcvdist with non-default port, David Levine, 2018/01/15
- Re: [Nmh-workers] rcvdist with non-default port, Steven Winikoff, 2018/01/15
- Re: [Nmh-workers] rcvdist with non-default port, David Levine, 2018/01/15
- Re: [Nmh-workers] rcvdist with non-default port,
Ralph Corderoy <=
- Re: [Nmh-workers] switches and smatch, David Levine, 2018/01/27
- Re: [Nmh-workers] switches and smatch, Ralph Corderoy, 2018/01/27
- Re: [Nmh-workers] switches and smatch, Ken Hornstein, 2018/01/28
- Re: [Nmh-workers] switches and smatch, Paul Vixie, 2018/01/28
- Re: [Nmh-workers] switches and smatch, Ken Hornstein, 2018/01/28
- Re: [Nmh-workers] switches and smatch, Ralph Corderoy, 2018/01/29
- Re: [Nmh-workers] switches and smatch, Bakul Shah, 2018/01/29
- Re: [Nmh-workers] switches and smatch, Ken Hornstein, 2018/01/29
- Re: [Nmh-workers] switches and smatch, Ralph Corderoy, 2018/01/30
- Re: [Nmh-workers] switches and smatch, Bakul Shah, 2018/01/30