From: sabrina downard Date: 15:53 on 25 May 2006 Subject: A simple hate today. If you do not have a better use for -h, and you have coded a usage statement for --help, then how damned difficult would it be to make -h equal to --help? Since you're already going to print a damned error telling me to use --help instead? I was clearly already asking for help, you idiots. Taunting me just pisses me off. bitchily, --s.
From: H.Merijn Brand Date: 17:19 on 25 May 2006 Subject: Re: A simple hate today. On Thu, 25 May 2006 09:53:56 -0500 (CDT), sabrina downard <sld@xxxxxxxx.xxx> wrote: > If you do not have a better use for -h, and you have coded a usage > statement for --help, then how damned difficult would it be to make -h > equal to --help? Since you're already going to print a damned error > telling me to use --help instead? Sorry, I disagree. And I have never liked -h to be help. IMHO it should be either -? or --help, and -help could be acceptable > I was clearly already asking for help, For me, -h never has meant help, nor do I expect *any* command or util to show me help on -h > you idiots. Taunting me just pisses me off. > > bitchily, > --s.
From: David Champion Date: 17:30 on 25 May 2006 Subject: Re: A simple hate today. * On 2006.05.25, in <20060525181940.6a7da5ce@pc09>, * "H.Merijn Brand" <h.m.brand@xxxxxx.xx> wrote: > On Thu, 25 May 2006 09:53:56 -0500 (CDT), sabrina downard <sld@xxxxxxxx.xxx> > wrote: > > > If you do not have a better use for -h, and you have coded a usage > > statement for --help, then how damned difficult would it be to make -h > > equal to --help? Since you're already going to print a damned error > > telling me to use --help instead? > > Sorry, I disagree. And I have never liked -h to be help. > IMHO it should be either -? or --help, and -help could be acceptable It's not a matter of taste. It's a matter of UNIX did that for twenty years before GNU came by and made all the kids think different. If you've been using UNIX long and have not completely soaked yourself in GNUisms, you still just automatically expect -h to work almost everywhere, and --help to work in many places (but not by any means most). But the larger point is that an exception occurs (-h is not recognized as an option) which triggers an error message. Given the history of -h, why should that error not be help itself, rather than metahelp? Yes: hate.
From: H.Merijn Brand Date: 17:44 on 25 May 2006 Subject: Re: A simple hate today. On Thu, 25 May 2006 11:30:20 -0500, David Champion <dgc@xxxxxxxx.xxx> wrote: > * On 2006.05.25, in <20060525181940.6a7da5ce@pc09>, > * "H.Merijn Brand" <h.m.brand@xxxxxx.xx> wrote: > > On Thu, 25 May 2006 09:53:56 -0500 (CDT), sabrina downard <sld@xxxxxxxx.xxx> > > wrote: > > > > > If you do not have a better use for -h, and you have coded a usage > > > statement for --help, then how damned difficult would it be to make -h > > > equal to --help? Since you're already going to print a damned error > > > telling me to use --help instead? > > > > Sorry, I disagree. And I have never liked -h to be help. > > IMHO it should be either -? or --help, and -help could be acceptable > > It's not a matter of taste. It's a matter of UNIX did that for twenty > years before GNU came by and made all the kids think different. If > you've been using UNIX long and have not completely soaked yourself > in GNUisms, you still just automatically expect -h to work almost > everywhere, and --help to work in many places (but not by any means > most). Huh? I've been using UNIX since 1982. Long enough? I started with System III, and then got cought in a job that involved writing Unic Device drivers for SLD disks. I've never seen a UNIX command from that time that did not support -? > But the larger point is that an exception occurs (-h is not recognized > as an option) which triggers an error message. Given the history of > -h, why should that error not be help itself, rather than metahelp? Not at all. It should be saved for something useful. > Yes: hate. Yes, to all those that only support info pages, and make their -?, --help, or even -h tell us to read the info pages. info pages are useless. Give me plain man pages please, *AND* a useful --help or -?
From: David Champion Date: 18:11 on 25 May 2006 Subject: Re: A simple hate today. * On 2006.05.25, in <20060525184438.0926fa9d@pc09>, * "H.Merijn Brand" <h.m.brand@xxxxxx.xx> wrote: > > Huh? I've been using UNIX since 1982. Long enough? I started with System III, > and then got cought in a job that involved writing Unic Device drivers for > SLD disks. I've never seen a UNIX command from that time that did not support > -? Many commands do/did -?. Many also do/did -h. I don't care which one anyone thinks is better, just that -h has a record. Maybe -? should still work, too, but it doesn't mean that -h should't. In 1982 on System III machines, how many programs that you recall supported "--help"? > > But the larger point is that an exception occurs (-h is not recognized > > as an option) which triggers an error message. Given the history of > > -h, why should that error not be help itself, rather than metahelp? > > Not at all. It should be saved for something useful. Help isn't useful? That's the single most useful option on any command. It should be easy to get, and "--help" isn't a first choice for many people and it's not supported on a HUGE number of programs. But we're talking about commands that don't use -h for anything else, anyway. > Yes, to all those that only support info pages, and make their -?, --help, or > even -h tell us to read the info pages. > > info pages are useless. Give me plain man pages please, *AND* a useful --help > or -? +1. Also hate on netpbm for not only removing the built-in usage statements and telling us to read the man page instead, but for additionally removing the actual man-page content and pointing us to a web site. unix$ pnmtops -h option `--height' requires an argument unix$ pnmtops --help pnmtops: Use 'man pnmtops' for help. unix$ man pnmtops pnmtops(1) Netpbm pointer man pages pnmtops(1) pnmtops is part of the Netpbm package. Netpbm documentation is kept in HTML format. Please refer to <http://netpbm.sourceforge.net/doc//pnmtops.html>;. If that doesn't work, also try <http://netpbm.sourceforge.net>; and emailing Bryan Henderson, bryanh@xxxxxxxxxxxx.xxx. 30 Mar 2003 Netpbm pnmtops(1) unix$ !! | mailx -s 'Are you responsible for this, you bloody retard?' bryanh@xxxxxxxxxxxx.xxx *seethe* The "net" in "netpbm" never used to mean that you needed a network to use the software.
From: H.Merijn Brand
Date: 18:32 on 25 May 2006
Subject: Re: A simple hate today.
On Thu, 25 May 2006 12:11:09 -0500, David Champion <dgc@xxxxxxxx.xxx> wrote:
> * On 2006.05.25, in <20060525184438.0926fa9d@pc09>,
> * "H.Merijn Brand" <h.m.brand@xxxxxx.xx> wrote:
> >
> > Huh? I've been using UNIX since 1982. Long enough? I started with System III,
> > and then got cought in a job that involved writing Unic Device drivers for
> > SLD disks. I've never seen a UNIX command from that time that did not support
> > -?
>
> Many commands do/did -?. Many also do/did -h. I don't care which one
> anyone thinks is better, just that -h has a record. Maybe -? should
> still work, too, but it doesn't mean that -h should't.
>
> In 1982 on System III machines, how many programs that you recall
> supported "--help"?
None :)
By that time double-dashed options were not yet common.
> > > But the larger point is that an exception occurs (-h is not recognized
> > > as an option) which triggers an error message. Given the history of
> > > -h, why should that error not be help itself, rather than metahelp?
> >
> > Not at all. It should be saved for something useful.
>
> Help isn't useful? That's the single most useful option on any command.
> It should be easy to get, and "--help" isn't a first choice for many
> people and it's not supported on a HUGE number of programs.
>
> But we're talking about commands that don't use -h for anything else,
> anyway.
But we're talking about commands that don't use -h *YET* for
anything else, anyway.
All my perl scripts that are for public consumption or are to stand time
start with something like
--8<---
#!/usr/bin/perl
use strict;
use warnings;
sub usage
{
print STDERR "usage: $0 [-i infile] [-o outfile]\n";
exit;
} # usage
@ARGV == 1 and $ARGV [0] eq "-?" || $ARGV[0] =~ m/^-+help$/ and usage;
:
-->8---
even if I plan to use Getopt::Long later on.
> > Yes, to all those that only support info pages, and make their -?, --help, or
> > even -h tell us to read the info pages.
> >
> > info pages are useless. Give me plain man pages please, *AND* a useful --help
> > or -?
>
> +1.
>
> Also hate on netpbm for not only removing the built-in usage statements
> and telling us to read the man page instead, but for additionally
> removing the actual man-page content and pointing us to a web site.
>
> unix$ pnmtops -h
> option `--height' requires an argument
>
> unix$ pnmtops --help
> pnmtops: Use 'man pnmtops' for help.
>
> unix$ man pnmtops
> pnmtops(1) Netpbm pointer man pages pnmtops(1)
>
> pnmtops is part of the Netpbm package. Netpbm documentation is kept in HTML
> format.
>
> Please refer to <http://netpbm.sourceforge.net/doc//pnmtops.html>;.
>
> If that doesn't work, also try <http://netpbm.sourceforge.net>; and emailing
> Bryan Henderson, bryanh@xxxxxxxxxxxx.xxx.
>
> 30 Mar 2003 Netpbm pnmtops(1)
>
> unix$ !! | mailx -s 'Are you responsible for this, you bloody retard?' bryanh@xxxxxxxxxxxx.xxx
>
> *seethe* The "net" in "netpbm" never used to mean that you needed a
> network to use the software.
This example is worth a separate HATE on itself !
From: jrodman
Date: 19:04 on 25 May 2006
Subject: Re: A simple hate today.
On Thu, May 25, 2006 at 06:44:38PM +0200, H.Merijn Brand wrote:
> On Thu, 25 May 2006 11:30:20 -0500, David Champion <dgc@xxxxxxxx.xxx> wrote:
>
> > * On 2006.05.25, in <20060525181940.6a7da5ce@pc09>,
> > * "H.Merijn Brand" <h.m.brand@xxxxxx.xx> wrote:
> > > On Thu, 25 May 2006 09:53:56 -0500 (CDT), sabrina downard <sld@xxxxxxxx.xxx>
> > > wrote:
> > >
> > > > If you do not have a better use for -h, and you have coded a
> > > > usage statement for --help, then how damned difficult would it
> > > > be to make -h equal to --help? Since you're already going to
> > > > print a damned error telling me to use --help instead?
> > >
> > > Sorry, I disagree. And I have never liked -h to be help.
> > > IMHO it should be either -? or --help, and -help could be acceptable
> >
> > It's not a matter of taste. It's a matter of UNIX did that for
> > twenty years before GNU came by and made all the kids think
> > different. If you've been using UNIX long and have not completely
> > soaked yourself in GNUisms, you still just automatically expect -h
> > to work almost everywhere, and --help to work in many places (but
> > not by any means most).
>
> Huh? I've been using UNIX since 1982. Long enough? I started with
> System III, and then got cought in a job that involved writing Unic
> Device drivers for SLD disks. I've never seen a UNIX command from that
> time that did not support -?
jrodman@Skonnos:~ >ls -?
ls: invalid option -- ?
Try `ls --help' for more information.
There we go. No -?, and the hateful taunt.
In my experience -? was more common in dos environents which I used
sometimes, so when I want help these days I tend to type
blah -h
blah -?
blah --help
while only barely glancing at the blah output to see if it might be
helpful.
> info pages are useless. Give me plain man pages please, *AND* a useful --help
> or -?
I know this is the hate list, and typical info content is rather
hateful, but consider installing 'pinfo', it's somewhat less hateful to
use than 'info', at least I can reliably quit it without ^Z to suspend
and kill %2. When a tool is complex enough that I actually need a bit
more of a manual and less of an options listing, sometimes this stuff is
useful.
-josh
From: peter (Peter da Silva) Date: 19:46 on 25 May 2006 Subject: Re: A simple hate today. > jrodman@Skonnos:~ >ls -? > ls: invalid option -- ? > Try `ls --help' for more information. % ls -? No match. % ls -\? usage: ls [-ABCFGHLPRSTWZabcdfghiklnoqrstuvx1] [file ...] THAT is what SHOULD happen.
From: H.Merijn Brand Date: 20:21 on 25 May 2006 Subject: Re: A simple hate today. On Thu, 25 May 2006 11:04:02 -0700, jrodman@xxxx.xxxxxxxxxx.xxx wrote: > On Thu, May 25, 2006 at 06:44:38PM +0200, H.Merijn Brand wrote: > > On Thu, 25 May 2006 11:30:20 -0500, David Champion <dgc@xxxxxxxx.xxx> wrote: > > > > > * On 2006.05.25, in <20060525181940.6a7da5ce@pc09>, > > > * "H.Merijn Brand" <h.m.brand@xxxxxx.xx> wrote: > > > > On Thu, 25 May 2006 09:53:56 -0500 (CDT), sabrina downard <sld@xxxxxxxx.xxx> > > > > wrote: > > > > > > > > > If you do not have a better use for -h, and you have coded a > > > > > usage statement for --help, then how damned difficult would it > > > > > be to make -h equal to --help? Since you're already going to > > > > > print a damned error telling me to use --help instead? > > > > > > > > Sorry, I disagree. And I have never liked -h to be help. > > > > IMHO it should be either -? or --help, and -help could be acceptable > > > > > > It's not a matter of taste. It's a matter of UNIX did that for > > > twenty years before GNU came by and made all the kids think > > > different. If you've been using UNIX long and have not completely > > > soaked yourself in GNUisms, you still just automatically expect -h > > > to work almost everywhere, and --help to work in many places (but > > > not by any means most). > > > > Huh? I've been using UNIX since 1982. Long enough? I started with > > System III, and then got cought in a job that involved writing Unic > > Device drivers for SLD disks. I've never seen a UNIX command from that > > time that did not support -? > > jrodman@Skonnos:~ >ls -? > ls: invalid option -- ? > Try `ls --help' for more information. > > There we go. No -?, and the hateful taunt. > > In my experience -? was more common in dos environents which I used > sometimes, so when I want help these days I tend to type > > blah -h > blah -? > blah --help > > while only barely glancing at the blah output to see if it might be > helpful. > > > info pages are useless. Give me plain man pages please, *AND* a useful --help > > or -? > > I know this is the hate list, and typical info content is rather > hateful, but consider installing 'pinfo', it's somewhat less hateful to > use than 'info', at least I can reliably quit it without ^Z to suspend > and kill %2. When a tool is complex enough that I actually need a bit > more of a manual and less of an options listing, sometimes this stuff is > useful. I'm always so annoyed by anything that looks like info, that I decided long ago to just throw it away after install. I have no info dir's anymore. If the man pages are not enough, I'm faster in finding what I need to know through google or other internet resources than when I would need to find it in info. I've tried most of the info clones, pinfo, info, tkinfo, and probably some more, but its format and ways to go through the docs is so revolting that I still just rather throw it away.
From: A. Pagaltzis Date: 04:43 on 26 May 2006 Subject: Re: A simple hate today. * jrodman@xxxx.xxxxxxxxxx.xxx <jrodman@xxxx.xxxxxxxxxx.xxx> [2006-05-25 20:05]: > I know this is the hate list, and typical info content is > rather hateful, but consider installing 'pinfo', it's somewhat > less hateful to use than 'info', at least I can reliably quit > it without ^Z to suspend and kill %2. http://www.vim.org/scripts/script.php?script_id=21 Or Emacs' info mode, I suppose, if you're of that persuasion. But the only info manual I've ever found useful is the one for GNU make. Regards,
From: peter (Peter da Silva) Date: 19:40 on 25 May 2006 Subject: Re: A simple hate today. > Huh? I've been using UNIX since 1982. Long enough? I started with System III, > and then got cought in a job that involved writing Unic Device drivers for > SLD disks. I've never seen a UNIX command from that time that did not support > -? Really? I've never used a UNIX system where that worked. And I started in the '70s with 5th Edition. "-?" is a MSDOS-ism. % foo -? No match. % got a light?
From: Aaron J. Grier Date: 00:24 on 28 May 2006 Subject: Re: A simple hate today. On Thu, May 25, 2006 at 01:40:39PM -0500, Peter da Silva wrote: > "-?" is a MSDOS-ism. you mean of course "/?" :)
From: Peter da Silva Date: 01:31 on 28 May 2006 Subject: Re: A simple hate today. On May 27, 2006, at 6:24 PM, Aaron J. Grier wrote: > On Thu, May 25, 2006 at 01:40:39PM -0500, Peter da Silva wrote: >> "-?" is a MSDOS-ism. > you mean of course "/?" :) From MS-DOS 2.11 through MS-DOS 5 there was a variable "SWITCHAR". If it was set to "/" (the default) the switch character was "/" and the path element separator was "\". If it was set to "-" then the switch character was "-" and the path element separator was "/".
From: spc (Sean 'Captain Napalm' Conner) Date: 08:28 on 29 May 2006 Subject: Re: A simple hate today. It was thus said that the Great Peter da Silva once stated: > > On May 27, 2006, at 6:24 PM, Aaron J. Grier wrote: > > On Thu, May 25, 2006 at 01:40:39PM -0500, Peter da Silva wrote: > >> "-?" is a MSDOS-ism. > > > you mean of course "/?" :) > > From MS-DOS 2.11 through MS-DOS 5 there was a variable "SWITCHAR". If > it was set to "/" (the default) the switch character was "/" and the > path element separator was "\". If it was set to "-" then the switch > character was "-" and the path element separator was "/". And internally (not COMMAND.COM, but in the kernel [1] itself) MS-DOS supported both '/' and '\' as a path separator. I'm not sure how many programs supported SWITCHAR [2], but I know that once I learned about it, my programs from then on checked. -spc (Gah! I still remember this stuff! Aaah!) [1] As much as a single tasking, non-reentrant interrupt controller could be called a "kernel". [2] There was semi-officially-documented system call to set and get the current setting of SWITCHAR.
From: David Cantrell Date: 17:46 on 25 May 2006 Subject: Re: A simple hate today. On Thu, May 25, 2006 at 11:30:20AM -0500, David Champion wrote: > But the larger point is that an exception occurs (-h is not recognized > as an option) which triggers an error message. Given the history of > -h, why should that error not be help itself, rather than metahelp? OTOH, you really want to know what the problem was, so something like this is needed first: ERROR: -h not recognised and you really don't want that to scroll off the top of the screen when you then spew the help text, because the user isn't expecting that. Yes, he could scroll back, or run the command again and capture its output (but I bet he'd first try capturing stdout, not stderr, so have to do it *again* when he realises his second mistake), but this is extra work, and you want your users to do minimal work.
From: David Champion Date: 17:57 on 25 May 2006 Subject: Re: A simple hate today. * On 2006.05.25, in <20060525164600.GC9327@xxxxxxxx.xxxxxxxx.xx.xx>, * "David Cantrell" <david@xxxxxxxx.xxx.xx> wrote: > > OTOH, you really want to know what the problem was, so something like > this is needed first: > > ERROR: -h not recognised > > and you really don't want that to scroll off the top of the screen when > you then spew the help text, because the user isn't expecting that. $ tool -h usage: tool [--this] [--that] [args ...] ... Lots of superfluous information that shouldn't really be in a help ... ... text at all (this is a manual's job) and lasts for 29 screen pages.... -h: option not recognized. $
From: peter (Peter da Silva) Date: 19:38 on 25 May 2006 Subject: Re: A simple hate today. > But the larger point is that an exception occurs (-h is not recognized > as an option) which triggers an error message. Given the history of > -h, why should that error not be help itself, rather than metahelp? A larger point is that the error message should be helpful no matter WHAT the option is. What REALLY takes the cake is when "--help" isn't helpful. The other day I got something like this... % foo --help You idiot, you need to type "foo --help commands" or "foo --help syntax". Not literally, but that's the subtext. HATE
From: Jarkko Hietaniemi
Date: 19:51 on 25 May 2006
Subject: Re: A simple hate today.
Peter da Silva wrote:
>> But the larger point is that an exception occurs (-h is not recognized
>> as an option) which triggers an error message. Given the history of
>> -h, why should that error not be help itself, rather than metahelp?
>
> A larger point is that the error message should be helpful no matter WHAT
> the option is. What REALLY takes the cake is when "--help" isn't helpful.
>
> The other day I got something like this...
>
> % foo --help
> You idiot, you need to type "foo --help commands" or "foo --help syntax".
>
> Not literally, but that's the subtext.
HA! Do not try to protect the guilty. cvs, I hate thee.
$ cvs -h
cvs: invalid option -- h
Usage: cvs [cvs-options] command [command-options-and-arguments]
where cvs-options are -q, -n, etc.
(specify --help-options for a list of options)
where command is add, admin, etc.
(specify --help-commands for a list of commands
or --help-synonyms for a list of command synonyms)
where command-options-and-arguments depend on the specific command
(specify -H followed by a command name for command-specific help)
Specify --help to receive this message
The Concurrent Versions System (CVS) is a tool for version control.
For CVS updates and additional information, see
the CVS home page at http://www.cvshome.org/ or
Pascal Molli's CVS site at http://www.loria.fr/~molli/cvs-index.html
> HATE
>
>
From: Phil Pennock Date: 13:13 on 26 May 2006 Subject: Re: A simple hate today. On 2006-05-25 at 11:30 -0500, David Champion wrote: > But the larger point is that an exception occurs (-h is not recognized > as an option) which triggers an error message. Given the history of > -h, why should that error not be help itself, rather than metahelp? Because if the program has to deal much with files and there's ever a risk of dealing differently with symlinks, you're looking at the BSDish semi-standardised options for dealing with that and having to learn "oops, no, this program uses that other letter instead" is a bitch. I'm not defending the choice of '-h' for this, merely pointing it out; a sometimes surprising number of commands end up needing special-casing for symlinks, which suggests a less than ideal design in and of itself. "-h" means "act on a symlink itself, rather than the file pointed to". "-H" means "follow symlinks pointing to directories, when traversing". etc etc. See chmod, chown, ln, etc etc. :^(
From: David Champion Date: 16:31 on 26 May 2006 Subject: Re: A simple hate today. * On 2006.05.26, in <20060526121357.GA6056@xxxxxxxxx.xxxxxxx.xxx>, * "Phil Pennock" <phil.pennock@xxxxxxx.xxx> wrote: > > "-h" means "act on a symlink itself, rather than the file pointed to". > "-H" means "follow symlinks pointing to directories, when traversing". > etc etc. Sure. It also means "show me data quantities in pretty MG and GB and TB values instead of large KB values", and it means "hostname" in a lot of programs. There are plenty of cases where it can't be used for help now, but it doesn't all add up to "-h should never key in help."
From: Michael Ahlers Date: 16:45 on 26 May 2006 Subject: Re: A simple hate today. As this debate has raged on for nearly a week (or more?) now, I think it is time we propose a new approach to resolve once and for all this "help switch dilemma." Common shells should recognize the event of a particular key sequence, such as "F1" and trigger a standardized function which exists in the context. That is, a blank command line will trigger a helpMessage function in the shell, and if a command is present, the helpMessage function will be invoked for the given binary. That function may choose to print a help message or load a manual. If command line arguments are present, it may provide specific details based on some strategy, such as which argument does the cursor currently rest on or maybe the last one on the list. If this hypothetical helpMessage function is not implemented, such as for legacy binaries, the shell could look up details from a web service which connects to a known database of manuals or other information. If that service is not available, it could simply query Google with the binary name and any command line arguments the user supplied, along with details about the platform and distribution. In the latter case, the user would be asked if they are feeling lucky. If not, a list of results is displayed in the shell which the user can select to display in a web browser. Should the user not find any relevant information from the query, a service "helpd" will then spam any related newsgroups with a "help, what does this do?" message, including the parameters of the Google query. The helpd process will then poll the newsgroups for responses to the thread and, ideally, use some spam filtering mechanism to block flames and requests the user RTFM (which, ironically, does not exist provided we have gotten this far) while allowing real responses through. Once received, these responses should be shuttled to the mailbox of the user making the request or sent as SMS messages to cell phone, et cetera. If all of this does not help, the user could try invoking the command with a -h, --help flag to get the arguments. Satirically yours
From: Tom Duff Date: 17:46 on 26 May 2006 Subject: Re: A simple hate today. On Fri, 26 May 2006, Michael Ahlers wrote: > As this debate has raged on for nearly a week (or more?) now, I think it > is time we propose a new approach to resolve once and for all this "help > switch dilemma." The way help should work is: At the shell prompt, prepend "man " to the name of the command; hit return; read the help. That's how it worked on every BTL UNIX, thanks mainly to Doug McIlroy. It's hateful that it doesn't work today.
From: Martin Ebourne Date: 17:54 on 26 May 2006 Subject: Re: A simple hate today. On Fri, 2006-05-26 at 09:46 -0700, Tom Duff wrote: > The way help should work is: > > At the shell prompt, prepend "man " to the name of the command; > hit return; > read the help. > > That's how it worked on every BTL UNIX, > thanks mainly to Doug McIlroy. It's > hateful that it doesn't work today. That's how zsh does it. (Well it prepends run-help which defaults to man but is user replaceable.) The zsh help key is meta-H, though you could bind it to F1 if you really wanted. These things are not exactly difficult eh? Even gives you your partial command back to edit after you've finished with man. Cheers, Martin.
From: Phil Pennock Date: 16:52 on 26 May 2006 Subject: Re: A simple hate today. On 2006-05-26 at 10:31 -0500, David Champion wrote: > Sure. It also means "show me data quantities in pretty MG and GB and > TB values instead of large KB values", and it means "hostname" in a lot > of programs. There are plenty of cases where it can't be used for help > now, but it doesn't all add up to "-h should never key in help." No, but it does argue against having a default action produce desirable behaviour, where that action is subject to implementation doing something else, resulting in an incompatible change. Semi-standardised sets of meanings which have claimed that letter merely make this more likely. If a program spews a short synopsis to stderr, and refers to the real help option, then exits non-zero, for all unimplemented options, whilst producing --help/whatever to stdout exiting zero, you have something which makes it clear that "this option isn't in use, see over there for help" and is much more amenable to later implementing Real Functionality under that option letter. But that's the behaviour which the OP hated. I'd just hate it if things did as the OP likes things. Each to their own.
From: David Champion Date: 17:44 on 26 May 2006 Subject: Re: A simple hate today. * On 2006.05.26, in <20060526155250.GA30730@xxxxxxxxx.xxxxxxx.xxx>, * "Phil Pennock" <phil.pennock@xxxxxxx.xxx> wrote: > > No, but it does argue against having a default action produce desirable > behaviour, where that action is subject to implementation doing OK, but... > But that's the behaviour which the OP hated. I'd just hate it if things > did as the OP likes things. Each to their own. I think you're conflating the OP with someone else, FTR. Suggesting that all unrecognized options map onto --help one to one was someone else's contribution. Not that this matters; your point is your point, no matter who you're disagreeing with. Myself, I like something along the lines of "-@ is an error. here is short usage text." But I never put long usage information in my built-in help anyway; that belongs in a manual, not in the last three screensful of my terminal window. In this arrangement, both interests are met. It's clear what is not a known option, and it's actually helpful rather than meta-helpful. The OP asked that -h be made a published interface to help, absent any particular reason not to. So what's a particular reason not to? I say: not much. Myself, I always make -h produce help up front. (And sometimes also --help, even though I rarely use --godawful-long-options.) It's easy; it makes sense; it can be anticipated by a broad set of users. I'm not concerned with BSD- or GNU-similitude should the program one day present alternative behaviors with respect to symbolic links or file sizes. My program will never be a BSD program or a GNU program, so why should I? And why should "use lstat/lchown/lwhatever" take precedence over "help, please" anyway? If you need lsomething(), you're familiar enough to find out how to do it, or to remember. If you need help, you know little to nothing. Help should be the easiest thing to get.
From: Phil Pennock Date: 17:48 on 26 May 2006 Subject: Re: A simple hate today. On 2006-05-26 at 11:44 -0500, David Champion wrote: > The OP asked that -h be made a published interface to help, absent any > particular reason not to. So what's a particular reason not to? I say: > not much. That's the one I was remembering, and presenting an argument against (as a specific instance). However: > And why should "use lstat/lchown/lwhatever" take precedence over "help, > please" anyway? If you need lsomething(), you're familiar enough to > find out how to do it, or to remember. If you need help, you know > little to nothing. Help should be the easiest thing to get. Good point; I concede -- I was wrong. OP: sorry David: thanks for the clue.
From: A. Pagaltzis Date: 18:04 on 26 May 2006 Subject: Re: A simple hate today. * David Champion <dgc@xxxxxxxx.xxx> [2006-05-26 18:50]: > And sometimes also --help, even though I rarely use > --godawful-long-options. I provide long synonyms for all short options. (Getopt::Long makes this very easy.) OT1H it helps self-document the code, OTOH it's also nice for the users. Short options are much nicer when working interactively with the shell, but I find the long forms preferrable when I want to use an obscure switch in a script. Much easier to decipher `xargs -0 --verbose --no-run-if-empty` five months down the road compared to `xargs -r0t` if I never use these options. (These are rhetoric examples. I use these switches a lot. For some hateful reason, xargs assigns `-t` for `--verbose` even though it does not use `-v` for anything, and for an more hateful reason, it will run the given command by default even when its input is empty, which is *never* *ever* what I want. That this combination of switches comes out as `-r0t` is quite apt. Hate.) Regards,
From: A. Pagaltzis Date: 17:51 on 26 May 2006 Subject: Re: A simple hate today. * Phil Pennock <phil.pennock@xxxxxxx.xxx> [2006-05-26 17:55]: > If a program spews a short synopsis to stderr, and refers to > the real help option, then exits non-zero, for all > unimplemented options, whilst producing --help/whatever to > stdout exiting zero, you have something which makes it clear > that "this option isn't in use, see over there for help" and is > much more amenable to later implementing Real Functionality > under that option letter. There is no reason not to produce a short synopsis of the far most common options along with that message, no? And heck, you could even spew the entire help message to stderr for `-h`, followed by "unrecognized option, use <whatever>" and a non-zero exit code, if you want to keep it free for later use. Regards,
From: Martin Ebourne Date: 18:05 on 25 May 2006 Subject: Re: A simple hate today. "H.Merijn Brand" <h.m.brand@xxxxxx.xx> wrote: > Sorry, I disagree. And I have never liked -h to be help. > IMHO it should be either -? or --help, and -help could be acceptable -help can't work or you'd never be able to combine options, which is a long standing unixism for lazy people. (I'm lazy, I like it.) -? is nasty. Not only does it look ugly but it's not even a letter. Furthermore I don't want a help option I have to escape to protect it from my shell, that's not helpful. I'm with the original poster. I try -h then --help. If that doesn't work I wouldn't even think to use -?, I'd be onto man and strings already. Cheers, Martin.
From: Philip Newton Date: 18:44 on 25 May 2006 Subject: Re: A simple hate today. On 5/25/06, Martin Ebourne <lists@xxxxxxx.xx.xx> wrote: > -? is nasty. Not only does it look ugly but it's not even a letter. > Furthermore I don't want a help option I have to escape to protect it > from my shell, that's not helpful. You have to escape it from your shell? Either you have files whose name is simply hyphen plus another character in the current directory, or you have a hateful shell. Cheers, --=20 Philip Newton <philip.newton@xxxxx.xxx>
From: Martin Ebourne Date: 19:08 on 25 May 2006 Subject: Re: A simple hate today. Philip Newton <philip.newton@xxxxx.xxx> wrote: > On 5/25/06, Martin Ebourne <lists@xxxxxxx.xx.xx> wrote: >> -? is nasty. Not only does it look ugly but it's not even a letter. > > You have to escape it from your shell? Either you have files whose > name is simply hyphen plus another character in the current directory, > or you have a hateful shell. % diff -? zsh: no match % There are things to hate about zsh for sure, but I've yet to try a =20 shell which has a shorter list of things to hate. ? is a pattern character as you know. And I have zsh to abort a =20 command if I enter a pattern that doesn't match anything. This is a =20 good thing, because if I entered a pattern I sure expect it to match, =20 so if it fails I don't want the command to run because it'll probably =20 only go off and do something stupid. As you say, if you have any file with a 2 character name starting with =20 a hyphen then all of a sudden -? doesn't work as a help option any =20 more, in any shell. That isn't hateful, it's downright vile. Options should be letters, not symbols. Cheers, Martin.
From: H.Merijn Brand Date: 20:14 on 25 May 2006 Subject: Re: A simple hate today. On Thu, 25 May 2006 19:08:17 +0100, Martin Ebourne <lists@xxxxxxx.xx.xx> wrote: > Philip Newton <philip.newton@xxxxx.xxx> wrote: > > On 5/25/06, Martin Ebourne <lists@xxxxxxx.xx.xx> wrote: > >> -? is nasty. Not only does it look ugly but it's not even a letter. > > > > You have to escape it from your shell? Either you have files whose > > name is simply hyphen plus another character in the current directory, > > or you have a hateful shell. > > % diff -? > zsh: no match > % (tcsh) % diff -? diff: invalid option -- ? diff: Try `diff --help' for more information. Exit 2 Bach to the original hate. HERE I would agree with Sabrina. Djeez, I'm asking for help, just give it. > There are things to hate about zsh for sure, but I've yet to try a > shell which has a shorter list of things to hate. > > ? is a pattern character as you know. And I have zsh to abort a > command if I enter a pattern that doesn't match anything. This is a > good thing, because if I entered a pattern I sure expect it to match, > so if it fails I don't want the command to run because it'll probably > only go off and do something stupid. > > As you say, if you have any file with a 2 character name starting with > a hyphen then all of a sudden -? doesn't work as a help option any > more, in any shell. That isn't hateful, it's downright vile. > > Options should be letters, not symbols. Nah, we should allow unicode too. And I stick to my opinion that -? is asking for help. That's how I grew up in Unix land.
From: Juerd
Date: 20:16 on 25 May 2006
Subject: Re: A simple hate today.
H.Merijn Brand skribis 2006-05-25 21:14 (+0200):
> % diff -?
> diff: invalid option -- ?
> diff: Try `diff --help' for more information.
> Exit 2
% touch -- -@
% diff -?
diff: invalid option -- @
diff: Try `diff --help' for more information.
Exit 2
:)
By the way:
juerd@nano:/usr/bin$ tcsh -?
Unknown option: `-?'
Usage: tcsh [ -bcdefilmnqstvVxX ] [ argument ... ].
Juerd
From: H.Merijn Brand Date: 16:32 on 09 Jun 2006 Subject: Re: A simple hate today. On Thu, 25 May 2006 21:16:16 +0200, Juerd <juerd@xxxxxxxxxxx.xx> wrote: > H.Merijn Brand skribis 2006-05-25 21:14 (+0200): > > % diff -? > > diff: invalid option -- ? > > diff: Try `diff --help' for more information. > > Exit 2 > > % touch -- -@ > % diff -? > diff: invalid option -- @ > diff: Try `diff --help' for more information. > Exit 2 These folk know how to please us: % pdftotext --help pdftotext version 3.00 Copyright 1996-2004 Glyph & Cog, LLC Usage: pdftotext [options] <PDF-file> [<text-file>] -f <int> : first page to convert -l <int> : last page to convert -layout : maintain original physical layout -raw : keep strings in content stream order -htmlmeta : generate a simple HTML file, including the meta information -enc <string> : output text encoding name -eol <string> : output end-of-line convention (unix, dos, or mac) -nopgbrk : don't insert page breaks between pages -opw <string> : owner password (for encrypted files) -upw <string> : user password (for encrypted files) -q : don't print any messages or errors -cfg <string> : configuration file to use in place of .xpdfrc -v : print copyright and version info -h : print usage information -help : print usage information --help : print usage information -? : print usage information
From: Chris Devers Date: 20:28 on 25 May 2006 Subject: Re: A simple hate today. On 25 May 2006, at 3:14 PM, H.Merijn Brand wrote: > Nah, we should allow unicode too. > And I stick to my opinion that -? is asking for help. > That's how I grew up in Unix land. There's a simple solution to this, of course. For all defined single-letter options, handle them appropriately. For all undefined single-letter options, return with help details. -h wouldn't have to be defined or special-cased, because the catchall would catch it. Likewise with -?, and everything else for that matter. Unicode, too! Good software should be helpful wherever possible, not playing some childish call & response game of "no, this is the secret handshake, no wait that is the secret handshake." But then, I've about given up on finding such pragmatic tools...
From: Juerd
Date: 20:37 on 25 May 2006
Subject: Re: A simple hate today.
Chris Devers skribis 2006-05-25 15:28 (-0400):
> There's a simple solution to this, of course.
> For all defined single-letter options, handle them appropriately.
> For all undefined single-letter options, return with help details.
-? is still dangerous:
touch -- -r
foo -?
Now, -? might resolve to -r, which in foo's case, means foo tries to
delete every file in your home directory. Too bad.
Juerd
From: H.Merijn Brand Date: 20:46 on 25 May 2006 Subject: Re: A simple hate today. On Thu, 25 May 2006 21:37:55 +0200, Juerd <juerd@xxxxxxxxxxx.xx> wrote: > Chris Devers skribis 2006-05-25 15:28 (-0400): > > There's a simple solution to this, of course. > > For all defined single-letter options, handle them appropriately. > > For all undefined single-letter options, return with help details. > > -? is still dangerous: > > touch -- -r > foo -? Your fault, core dumped. > Now, -? might resolve to -r, which in foo's case, means foo tries to > delete every file in your home directory. Too bad. That is unix. You asked for it, you got it. That is why many shells have aliasses and completions. rm -i is the default for a lot of users
From: Juerd Date: 20:48 on 25 May 2006 Subject: Re: A simple hate today. H.Merijn Brand skribis 2006-05-25 21:46 (+0200): > > -? is still dangerous: > > touch -- -r > > foo -? > Your fault, core dumped. Well, imagine "touch -- -r" was actually "bar /etc/passwd", which is closed source and left a file called "-r" in the cwd. Do you encourage running "ls" before using "-?", for security reasons? > > Now, -? might resolve to -r, which in foo's case, means foo tries to > > delete every file in your home directory. Too bad. > That is unix. You asked for it, you got it. > That is why many shells have aliasses and completions. > rm -i is the default for a lot of users An alias for rm won't help against a direct unlink call. Juerd
From: H.Merijn Brand Date: 20:58 on 25 May 2006 Subject: Re: A simple hate today. On Thu, 25 May 2006 21:48:32 +0200, Juerd <juerd@xxxxxxxxxxx.xx> wrote: > H.Merijn Brand skribis 2006-05-25 21:46 (+0200): > > > -? is still dangerous: > > > touch -- -r > > > foo -? > > Your fault, core dumped. > > Well, imagine "touch -- -r" was actually "bar /etc/passwd", which is > closed source and left a file called "-r" in the cwd. Sew them! Sew them! Priority 1 security problem :) > Do you encourage running "ls" before using "-?", for security reasons? I can't count the number of times I do 'l' (an alias for 'ls -a') > > > Now, -? might resolve to -r, which in foo's case, means foo tries to > > > delete every file in your home directory. Too bad. > > That is unix. You asked for it, you got it. > > That is why many shells have aliasses and completions. > > rm -i is the default for a lot of users > > An alias for rm won't help against a direct unlink call. true
From: A. Pagaltzis Date: 04:54 on 26 May 2006 Subject: Re: A simple hate today. * H.Merijn Brand <h.m.brand@xxxxxx.xx> [2006-05-25 21:50]: > rm -i is the default for a lot of users That is hateful. Get used to it, then run into a system where `-i` is not the default. Oops. Regards,
From: jrodman
Date: 11:35 on 26 May 2006
Subject: Re: A simple hate today.
On Thu, May 25, 2006 at 09:46:46PM +0200, H.Merijn Brand wrote:
> On Thu, 25 May 2006 21:37:55 +0200, Juerd <juerd@xxxxxxxxxxx.xx> wrote:
>
> > Chris Devers skribis 2006-05-25 15:28 (-0400):
> > > There's a simple solution to this, of course.
> > > For all defined single-letter options, handle them appropriately.
> > > For all undefined single-letter options, return with help details.
> >
> > -? is still dangerous:
> >
> > touch -- -r
> > foo -?
>
> Your fault, core dumped.
Here, let me plant a bomb on our shared computer:
touch /tmp/-r
It may take a long time to go off, but if it does, I guess that's your
fault too?
From: H.Merijn Brand Date: 13:04 on 26 May 2006 Subject: Re: A simple hate today. On Fri, 26 May 2006 03:35:44 -0700, jrodman@xxxx.xxxxxxxxxx.xxx wrote: > On Thu, May 25, 2006 at 09:46:46PM +0200, H.Merijn Brand wrote: > > On Thu, 25 May 2006 21:37:55 +0200, Juerd <juerd@xxxxxxxxxxx.xx> wrote: > > > > > Chris Devers skribis 2006-05-25 15:28 (-0400): > > > > There's a simple solution to this, of course. > > > > For all defined single-letter options, handle them appropriately. > > > > For all undefined single-letter options, return with help details. > > > > > > -? is still dangerous: > > > > > > touch -- -r > > > foo -? > > > > Your fault, core dumped. > > Here, let me plant a bomb on our shared computer: > > touch /tmp/-r > > It may take a long time to go off, but if it does, I guess that's your > fault too? No it's yours. *YOU* planted the bomb. You cannot blame someone using a train for a bomb a terrorist planted.
From: jrodman Date: 13:50 on 26 May 2006 Subject: Re: A simple hate today. On Fri, May 26, 2006 at 02:04:53PM +0200, H.Merijn Brand wrote: > On Fri, 26 May 2006 03:35:44 -0700, jrodman@xxxx.xxxxxxxxxx.xxx wrote: > > > On Thu, May 25, 2006 at 09:46:46PM +0200, H.Merijn Brand wrote: > > > On Thu, 25 May 2006 21:37:55 +0200, Juerd <juerd@xxxxxxxxxxx.xx> wrote: > > > > > > > Chris Devers skribis 2006-05-25 15:28 (-0400): > > > > > There's a simple solution to this, of course. > > > > > For all defined single-letter options, handle them appropriately. > > > > > For all undefined single-letter options, return with help details. > > > > > > > > -? is still dangerous: > > > > > > > > touch -- -r > > > > foo -? > > > > > > Your fault, core dumped. > > > > Here, let me plant a bomb on our shared computer: > > > > touch /tmp/-r > > > > It may take a long time to go off, but if it does, I guess that's your > > fault too? > > No it's yours. *YOU* planted the bomb. Thanks for connecting the dots. The user gets screwed and it's not their fault, and they are exposed by this dumb -? pattern. Hate. -josh
From: Abigail Date: 20:19 on 26 May 2006 Subject: Re: A simple hate today. --imjhCm/Pyz7Rq5F2 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Fri, May 26, 2006 at 05:50:05AM -0700, jrodman@xxxx.xxxxxxxxxx.xxx wrote: > On Fri, May 26, 2006 at 02:04:53PM +0200, H.Merijn Brand wrote: > > On Fri, 26 May 2006 03:35:44 -0700, jrodman@xxxx.xxxxxxxxxx.xxx wrote: > >=20 > > > On Thu, May 25, 2006 at 09:46:46PM +0200, H.Merijn Brand wrote: > > > > On Thu, 25 May 2006 21:37:55 +0200, Juerd <juerd@xxxxxxxxxxx.xx> wr= ote: > > > >=20 > > > > > Chris Devers skribis 2006-05-25 15:28 (-0400): > > > > > > There's a simple solution to this, of course. > > > > > > For all defined single-letter options, handle them appropriatel= y. > > > > > > For all undefined single-letter options, return with help detai= ls. > > > > >=20 > > > > > -? is still dangerous: > > > > >=20 > > > > > touch -- -r > > > > > foo -? > > > >=20 > > > > Your fault, core dumped. > > >=20 > > > Here, let me plant a bomb on our shared computer: > > >=20 > > > touch /tmp/-r > > >=20 > > > It may take a long time to go off, but if it does, I guess that's your > > > fault too? > >=20 > > No it's yours. *YOU* planted the bomb. >=20 > Thanks for connecting the dots. The user gets screwed and it's not > their fault, and they are exposed by this dumb -? pattern. I disagree. If you use 'rm', you ought to be aware of the '-r' option. If you've=20 choosen to use a shell that expands '-?' for you, and you choose to use '-?' unquoted, and get bitten, well, boohoo. I don't find that hateful. Abigail --imjhCm/Pyz7Rq5F2 Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) iD8DBQFEd1S8BOh7Ggo6rasRAoV6AJ92Hzo3ERwZ3ZxHCmLqJE2ocHDanACeM4fK 8Ji1lGjKeu8t3wmslkQRpbg= =pLxK -----END PGP SIGNATURE----- --imjhCm/Pyz7Rq5F2--
From: A. Pagaltzis Date: 20:42 on 26 May 2006 Subject: Re: A simple hate today. * Abigail <abigail@xxxxxxx.xx> [2006-05-26 21:20]: > If you use 'rm', you ought to be aware of the '-r' option. If > you've choosen to use a shell that expands '-?' for you, Err, is there any shell which *won't* expand `-?` for you? > and you choose to use '-?' unquoted, and get bitten, well, > boohoo. I don't find that hateful. So asking the user to type a switch which can *never* be issued safely without escaping it is not hateful? This freakin' `-?` meme should just roll over and die. Regards,
From: Abigail Date: 21:14 on 26 May 2006 Subject: Re: A simple hate today. --l+goss899txtYvYf Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Fri, May 26, 2006 at 09:42:40PM +0200, A. Pagaltzis wrote: > * Abigail <abigail@xxxxxxx.xx> [2006-05-26 21:20]: > > If you use 'rm', you ought to be aware of the '-r' option. If > > you've choosen to use a shell that expands '-?' for you, >=20 > Err, is there any shell which *won't* expand `-?` for you? No doubt there is. > > and you choose to use '-?' unquoted, and get bitten, well, > > boohoo. I don't find that hateful. >=20 > So asking the user to type a switch which can *never* be issued > safely without escaping it is not hateful? Never? I'd say that using -? is risk free most of the time. Even when used with rm. > This freakin' `-?` meme should just roll over and die. That I don't disagree with. Abigail --l+goss899txtYvYf Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) iD8DBQFEd2G8BOh7Ggo6rasRAsGEAJsGYLw5WxLX4IITtWFhaPlZqvxf9ACfYjCu X/39I8yEYi3qdyaEGfwsfak= =p9bl -----END PGP SIGNATURE----- --l+goss899txtYvYf--
From: Bruce Richardson Date: 15:05 on 26 May 2006 Subject: Re: A simple hate today. On Fri, May 26, 2006 at 02:04:53PM +0200, H.Merijn Brand wrote: > On Fri, 26 May 2006 03:35:44 -0700, jrodman@xxxx.xxxxxxxxxx.xxx wrote: > > Here, let me plant a bomb on our shared computer: > > > > touch /tmp/-r > > > > It may take a long time to go off, but if it does, I guess that's your > > fault too? > > No it's yours. *YOU* planted the bomb. > You cannot blame someone using a train for a bomb a terrorist planted. That makes sense. Let's all go put . back into our PATHs to celebrate.
From: A. Pagaltzis Date: 15:31 on 26 May 2006 Subject: Re: A simple hate today. * H.Merijn Brand <h.m.brand@xxxxxx.xx> [2006-05-26 14:05]: > On Fri, 26 May 2006 03:35:44 -0700, jrodman@xxxx.xxxxxxxxxx.xxx wrote: > > Here, let me plant a bomb on our shared computer: > > > > touch /tmp/-r > > > > It may take a long time to go off, but if it does, I guess > > that's your fault too? > > No it's yours. *YOU* planted the bomb. You cannot blame someone > using a train for a bomb a terrorist planted. But engineers deliberately designing trains such that even tiny, hard to spot charges can cause catastrophic failure are to be encouraged, yes? Regards,
From: David Cantrell Date: 21:42 on 26 May 2006 Subject: Re: A simple hate today. On Fri, May 26, 2006 at 02:04:53PM +0200, H.Merijn Brand wrote: > On Fri, 26 May 2006 03:35:44 -0700, jrodman@xxxx.xxxxxxxxxx.xxx wrote: > > Here, let me plant a bomb on our shared computer: > > touch /tmp/-r > > It may take a long time to go off, but if it does, I guess that's your > > fault too? > No it's yours. *YOU* planted the bomb. > You cannot blame someone using a train for a bomb a terrorist planted. It's not about blame. It's about who suffers. In this case *you* suffer from *his* actions because you use a character which has special meaning to the shell, and you don't want that.
From: H.Merijn Brand
Date: 11:01 on 27 May 2006
Subject: Re: A simple hate today.
On Fri, 26 May 2006 21:42:25 +0100, David Cantrell <david@xxxxxxxx.xxx.xx>
wrote:
> On Fri, May 26, 2006 at 02:04:53PM +0200, H.Merijn Brand wrote:
> > On Fri, 26 May 2006 03:35:44 -0700, jrodman@xxxx.xxxxxxxxxx.xxx wrote:
> > > Here, let me plant a bomb on our shared computer:
> > > touch /tmp/-r
> > > It may take a long time to go off, but if it does, I guess that's your
> > > fault too?
> > No it's yours. *YOU* planted the bomb.
> > You cannot blame someone using a train for a bomb a terrorist planted.
>
> It's not about blame. It's about who suffers. In this case *you*
> suffer from *his* actions because you use a character which has special
> meaning to the shell, and you don't want that.
I don't suffer, because I have protected myself (as much as possible) against
that. At the end, no-one really cares about the victim, and all blame the
attacker.
Anyway, My plea was not about how shells interpret -?, but that programs
should support it as alias for --help. -\? is still a lot less characters
than --help, and IMHO still evenly clear.
OK, I've learned a lot from all the opinions raised in this hate, and I will
continue to support *both* --help and -? for all my scripts. There is no
harm nor security issue from the script/program side of that.
I still have no intention to have support for -h. To *me* it is an illogical
choice and I am likely to choose that letter for something else.
I will probably review my scripts to
* send a requested help to STDOUT, and exit with 0
* send error help to STDERR, and exit with non-null
* think about the error message regarding how help should be asked for
Most likely that will end up with a construct like
--8<---
use strict;
use warnings;
sub usage ($)
{
my $err = shift and select STDERR;
print "usage: ...";
exit $err;
} # usage
@ARGV == 1 and $ARGV[0] eq "-?" || $ARGV[0] =~ m/^-+help$/ and usage (0);
use Getopt::Long qw(:config bundling nopermute);
my $opt_v = 0;
GetOptions (
"v:1" => \$opt_v,
) or usage (1);
-->8---
Sounds sane?
if someone sees the need to extend that with -h, for whatever reason, that is
easy done in the GetOptions
"h|aide|hilfe" => sub { usage (0) },
From: demerphq Date: 11:16 on 27 May 2006 Subject: Re: A simple hate today. On 5/27/06, H.Merijn Brand <h.m.brand@xxxxxx.xx> wrote: > On Fri, 26 May 2006 21:42:25 +0100, David Cantrell <david@xxxxxxxx.xxx.xx> > wrote: > > > On Fri, May 26, 2006 at 02:04:53PM +0200, H.Merijn Brand wrote: > > > On Fri, 26 May 2006 03:35:44 -0700, jrodman@xxxx.xxxxxxxxxx.xxx wrote: > > > > Here, let me plant a bomb on our shared computer: > > > > touch /tmp/-r > > > > It may take a long time to go off, but if it does, I guess that's your > > > > fault too? > > > No it's yours. *YOU* planted the bomb. > > > You cannot blame someone using a train for a bomb a terrorist planted. > > > > It's not about blame. It's about who suffers. In this case *you* > > suffer from *his* actions because you use a character which has special > > meaning to the shell, and you don't want that. > > I don't suffer, because I have protected myself (as much as possible) against > that. At the end, no-one really cares about the victim, and all blame the > attacker. > > Anyway, My plea was not about how shells interpret -?, but that programs > should support it as alias for --help. -\? is still a lot less characters > than --help, and IMHO still evenly clear. > > OK, I've learned a lot from all the opinions raised in this hate, and I will > continue to support *both* --help and -? for all my scripts. There is no > harm nor security issue from the script/program side of that. > > I still have no intention to have support for -h. To *me* it is an illogical > choice and I am likely to choose that letter for something else. > > I will probably review my scripts to > > * send a requested help to STDOUT, and exit with 0 > * send error help to STDERR, and exit with non-null Doesnt Pod::Usage make this easy? Yves
From: Peter da Silva Date: 11:46 on 27 May 2006 Subject: Re: A simple hate today. On May 27, 2006, at 5:01 AM, H.Merijn Brand wrote: > OK, I've learned a lot from all the opinions raised in this hate, and > I will > continue to support *both* --help and -? for all my scripts. There is > no > harm nor security issue from the script/program side of that. And I consider even supporting it Hateful in the same way that ActiveX is Hateful.
From: Juerd Date: 14:11 on 27 May 2006 Subject: Re: A simple hate today. H.Merijn Brand skribis 2006-05-27 12:01 (+0200): > Anyway, My plea was not about how shells interpret -?, but that programs > should support it as alias for --help. -\? is still a lot less characters > than --help, and IMHO still evenly clear. Characters, yes, but my keyboard doesn't have a single key for each character. The "?" is shifted, it requires another keystroke. -\? is 4 keystrokes, with all keys but the "-" in a nice place. On QWERTY, it would be even worse: all keys would be awkwardly placed. "help" on the other hand, is in the alphabetical area, where my hands already are. "--help" is 6 keystrokes, but only 5 different characters, of which 4 alphabetical. "-\?" is 4 keystrokes, all in annoying places. (Dvorak does "-" a little better than QWERTY.) > OK, I've learned a lot from all the opinions raised in this hate, and I will > continue to support *both* --help and -? for all my scripts. There is no > harm nor security issue from the script/program side of that. If you maintain that there's no harm nor security issue, then you have not learned enough :) Will you document "-?" as "-\?"...? > I will probably review my scripts to > * send a requested help to STDOUT, and exit with 0 > * send error help to STDERR, and exit with non-null That's great. I like software that behaves like that. > sub usage ($) > (...) > Sounds sane? Apart from the prototype, yep. Juerd
From: Rhesa Rozendaal Date: 16:20 on 27 May 2006 Subject: Re: A simple hate today. all@xxxxxxxxxxxxxx.xxx % bs -? bs: invalid option -- ? Try `bs --help' for more information. all@xxxxxxxxxxxxxx.xxx % bs -h bs: invalid option -- h Try `bs --help' for more information. all@xxxxxxxxxxxxxx.xxx % bs --help bs: the color of the Bike Shed is a greenish, yellowish sort of brown. Now cut it out already!
From: David Champion Date: 18:43 on 27 May 2006 Subject: Re: A simple hate today. * On 2006.05.27, in <44786E2D.50306@xxxxxx.xx>, * "Rhesa Rozendaal" <rhesa@xxxxxx.xx> wrote: > all@xxxxxxxxxxxxxx.xxx % bs -? > bs: invalid option -- ? > Try `bs --help' for more information. > > all@xxxxxxxxxxxxxx.xxx % bs -h > bs: invalid option -- h > Try `bs --help' for more information. > > all@xxxxxxxxxxxxxx.xxx % bs --help > bs: the color of the Bike Shed is a greenish, yellowish sort of brown. And... iterate.
From: Bill Page Date: 19:05 on 28 May 2006 Subject: Re: A simple hate today. > "-\?" is 4 keystrokes, all in annoying places. (Dvorak does "-" a little > better than QWERTY.) you sicken me
From: Juerd Date: 16:42 on 28 May 2006 Subject: Re: A simple hate today. Bill Page skribis 2006-05-29 3:35 (+0930): > > "-\?" is 4 keystrokes, all in annoying places. (Dvorak does "-" a little > > better than QWERTY.) > you sicken me Ahhh, the satisfaction... :) Juerd
From: David Landgren Date: 21:26 on 28 May 2006 Subject: Re: A simple hate today. Juerd wrote: > -\? is 4 keystrokes, with all keys but the "-" in a nice place. On > QWERTY, it would be even worse: all keys would be awkwardly placed. And on a French azerty keyboard layout, the \ needs a right-alt to get it. Of course, the azerty layout is hateful in many other ways too. David
From: A. Pagaltzis
Date: 22:05 on 28 May 2006
Subject: Re: A simple hate today.
* David Landgren <david@xxxxxxxx.xxx> [2006-05-28 22:25]:
> And on a French azerty keyboard layout, the \ needs a right-alt
> to get it.
Same holds true for the German QWERTZ layout; it's actually on
the same key as ?, except the ? needs Shift.
> Of course, the azerty layout is hateful in many other ways too.
Same holds true for the German QWERTZ layout... the worst, for a
programmer, is that {[]}\ are located on AltGr-7 thru -ß (which
is right of the 0). These had always been a pain in the neck, but
after I had to get used to a US-like layout for a few weeks I
just couldn't go back anymore.
Regards,
From: demerphq
Date: 22:24 on 28 May 2006
Subject: Re: A simple hate today.
On 5/28/06, A. Pagaltzis <pagaltzis@xxx.xx> wrote:
> * David Landgren <david@xxxxxxxx.xxx> [2006-05-28 22:25]:
> > And on a French azerty keyboard layout, the \ needs a right-alt
> > to get it.
>
> Same holds true for the German QWERTZ layout; it's actually on
> the same key as ?, except the ? needs Shift.
>
> > Of course, the azerty layout is hateful in many other ways too.
>
> Same holds true for the German QWERTZ layout... the worst, for a
> programmer, is that {[]}\ are located on AltGr-7 thru -=DF (which
> is right of the 0). These had always been a pain in the neck, but
> after I had to get used to a US-like layout for a few weeks I
> just couldn't go back anymore.
I agree totally. I always hated the german keyboard and would do
anything (including mail order purchasing) to get a US or UK keyboard.
UK is a little nicer as it has a convenient euro key, but they both
are vastly superior for coding than the German keyboard is. With C
like languages you end typing with your thumb almost permanently
pressed on the alt-gr key...
Yves
--=20
perl -Mre=3Ddebug -e "/just|another|perl|hacker/"
From: Abigail Date: 21:03 on 25 May 2006 Subject: Re: A simple hate today. --zbGR4y+acU1DwHSi Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Thu, May 25, 2006 at 03:28:10PM -0400, Chris Devers wrote: >=20 > -h wouldn't have to be defined or special-cased, because the catchall =20 > would catch it. Disagree. Even if the generated messages are the same, use of an undefined option should write a message to stderr, and exit with a non-zero value, while if the user asks for help, the message should be written to stdout, and the program should exit with a zero value. Abigail --zbGR4y+acU1DwHSi Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) iD8DBQFEdg17BOh7Ggo6rasRAnqTAJ9aUrg4TtzqeHwwyj6gj34Q3k5EMACfR54W +bvgjSGF2fDhVidwmvA7G24= =rDfw -----END PGP SIGNATURE----- --zbGR4y+acU1DwHSi--
From: peter (Peter da Silva) Date: 19:45 on 25 May 2006 Subject: Re: A simple hate today. > You have to escape it from your shell? Either you have files whose > name is simply hyphen plus another character in the current directory, > or you have a hateful shell. I have a shell that allows me to "set nonomatch" to turn this off. I do not do this, because having metacharacters sometimes working without quoting and sometimes not is hateful. I have to quote them anyway, even if I think I'm safe, so there's no point to ever treating them as normal characters. See also, hateful automatic command abbreviations.
From: A. Pagaltzis Date: 04:48 on 26 May 2006 Subject: Re: A simple hate today. * Peter da Silva <peter@xxxxxxx.xxx> [2006-05-25 20:50]: > I have a shell that allows me to "set nonomatch" to turn this > off. A "9 times matchable" option? Regards,
From: spc (Sean 'Captain Napalm' Conner) Date: 19:30 on 25 May 2006 Subject: Re: A simple hate today. It was thus said that the Great H.Merijn Brand once stated: > > On Thu, 25 May 2006 09:53:56 -0500 (CDT), sabrina downard <sld@xxxxxxxx.xxx> > wrote: > > > If you do not have a better use for -h, and you have coded a usage > > statement for --help, then how damned difficult would it be to make -h > > equal to --help? Since you're already going to print a damned error > > telling me to use --help instead? > > Sorry, I disagree. And I have never liked -h to be help. > IMHO it should be either -? or --help, and -help could be acceptable Notice she said "if you do not have a better use for -h." Me? The stuff I write will print a help message if any options given aren't supported. But I agree with Sabrina: GenericUnixPrompt> mv -h mv: invalid option -- h Try `mv --help' for more information. Um? Okay, it's an invalid switch. You detected it's an invalid switch. You printed an error as such. -spc (What I hate is when the man page says "Use info")
From: H.Merijn Brand Date: 20:55 on 25 May 2006 Subject: Re: A simple hate today. On Thu, 25 May 2006 14:30:35 -0400 (EDT), spc@xxxxxx.xxx (Sean 'Captain Napalm' Conner) wrote: > It was thus said that the Great H.Merijn Brand once stated: > > > > On Thu, 25 May 2006 09:53:56 -0500 (CDT), sabrina downard <sld@xxxxxxxx.xxx> > > wrote: > > > > > If you do not have a better use for -h, and you have coded a usage > > > statement for --help, then how damned difficult would it be to make -h > > > equal to --help? Since you're already going to print a damned error > > > telling me to use --help instead? > > > > Sorry, I disagree. And I have never liked -h to be help. > > IMHO it should be either -? or --help, and -help could be acceptable > > Notice she said "if you do not have a better use for -h." And I added *yet*. Most of the programs I write are never done, and get more and more options along the lifetime of the utility. > Me? The stuff I write will print a help message if any options given > aren't supported. But I agree with Sabrina: > > GenericUnixPrompt> mv -h > mv: invalid option -- h > Try `mv --help' for more information. > > Um? Okay, it's an invalid switch. You detected it's an invalid switch. > You printed an error as such. > > -spc (What I hate is when the man page says "Use info") that sucks.
From: Juerd
Date: 20:12 on 25 May 2006
Subject: Re: A simple hate today.
H.Merijn Brand skribis 2006-05-25 18:19 (+0200):
> IMHO it should be either -? or --help, and -help could be acceptable
-? is silly. In most shells, it's a glob match, and is passed to the
program only if it didn't match.
I couldn't easily find any program that supports -?...
juerd@nano:~$ ls -?
ls: invalid option -- ?
Try `ls --help' for more information.
juerd@nano:~$ grep -?
grep: invalid option -- ?
Usage: grep [OPTION]... PATTERN [FILE]...
Try `grep --help' for more information.
juerd@nano:~$ du -?
du: invalid option -- ?
Try `du --help' for more information.
juerd@nano:~$ cp -?
cp: invalid option -- ?
Try `cp --help' for more information.
juerd@nano:~$ mv -?
mv: invalid option -- ?
Try `mv --help' for more information.
juerd@nano:~$ mkdir -?
mkdir: invalid option -- ?
Try `mkdir --help' for more information.
juerd@nano:~$ perl -?
Unrecognized switch: -? (-h will show valid options).
juerd@nano:~$ ed -?
ed: invalid option -- ?
Try `ed --help' for more information.
juerd@nano:~$ vi -?
VIM - Vi IMproved 6.4 (2005 Oct 15, compiled Apr 28 2006 01:45:37)
Unknown option: "-?"
More info with: "vim -h"
juerd@nano:~$ yes -?
yes: invalid option -- ?
Try `yes --help' for more information.
juerd@nano:~$ awk -?
awk: not an option: -?
juerd@nano:~$ python -?
Unknown option: -?
usage: python [option] ... [-c cmd | -m mod | file | -] [arg] ...
Try `python -h' for more information.
juerd@nano:~$ ruby -?
ruby: invalid option -? (-h will show valid options)
I found a few things that do support it, for some values of support.
The most likely reason for -? to work is that a program always gives
help when you use an unknown option (mount, ps). And sometimes it works,
but isn't documented (zip, gzip, firefox).
The only programs I could find that officially support this are abiword,
activation-client (some gnome thing), aspell, botti, epydocgui,
fc-{cache,list,match}, gconftool, gdm-signal, getent, gij,
gtk-update-icon-cache, h2xs, iconv, inkscape, irssi, java, less,
locale{,def}, nano, nmblookup, oo{base,calc,draw,ffice,fromtemplate,
impress,math,web}, pango-{querymodules,view}, pdbedit, pdf{fonts,images,
info,toppm,tops,totext}, pico, pon, profiles, rpcclient,
smb{cacls,control,get,mount,spool,client,cquotas,mnt,status,tree},
synclient, wvdial{,conf}, xkb{comp,evd,print}
I suspect you have something to do with the introduction of -? into
h2xs, in between perl 5.7.2 and 5.7.3 :)
So that's about 60 programs that do support -?, against over 1200 that
do not. Sanity wins!
> For me, -h never has meant help, nor do I expect *any* command or util to
> show me help on -h
Many, many more programs have -h bound to help, compared to those that
support -?.
Juerd
From: Philip Newton Date: 20:17 on 25 May 2006 Subject: Re: A simple hate today. On 5/25/06, Juerd <juerd@xxxxxxxxxxx.xx> wrote: > I couldn't easily find any program that supports -?... Presumably you looked at the GNU versions of most of those tools, though, rather than original UNIX versions. Apples and oranges. I thought it had been established that GNU tools don't support -? on general principles. Cheers,
From: Juerd Date: 20:21 on 25 May 2006 Subject: Re: A simple hate today. Philip Newton skribis 2006-05-25 21:17 (+0200): > On 5/25/06, Juerd <juerd@xxxxxxxxxxx.xx> wrote: > >I couldn't easily find any program that supports -?... > Presumably you looked at the GNU versions of most of those tools, > though, rather than original UNIX versions. Apples and oranges. I > thought it had been established that GNU tools don't support -? on > general principles. Not all of these over 1200 programs are GNU. There are numerous non-GNU programs that do not support -?. But if GNU caused all these programs to not support -?, then I thank them for it. Juerd
From: H.Merijn Brand
Date: 20:43 on 25 May 2006
Subject: Re: A simple hate today.
On Thu, 25 May 2006 21:12:39 +0200, Juerd <juerd@xxxxxxxxxxx.xx> wrote:
> H.Merijn Brand skribis 2006-05-25 18:19 (+0200):
> > IMHO it should be either -? or --help, and -help could be acceptable
>
> -? is silly. In most shells, it's a glob match, and is passed to the
> program only if it didn't match.
>
> I couldn't easily find any program that supports -?...
>
> juerd@nano:~$ ls -?
> ls: invalid option -- ?
> Try `ls --help' for more information.
> juerd@nano:~$ grep -?
> grep: invalid option -- ?
> Usage: grep [OPTION]... PATTERN [FILE]...
> Try `grep --help' for more information.
> juerd@nano:~$ du -?
> du: invalid option -- ?
> Try `du --help' for more information.
> juerd@nano:~$ cp -?
> cp: invalid option -- ?
> Try `cp --help' for more information.
> juerd@nano:~$ mv -?
> mv: invalid option -- ?
> Try `mv --help' for more information.
> juerd@nano:~$ mkdir -?
> mkdir: invalid option -- ?
> Try `mkdir --help' for more information.
> juerd@nano:~$ perl -?
> Unrecognized switch: -? (-h will show valid options).
> juerd@nano:~$ ed -?
> ed: invalid option -- ?
> Try `ed --help' for more information.
> juerd@nano:~$ vi -?
> VIM - Vi IMproved 6.4 (2005 Oct 15, compiled Apr 28 2006 01:45:37)
> Unknown option: "-?"
> More info with: "vim -h"
> juerd@nano:~$ yes -?
> yes: invalid option -- ?
> Try `yes --help' for more information.
> juerd@nano:~$ awk -?
> awk: not an option: -?
> juerd@nano:~$ python -?
> Unknown option: -?
> usage: python [option] ... [-c cmd | -m mod | file | -] [arg] ...
> Try `python -h' for more information.
> juerd@nano:~$ ruby -?
> ruby: invalid option -? (-h will show valid options)
All GNU, as Philip said.
d3:/usr/bin 105 > ls -?
ls: illegal option -- ?
usage: ls -1ARadeCxmnlogrtucpFLbqisf [files]
d3:/usr/bin 106 > grep -?
grep: illegal option -- ?
usage: grep [-E|-F] [-c|-l|-q] [-binsvx] -e pattern_list...
[-f pattern_file...] [file...]
usage: grep [-E|-F] [-c|-l|-q] [-binsvx] [-e pattern_list...]
-f pattern_file... [file...]
usage: grep [-E|-F] [-c|-l|-q] [-binsvx] pattern [file...]
d3:/usr/bin 107 > du -?
du: illegal option -- ?
usage: du [-a|-s] [-kbrx] [-t type] [name ...]
d3:/usr/bin 108 > cp -?
cp: illegal option -- ?
Usage: cp [-f|-i] [-p] [-S] [-e warn|force|ignore] source_file target_file
cp [-f|-i] [-p] [-S] [-e warn|force|ignore] source_file ... target_directory
cp [-f|-i] [-p] [-S] -R|-r [-e warn|force|ignore] source_directory ... target_directory
d3:/usr/bin 109 > mkdir -?
mkdir: illegal option -- ?
usage: mkdir [-p] [-m mode] dirname ...
d3:/usr/bin 109 > mkdir -?
mkdir: illegal option -- ?
usage: mkdir [-p] [-m mode] dirname ...
d3:/usr/bin 110 > ed -?
ed: illegal option -- ?
d3:/usr/bin 111 > ./vi -?
vi: illegal option -- ?
d3:/usr/bin 112 > sh -?
-?: A specified flag is not valid for this command.
Look, here is a genuine old awk and sed :)
d3:/usr/bin 113 > awk -?
Usage: awk [-F fs][-v Assignment][-f Progfile|Program][Assignment|File] ...
d3:/usr/bin 114 > sed -?
Usage: sed [-n] [-e script] [-f source_file] [file...]
I don't have Ruby or Python installed
csh is funny. It just ignores it and starts
d3:/usr/bin 117 > csh -?
%
Echo is also OK
d3:/usr/bin 123 > echo -?
-?
Some *very* old commands:
d3:/usr/bin 133 > metamail -?
Usage: metamail [-b] [-B] [-d] [-e] [-h] [-r] [-R] [-p] [-P] [-x] [-y] [-z] [-c content-type] [-E content-transfer-encoding] [-f from-name] [-m mailername] [-s subject] [message-file-name]
d3:/usr/bin 134 > elmalias -?
Usage: elmalias [ -c alias-list | -q ]
or elmalias -l [ regular-expression ]
-c alias-list checkalias - check to see if an alias is defined.
-l expression listalias - list user and system alias.
-q systemalias - install system-aliases.
d3:/usr/bin 142 > rup -?
Usage: rup [-h] [-l] [-t] [host ...]
d3:/usr/bin 147 > jobs -?
Usage: jobs [ -l ].
> I found a few things that do support it, for some values of support.
> The most likely reason for -? to work is that a program always gives
> help when you use an unknown option (mount, ps). And sometimes it works,
> but isn't documented (zip, gzip, firefox).
>
> The only programs I could find that officially support this are abiword,
> activation-client (some gnome thing), aspell, botti, epydocgui,
> fc-{cache,list,match}, gconftool, gdm-signal, getent, gij,
> gtk-update-icon-cache, h2xs, iconv, inkscape, irssi, java, less,
> locale{,def}, nano, nmblookup, oo{base,calc,draw,ffice,fromtemplate,
> impress,math,web}, pango-{querymodules,view}, pdbedit, pdf{fonts,images,
> info,toppm,tops,totext}, pico, pon, profiles, rpcclient,
> smb{cacls,control,get,mount,spool,client,cquotas,mnt,status,tree},
> synclient, wvdial{,conf}, xkb{comp,evd,print}
>
> I suspect you have something to do with the introduction of -? into
> h2xs, in between perl 5.7.2 and 5.7.3 :)
No, I didn't, but I proposed a change to perl a very long time ago, and I
still don't understand why it was turned down.
> So that's about 60 programs that do support -?, against over 1200 that
> do not. Sanity wins!
Again, you're in a GNU environment.
> > For me, -h never has meant help, nor do I expect *any* command or util to
> > show me help on -h
>
> Many, many more programs have -h bound to help, compared to those that
> support -?.
I'll remember that. Maybe even IBM and HP will listen :p
Now, come up with a good reason why -? should not map to --help
Instead of barfing with unsupported option. Shell escaping is NOT a good
reason, as the people that suffer from that don't use it anyway.
-? as an alias for --help makes sense. I ask for help: just give it.
From: Juerd
Date: 21:00 on 25 May 2006
Subject: Re: A simple hate today.
H.Merijn Brand skribis 2006-05-25 21:43 (+0200):
> > juerd@nano:~$ perl -?
> > Unrecognized switch: -? (-h will show valid options).
> > juerd@nano:~$ vi -?
> > VIM - Vi IMproved 6.4 (2005 Oct 15, compiled Apr 28 2006 01:45:37)
> > Unknown option: "-?"
> > More info with: "vim -h"
> > juerd@nano:~$ python -?
> > Unknown option: -?
> > usage: python [option] ... [-c cmd | -m mod | file | -] [arg] ...
> > Try `python -h' for more information.
> > juerd@nano:~$ ruby -?
> > ruby: invalid option -? (-h will show valid options)
> All GNU, as Philip said.
GNU Perl?
GNU VIM?
GNU Python?
GNU Ruby?
> Look, here is a genuine old awk and sed :)
> d3:/usr/bin 113 > awk -?
> Usage: awk [-F fs][-v Assignment][-f Progfile|Program][Assignment|File] ...
> d3:/usr/bin 114 > sed -?
> Usage: sed [-n] [-e script] [-f source_file] [file...]
Interesting that they don't list -? in the usage information. What do
these awk and sed do if you run
awk -@
sed -@
If they then too give usage info, -? isn't a feature, but a useful side
effect.
> d3:/usr/bin 133 > metamail -?
How about
metamail -@
> d3:/usr/bin 134 > elmalias -?
elmalias -@
> d3:/usr/bin 142 > rup -?
> Usage: rup [-h] [-l] [-t] [host ...]
rup -@
> d3:/usr/bin 147 > jobs -?
> Usage: jobs [ -l ].
jobs -@
> No, I didn't, but I proposed a change to perl a very long time ago, and I
> still don't understand why it was turned down.
I do, especially after reading this thread.
> > So that's about 60 programs that do support -?, against over 1200 that
> > do not. Sanity wins!
> Again, you're in a GNU environment.
OpenBSD 3.2, not quite a GNU environment:
$ ls -?
ls: illegal option -- ?
usage: ls [-1ACFLRSTWacdfiklmnopqrstux] [file ...]
$ grep -?
grep: invalid option -- ?
Usage: grep [OPTION]... PATTERN [FILE]...
Try `grep --help' for more information.
$ mv -?
mv: illegal option -- ?
usage: mv [-fi] source target mv [-fi] source ... directory
$ vi -?
vi: illegal option -- ?
usage: ex [-eFRrSsv] [-c command] [-t tag] [-w size] [file ...]
usage: vi [-eFlRrSv] [-c command] [-t tag] [-w size] [file ...]
$ mkdir -?
mkdir: illegal option -- ?
usage: mkdir [-p] [-m mode] dirname ...
Blaming GNU is too easy.
> Now, come up with a good reason why -? should not map to --help
Oh, I'm not arguing that -? shouldn't be supported. I am arguing that
expecting that it is supported, is rather silly. I am also warning that
using -? may be unsafe, and that -h or --help is more likely to work
anyway.
Supporting -? and documenting it is bad, though, because of
aforementioned security risks and because people might come to expect
-? to work from other applications.
> -? as an alias for --help makes sense. I ask for help: just give it.
I'll give you help info, but only because you're feeding me something I
don't understand. Not because I think "?" looks like "help". It looks
more like "huh?", but that isn't something I'd shout when drowning.
Juerd
From: Richard Clamp Date: 22:15 on 28 May 2006 Subject: Re: A simple hate today. From a fine hate we seem to have devolved into some kind of talking shop with it's own brand of retarded one-upmanship. Enough.
Generated at 12:27 on 27 Sep 2007 by mariachi