Thread: proposal: new long psql parameter --on-error-stop

Pavel Stehule <pavel.stehule@gmail.com> writes:
> for beginners is relative difficult to set psql variable from command line
> and option
> -v ON_ERROR_STOP=1 is little bit unclean.
> and for any user it is not comfortable. So I propose a new psql option
> --on-error-stop

That saves a whole three characters ... not sure it's really worth it,
especially since it will not work on older versions.

Also, it's not like this will save people from having to know about the
ON_ERROR_STOP variable, because there will still be plenty of contexts
where they need to know that (eg, inspecting the setting or changing it
mid-session).  So I don't buy the argument that this simplifies what
beginners need to learn.

                        regards, tom lane

On Fri, Feb 28, 2014 at 3:17 PM, Pavel Stehule <pavel.stehule@gmail.com> wrote:
>
> 2014-02-28 19:13 GMT+01:00 Tom Lane <tgl@sss.pgh.pa.us>:
>
>> Pavel Stehule <pavel.stehule@gmail.com> writes:
>> > for beginners is relative difficult to set psql variable from command line
>> > and option
>> > -v ON_ERROR_STOP=1 is little bit unclean.
>> > and for any user it is not comfortable. So I propose a new psql option
>> > --on-error-stop
>>
>> That saves a whole three characters ... not sure it's really worth it,
>> especially since it will not work on older versions.
>>
>> Also, it's not like this will save people from having to know about the
>> ON_ERROR_STOP variable, because there will still be plenty of contexts
>> where they need to know that (eg, inspecting the setting or changing it
>> mid-session).  So I don't buy the argument that this simplifies what
>> beginners need to learn.
>
>
> three chars is not important
>
> important is a  placing in --help output
>

Well, then we just have to add more info to --help

Regards,

--
Fabrízio de Royes Mello
Consultoria/Coaching PostgreSQL
>> Timbira: http://www.timbira.com.br
>> Blog sobre TI: http://fabriziomello.blogspot.com
>> Perfil Linkedin: http://br.linkedin.com/in/fabriziomello
>> Twitter: http://twitter.com/fabriziomello

On Fri, Feb 28, 2014 at 3:44 PM, Alvaro Herrera <alvherre@2ndquadrant.com> wrote:
>
> Fabrízio de Royes Mello escribió:
> > On Fri, Feb 28, 2014 at 3:17 PM, Pavel Stehule <pavel.stehule@gmail.com>
> > wrote:
>
> > > important is a  placing in --help output
> >
> > Well, then we just have to add more info to --help
>
> I think psql could do with some lines for the possible options for
> --pset (14) and --variable (13).  Not sure how to do that without having
> it become too cumbersome while not hiding useful variables.
>

Maybe by adding something like "--help-pset" and "--help-variables".

Regards,

--
Fabrízio de Royes Mello
Consultoria/Coaching PostgreSQL
>> Timbira: http://www.timbira.com.br
>> Blog sobre TI: http://fabriziomello.blogspot.com
>> Perfil Linkedin: http://br.linkedin.com/in/fabriziomello
>> Twitter: http://twitter.com/fabriziomello

Andrew Dunstan <andrew@dunslane.net> writes:
>> Well, then we just have to add more info to --help

> +1 for at least doing that. I found it annoying just the other day not
> to find it in plsql's --help output, in a moment of brain fade when I
> forgot how to spell it. So it's not just beginners who can benefit, it's
> people like me whose memory occasionally goes awry.

No objection in principle, but what are we talking about exactly?
Adding some new backslash command that lists all the variables that have
special meanings?  I'm not sure that the main --help output is the place
for this, because that only covers psql's command line switches (and
is plenty long enough already).

                        regards, tom lane

On Fri, February 28, 2014 22:38, Tom Lane wrote:
> Andrew Dunstan <andrew@dunslane.net> writes:
>>> Well, then we just have to add more info to --help
>
>> +1 for at least doing that. I found it annoying just the other day not
>> to find it in plsql's --help output, in a moment of brain fade when I
>> forgot how to spell it. So it's not just beginners who can benefit, it's
>> people like me whose memory occasionally goes awry.
>
> No objection in principle, but what are we talking about exactly?
> Adding some new backslash command that lists all the variables that have
> special meanings?  I'm not sure that the main --help output is the place
> for this, because that only covers psql's command line switches (and
> is plenty long enough already).
>
>                       regards, tom lane

Perhaps this compromise:

  -v, --set=, --variable=NAME=VALUE
                           set psql variable NAME to VALUE e.g.: -v ON_ERROR_STOP=1

this would not lengthen and not broaden the output of psql --help

On 02/28/2014 04:38 PM, Tom Lane wrote:

Andrew Dunstan <andrew@dunslane.net> writes:

Well, then we just have to add more info to --help

+1 for at least doing that. I found it annoying just the other day not
to find it in plsql's --help output, in a moment of brain fade when I
forgot how to spell it. So it's not just beginners who can benefit, it's
people like me whose memory occasionally goes awry.

No objection in principle, but what are we talking about exactly?
Adding some new backslash command that lists all the variables that have
special meanings?

That's a pretty good idea, especially if we give that command a command line option too, so something like

   psql --special-variables

would run that command and exit.

Maybe I'm over-egging the pudding a bit ;-)

cheers

andrew

On Sat, Mar 1, 2014 at 5:37 AM, Pavel Stehule <pavel.stehule@gmail.com> wrote:
>
> Hello
>
> here is a prototype:
>
> bash-4.1$ /usr/local/pgsql/bin/psql --help-variables
> List of some variables (options) for use from command line.
> Complete list you find in psql section in the PostgreSQL documentation.
>
> psql variables:
> Usage:
>   psql --set=NAME=VALUE
>   or \set NAME VALUE in interactive mode
>
>   AUTOCOMMIT         when is on, successful SQL command is automatically commited
>   COMP_KEYWORD_CASE  determines which letter case to use when completing an SQL key word
>   ECHO               all lines from input can be written to standard output
>   ECHO_HIDDEN        display queries for internal commands (same as -E option)
>   FETCH_COUNT        how many rows should be for one page (default 0 unlimited)
>   HISTFILE           file name that be used for store history list
>   HISTSIZE           the number of commands to store in the command history
>   ON_ERROR_ROLLBACK  when is on, raise ROLLBACK on error automatically
>   ON_ERROR_STOP      when is set, then batch execution stop immediately after error
>   VERBOSITY          control verbosity of error reports [default, verbose, terse]
>
> Printing options:
> Usage:
>   psql --pset=NAME[=VALUE]
>   or \pset NAME [VALUE] in interactive mode
>
>   border             number of border style
>   fieldsep           specify field separator for unaligned output
>   fieldsep_zero      field separator in unaligned mode will be zero
>   format             set output format [unaligned, aligned, wrapped, html, latex, ..]
>   linestyle          sets the border line drawing style [ascii, old-ascii, unicode]
>   null               sets the string to be printed in place of a null value
>   pager              when the pager option is off, the pager program is not used
>   recordsep          specifies the record (line) separator to use in unaligned output format
>   recordsep_zero     record separator be in unaligned output format a zero byte
>   title              sets the table title for any subsequently printed tables
>   tuples_only        in tuples-only mode, only actual table data is shown
>
> Environment options:
> Usage:
>   NAME=VALUE, [NAME=VALUE] psql ...
>   or \setenv NAME [VALUE] in interactive mode
>
>   COLUMNS            number of columns for wrapped format
>   PAGER              used pager
>   PGHOST             same as the host connection parameter
>   PGDATABASE         same as the dbname connection parameter
>   PGUSER             same as the user connection parameter
>   PGPASSWORD         possibility to set password
>   PSQL_EDITOR, EDITOR, VISUAL  editor used by \e \ef commands
>   PSQL_EDITOR_LINE_NUMBER_ARG  style how to line number is used in editor
>   PSQL_HISTORY       alternative location for the command history file
>   PSQL_RC            alternative location of the user's .psqlrc file
>   SHELL              command executed by the \! command
>   TMPDIR             directory for storing temporary files
>
> For more information consult the psql section in the PostgreSQL
> documentation.
>

The patch is ok (apply to master and apply to master without errors).

Maybe we must show the possible values for each variable/option too.

Thinking more about it, would be nice if we have the possibility to show help for commands too. Some like that:

$ psql -H vacuum
Command:     VACUUM
Description: garbage-collect and optionally analyze a database
Syntax:
VACUUM [ ( { FULL | FREEZE | VERBOSE | ANALYZE } [, ...] ) ] [ table_name [ (column_name [, ...] ) ] ]
VACUUM [ FULL ] [ FREEZE ] [ VERBOSE ] [ table_name ]
VACUUM [ FULL ] [ FREEZE ] [ VERBOSE ] ANALYZE [ table_name [ (column_name [, ...] ) ] ]

$ psql --help-command=vacuum
Command:     VACUUM
Description: garbage-collect and optionally analyze a database
Syntax:
VACUUM [ ( { FULL | FREEZE | VERBOSE | ANALYZE } [, ...] ) ] [ table_name [ (column_name [, ...] ) ] ]
VACUUM [ FULL ] [ FREEZE ] [ VERBOSE ] [ table_name ]
VACUUM [ FULL ] [ FREEZE ] [ VERBOSE ] ANALYZE [ table_name [ (column_name [, ...] ) ] ]

It's only an idea that occurred to me reading this thread!

Grettings,

--
Fabrízio de Royes Mello
Consultoria/Coaching PostgreSQL
>> Timbira: http://www.timbira.com.br
>> Blog sobre TI: http://fabriziomello.blogspot.com
>> Perfil Linkedin: http://br.linkedin.com/in/fabriziomello
>> Twitter: http://twitter.com/fabriziomello

From: "Pavel Stehule" <pavel.stehule@gmail.com>

I am not sure in this point. It is aligned left with all long options:

 -?, --help               show this help, then exit
     --help-variables     list of available configuration variables
(options), then exit

pg_dumpall aligns all options left with each other, whether they are short or long.

 -x, --no-privileges          do not dump privileges (grant/revoke)
 --binary-upgrade             for use by upgrade utilities only
 --column-inserts             dump data as INSERT commands with column names

Regards
MauMau

From: "Pavel Stehule" <pavel.stehule@gmail.com>

pg_dumpall aligns all options left with each other, whether they are short
or long.

 -x, --no-privileges          do not dump privileges (grant/revoke)
 --binary-upgrade             for use by upgrade utilities only
 --column-inserts             dump data as INSERT commands with column
names

ok

I fixed it

Thank you.  I marked this patch as ready for committer.

Regards
MauMau