Re: [Info] Get more from the command line

[Harry Putnam]

Eli Zaretskii <address@hidden> writes:

>> ??? What does this have to do with Emacs?

>From your perspective as an author and developer, the difference
between them may loom large.  But for common users I believe the two
programs are well intertwined.  Using shared commands and proceedures.
For us the similarities are center stage.  Futher I believe many if not
most emacs users are also info users and may well be interested in this
subject as it relates to emacs.

If I here complaints, I will reconsider.

>> --subnodes
>>    recurses and outputs the subnodes and most often is too much
>>    information at one time. Unless one brings to bear other tools to
>>    filter it. No problem there, but if one could get a list of the
>>    subnodes available it would be more practical to ask for certain
>>    ones.
>
> Noted.

Thanks for noticing... a command that spits out all the nodes of a
section would be very handy.
>
>> --output=FILENAME
>>    output selected nodes to FILENAME
>> 
>>    This would be more usefull if there was a choice of stdout or a
>>    file.  I guess one could name tty as the file but that is kind of
>>    hackish and not something one thinks of immediately.
>
> Doesn't "--output=-" work?

It may very well do so, but not too easy to find how.  I tried quite
a few formulas and still have not hit it.  Maybe a few examples would
be in order. I tried way too many of the possible combinations to list,
but never did get the output. 
Using several variations --output=- --output - --output-
And with some 5 pieces of cl argument there are literally hundreds
of possible ways this could be arranged.

 $ info --output - --node dired emacs    
 $ info --output=- --node dired emacs
 $ info --node dired emacs --output-

Might be handy to have a few working examples at the usage section.
The ones that are there don't cover outputing stuff to stdout much.

>>  --show-options, --usage
>>    go to command-line options node.
>> 
>>    This would be much more useful if it could be directed to stdout
>>    rather than invoking the reader.  Or better still would be some
>>    switch to invoke stdout or open the reader.  
>>   
>>    I guess it could be coupled with  -output=filename but again that
>>    assumes the user knows what the section is called.
>
> No, you don't need to know the name of the section, --usage find it 
> automatically; that's the reason for its existence.
>
> So yes, --usage coupled with --output= should take care of this issue.

Again there is a shortage of actual examples.  I don't mean the kind
that programmers love, where you have to guess the real content.  I
mean actual working examples.

$ info --output= --usage standalone
info: Could not create output file `'.

info --output - --usage standalone
Shows the whole section.  Not the usage

info --usage  --output=-  emacs
Shows the whole first outline page, no usage at all.
But seems to be about what I wanted regarding a full listing of 
nodes.

Whether my tries are well informed may be a problem but I think it shows
that at the very best it is not an easy thing to figure out.
People have learned for thousands of years by concrete example.  If
ever something has stood the test of time, that is it.  Why do we 
discard a known reliable method?

I don't see any of the examples under `Examples' that show how to get
things on stdout.

Just a general comment here: There comes a time, when working on
something engrossing where it is no longer fun or interesting to
fiddle and experiment to see how an app can work.  Especially a help
application since when you need it, its usally to get going on
something else.  That is the one application that should be laden with
examples.

> Info supports PageUp and PageDown (as well as arrows and other popular 
> edit keys).  So I don't see how this could be a problem.

Arrows yes, but I must have terminal interference.  My page up/down
walks right out of the reader and up through what ever is in my term buffer.

> Besides, paging through the manual is easier with SPC and DEL keys, 
> rather than with PageUp/PageDown.

I'd forgotten that one.  Something about the reader makes you
forget....:-)


>> We need keys that do what emacs does with C-h m.  In info the C-h or ?
>> lead to the full documentation rather than the documentatin about the
>> mode you are in.  
>
> There's no modes in the stand-alone Info.  What would you like the 
> equivalent of C-h m to display?

Sloppy use of the term `modes' here, but there are sections:

I see several broad grouping therez:
  General navigaiton
  Commands available in Info windows:
  The search path
  Command for the echo area
  commands invokable by M-x

This is quite a lot to go thru.  Can it not be arraged to be accessed
in those general lumps or all at once?  

Maybe begin with a menu similar to what emacs does with C-h.
hypertext links to the general categories above.

>

>> There should be some easily findable way to quit out of the help menu
>> documentaiton. Its almost as if someone set a booby trap.
>
> Doesn't "C-x 0" work?

Not here.  `?' switches to help menu full buffer. C-x 0 tells me 
`only one window'

There is the saving grace of the `l' command though.  Maybe if I spent
more time in the reader it would finally click into my fingers.