trying out icicles

trying out icicles

[Rustom Mody]

I started trying out icicles and I get:
(<S-iso-lefttab>, TAB: list, C-?: help)

What in the name of ??? is S-iso-lefttab ???
[I am posting this after trying google and emacs-info]

Re: trying out icicles

[Glauber Alex Dias Prado]

its shift-tab here, it depends on your terminal i think, btw on trying
icicles persistent sets i can only access then with a C-x b which is
bound to icicle-buffer(that is after i have defined and saved a named
set(project)), is this right, i thought i could access then by the means
of C-x C-f but it doesnt seem to work.

RE: trying out icicles

[Drew Adams]

> on trying icicles persistent sets i can only access then with
> a C-x b which is bound to icicle-buffer (that is after i have
> defined and saved a named set (project)), is this right,
> i thought i could access then by the means of C-x C-f but it
> doesnt seem to work.

I'm not sure I understand you.

You can define any set of completion candidates, of any type, as a "saved
completion set". This tells how:
http://www.emacswiki.org/cgi-bin/wiki/Icicles_-_Candidate_Sets

You can save any such set persistently (when you create it or later). This tells
how:
http://www.emacswiki.org/cgi-bin/wiki/Icicles_-_Persistent_Completions. (If a
"saved set" is not saved persistently, it is saved in a variable.)

You can retrieve any saved set (persistent or not) during completion for any
type of object (see the first URL for how). It makes sense to retrieve a set
only in a context where you can complete to names of objects of the type in the
set. For example, you could retrieve a set of command names when doing
buffer-name completion, but that wouldn't make much sense. In that context, it
makes sense to retrieve a set of buffer names. (You can have as many saved sets
as you like.)

Emacs file-name completion is special. The completion candidates are relative to
the current `default-directory'; they are not absolute file names. If you save a
set of these, it will look like this: ("foo.el" "bar.txt" "titi.el"). If you
retrieve such a candidate set when you are completing (e.g. with `C-x C-f'),
then you had better be in a directory that has such files, or else you will be
opening new files there with those names.

Icicles has two commands that do file-name completion differently from this.
They use absolute file names, so you can complete against any parts of the full
name (path). These commands are `icicle-locate-file' and `icicle-recent-file'
(plus `*-other-window' versions). If you save a set of completion candidates
when you use these commands, you save absolute file names. And the files whose
names you save need not be in the same directory. Example set:
("/projects/alpha/foo.el" "/projects/beta/foo.el" "/usr/local/whatever/titi.c").
This is what you want to do to create project file sets for different purposes.

If you save a set of buffer names when you use `icicle-buffer', the result is a
list of buffer-name strings, e.g. ("foo.el" "*Compile-Log*" "*grep*"). You can
retrieve this during file-name completion (e.g. `C-x C-f'), as you mentioned,
but unless you somehow have a buffer name that is an absolute file name, you
will have the same problem as for a set saved during `C-x C-f' (see above).

If you do have a saved set of absolute file names, you can retrieve it during
`C-x C-f'. That will work, even though `C-x C-f' uses relative names itself as
completion candidates. That is, even if the current `default-directory' is
unrelated to the directories in the saved absolute file names, the correct files
will be opened: an absolute file name means the `default-directory' is ignored.
(This is standard behavior for `C-x C-f'.)

If what I describe above is not what you see, and you think what you're seeing
is a bug, then please file a bug report: `M-x icicle-send-bug-report'.

Resp.: trying out icicles

[glauber M4]

>> on trying icicles persistent sets i can only access then with
>> a C-x b which is bound to icicle-buffer (that is after i have
>> defined and saved a named set (project)), is this right,
>> i thought i could access then by the means of C-x C-f but it
>> doesnt seem to work.
>
> I'm not sure I understand you.
>
You understood yes.

> You can define any set of completion candidates, of any type, as a "saved
> completion set". This tells how:
> http://www.emacswiki.org/cgi-bin/wiki/Icicles_-_Candidate_Sets
>
> You can save any such set persistently (when you create it or later). This tells
> how:
> http://www.emacswiki.org/cgi-bin/wiki/Icicles_-_Persistent_Completions. (If a
> "saved set" is not saved persistently, it is saved in a variable.)
>
> You can retrieve any saved set (persistent or not) during completion for any
> type of object (see the first URL for how). It makes sense to retrieve a set
> only in a context where you can complete to names of objects of the type in the
> set. For example, you could retrieve a set of command names when doing
> buffer-name completion, but that wouldn't make much sense. In that context, it
> makes sense to retrieve a set of buffer names. (You can have as many saved sets
> as you like.)
>
> Emacs file-name completion is special. The completion candidates are relative to
> the current `default-directory'; they are not absolute file names. If you save a
> set of these, it will look like this: ("foo.el" "bar.txt" "titi.el"). If you
> retrieve such a candidate set when you are completing (e.g. with `C-x C-f'),
> then you had better be in a directory that has such files, or else you will be
> opening new files there with those names.
>
> Icicles has two commands that do file-name completion differently from this.
> They use absolute file names, so you can complete against any parts of the full
> name (path). These commands are `icicle-locate-file' and `icicle-recent-file'
> (plus `*-other-window' versions). If you save a set of completion candidates
> when you use these commands, you save absolute file names. And the files whose
> names you save need not be in the same directory. Example set:
> ("/projects/alpha/foo.el" "/projects/beta/foo.el" "/usr/local/whatever/titi.c").
> This is what you want to do to create project file sets for different purposes.
>
> If you save a set of buffer names when you use `icicle-buffer', the result is a
> list of buffer-name strings, e.g. ("foo.el" "*Compile-Log*" "*grep*"). You can
> retrieve this during file-name completion (e.g. `C-x C-f'), as you mentioned,
> but unless you somehow have a buffer name that is an absolute file name, you
> will have the same problem as for a set saved during `C-x C-f' (see above).
>
> If you do have a saved set of absolute file names, you can retrieve it during
> `C-x C-f'. That will work, even though `C-x C-f' uses relative names itself as
> completion candidates. That is, even if the current `default-directory' is
> unrelated to the directories in the saved absolute file names, the correct files
> will be opened: an absolute file name means the `default-directory' is ignored.
> (This is standard behavior for `C-x C-f'.)
>
> If what I describe above is not what you see, and you think what you're seeing
> is a bug, then please file a bug report: `M-x icicle-send-bug-report'.
      I guess it is a bug, if i have a saved set and do a C-x C-f, C-{,
RET(to select the last set saved), I can even see the set of files that
was stored and if i click on them using the mouse before typing anything
they are going to open, but as soon as i type anything they disappear,
even if i type something like * to match everything no go, while they
keep and stay on the C-x b, but you are right it is kind of pointless
doing it cause the better you will get is a buffer named
/abal/alfjlk/anmlakjs/blakjsdf assuming that was the name of the
filepath+file, not the real file, anyway it is a great feature, I'm
going to send this bug report, I'm using latest icicles with latest
emacs.Except for this issue everything else you described is working
like you said.

cheers,

glauber.


2008/8/19, Drew Adams <drew.adams@oracle.com>:
>> on trying icicles persistent sets i can only access then with
>> a C-x b which is bound to icicle-buffer (that is after i have
>> defined and saved a named set (project)), is this right,
>> i thought i could access then by the means of C-x C-f but it
>> doesnt seem to work.
>
> I'm not sure I understand you.
>
> You can define any set of completion candidates, of any type, as a "saved
> completion set". This tells how:
> http://www.emacswiki.org/cgi-bin/wiki/Icicles_-_Candidate_Sets
>
> You can save any such set persistently (when you create it or later). This
> tells
> how:
> http://www.emacswiki.org/cgi-bin/wiki/Icicles_-_Persistent_Completions. (If
> a
> "saved set" is not saved persistently, it is saved in a variable.)
>
> You can retrieve any saved set (persistent or not) during completion for any
> type of object (see the first URL for how). It makes sense to retrieve a set
> only in a context where you can complete to names of objects of the type in
> the
> set. For example, you could retrieve a set of command names when doing
> buffer-name completion, but that wouldn't make much sense. In that context,
> it
> makes sense to retrieve a set of buffer names. (You can have as many saved
> sets
> as you like.)
>
> Emacs file-name completion is special. The completion candidates are
> relative to
> the current `default-directory'; they are not absolute file names. If you
> save a
> set of these, it will look like this: ("foo.el" "bar.txt" "titi.el"). If you
> retrieve such a candidate set when you are completing (e.g. with `C-x C-f'),
> then you had better be in a directory that has such files, or else you will
> be
> opening new files there with those names.
>
> Icicles has two commands that do file-name completion differently from this.
> They use absolute file names, so you can complete against any parts of the
> full
> name (path). These commands are `icicle-locate-file' and
> `icicle-recent-file'
> (plus `*-other-window' versions). If you save a set of completion candidates
> when you use these commands, you save absolute file names. And the files
> whose
> names you save need not be in the same directory. Example set:
> ("/projects/alpha/foo.el" "/projects/beta/foo.el"
> "/usr/local/whatever/titi.c").
> This is what you want to do to create project file sets for different
> purposes.
>
> If you save a set of buffer names when you use `icicle-buffer', the result
> is a
> list of buffer-name strings, e.g. ("foo.el" "*Compile-Log*" "*grep*"). You
> can
> retrieve this during file-name completion (e.g. `C-x C-f'), as you
> mentioned,
> but unless you somehow have a buffer name that is an absolute file name, you
> will have the same problem as for a set saved during `C-x C-f' (see above).
>
> If you do have a saved set of absolute file names, you can retrieve it
> during
> `C-x C-f'. That will work, even though `C-x C-f' uses relative names itself
> as
> completion candidates. That is, even if the current `default-directory' is
> unrelated to the directories in the saved absolute file names, the correct
> files
> will be opened: an absolute file name means the `default-directory' is
> ignored.
> (This is standard behavior for `C-x C-f'.)
>
> If what I describe above is not what you see, and you think what you're
> seeing
> is a bug, then please file a bug report: `M-x icicle-send-bug-report'.
>
>
>

RE: Resp.: trying out icicles

[Drew Adams]

> > If what I describe above is not what you see, and you think 
> > what you're seeing is a bug, then please file a bug report: `M-x 
> > icicle-send-bug-report'.
>
> I guess it is a bug, if i have a saved set and do a C-x 
> C-f, C-{, RET (to select the last set saved), I can even see the set of 
> files that was stored and if i click on them using the mouse before 
> typing anything they are going to open,

That means that they were correctly restored as the current set of completion
candidates.

> but as soon as i type anything they disappear,

Anything? What you type needs to match at least one of the candidates, or it
will say "No match".

> even if i type something like * to match everything no go,

`*' matches only names that contain the character `*'. Perhaps that's your
problem?

Empty input matches everything, and if you use apropos completion (`S-TAB') `.*'
matches everything. 

For apropos completion, the regexp `*' matches only names that contain `*'. For
prefix completion (`TAB'), `*' matches only names that start with `*'. 

You can of course use just `*' with `RET' to get normal Emacs file-name globbing
(`*' matches all file names in this context), but for completion `*' is a regexp
that matches `*'.

See
http://www.emacswiki.org/cgi-bin/wiki/Icicles_-_Special_Characters_in_Input_Patt
erns.

> I'm going to send this bug report, I'm using latest icicles with latest
> emacs.Except for this issue everything else you described is working
> like you said.

When you send the complete bug report, please give an exact recipe, step by
step, starting with emacs -Q.

Thx - Drew

RE: trying out icicles

[Drew Adams]

> I started trying out icicles and I get:
> (<S-iso-lefttab>, TAB: list, C-?: help)
> 
> What in the name of ??? is S-iso-lefttab ???
> [I am posting this after trying google and emacs-info]

1. Sounds like you have an old copy of Icicles. That text used to be added to
the prompt for your first few Icicles sessions, as a reminder, but that has not
been the case since March, 2008.

2. Would it make more sense if you saw `S-tab' instead of `S-iso-lefttab'? These
are both possible names for the key combination Shift + Tab. Different keyboards
send different logical keys (names) for the physical key combination Shift +
Tab.

3. When multiple keys (e.g. `S-tab' and `S-iso-lefttab') are bound to the same
command (e.g. `icicle-apropos-complete'), and a message prints the current
binding for the command (so the message will always be correct, even if you
rebind the command), only one of those multiple keys appears in the message, and
which one appears is not under program control. This is too bad, but it is the
way things are. The choices for the programmer are either hard-code a key name
(e.g. `S-TAB') in the message or put up with Emacs's pseudo-arbitrary choice.

Bottom line: Don't worry about this, but do download the latest version of
Icicles:
http://www.emacswiki.org/cgi-bin/wiki/Icicles_-_Libraries