>>>> Is it possible to complete zero length string with icicles
>>>> in other words , is it possible to have in the completion
>>>> buffer ALL the words avaialbe in ALL buffers and from there
>>>> starting filtering ?

>

>>> Not sure what you mean, exactly.
>>> But yes, if you type nothing in the minibuffer when it asks
>>> for completion input, and you then hit TAB or S-TAB, you
>>> should get all available completion candidates. This is a
>>> common action - it's how you see all possible candidates.
>>> If you type something, then what you type filters the
>>> possible candidates; otherwise, you see them all.

>

>>> However, I don't know what you mean by the last part: "ALL
>>> the words available in ALL buffers".  Minibuffer completion
>>> is available for lots of different kinds of object (buffer
>>> name, file name, function name, search hit,...). What is the
>>> context here? What command are you using?

>

>>> It sounds like you want to complete a word (in a buffer
>>> other than the minibuffer), using the words that are present
>>> in all buffers as candidates. Is that correct? Are you
>>> referring to `C-M-/', which (in Icicle mode) is
>>> `icicle-dabbrev-completion'? If so, use `C-h k C-M-/' to see
>>> the description of this command. In that description, you see
>>> that `C-u C-u C-M-/' uses all words
>>> from all buffers as candidates. (`C-h k' is your friend.)

>

>> Thanks for the detailed answer.
>> first off all, thanks for this amazing extention.
>> it is dream come true .

>

>> you are correct for my intention but i didnt understand the solution.

>

>> i want to use dabbrev of icicle and also to see ALL words in ALL
>> buffers in the completion list.
>> the use for this is obvious , suppose i do remeber only the middle of
>> word that exist in other buffer.

>

>> in my current situation i dont know how to "tell" icicle to show all
>> words in all buffers without a prefix.

>

>> if i use icicle-dabbrev-completion with no prefix i get the message "
>> no dynamic exit"
>> the same message occur with c-u c-u icicle-dabbrev-completion.

>
> I understand now. The Icicles version of `C-M-/', like the vanilla Emacs
> version, required you to have a non-empty prefix as a starting point, even
> though you could then use apropos (e.g. substring) completion to choose among
> the candidate completions.
>
> I've changed this to accommodate your expectation, which was reasonable. You can
> now complete immediately even from an empty string (e.g. at the beginning of a
> buffer). Try the latest version. HTH.- Hide quoted text -
>
> - Show quoted text -