How to speed up ido?

How to speed up ido?

[senny]

Hey there

I wrote a function to open any file in my project using ido. The
Problem is, that the project has a lot of files and a deeply nested
directory structure. When i then complete the file name in ido using
fuzzy-matching it is extremely slow and unusable. Is there a way to
increase the speed of ido? or should i take another approach to this
problem? I am currently using the method "ido-completing-read".

regards,
Senny

Re: How to speed up ido?

[Andreas Politz]

senny wrote:
> Hey there
> 
> I wrote a function to open any file in my project using ido. The
> Problem is, that the project has a lot of files and a deeply nested
> directory structure. When i then complete the file name in ido using
> fuzzy-matching it is extremely slow and unusable. Is there a way to
> increase the speed of ido? or should i take another approach to this
> problem? I am currently using the method "ido-completing-read".
> 
> regards,
> Senny


A while ago I wrote a mode called ido-hacks. As the name
suggests it contains a couple of more or less dirty
modifications to ido, among them :

* faster display of prospect (streamlined one loop)
* caching of already computed flex-matches
* use ido for every completing-read call

If you're interested I could mail/post it and you could maybe
extract the optimization parts (It's one advice and one redefinition.)
Btw, I am using it all the time.

-ap

Re: How to speed up ido?

[Seweryn Kokot]

Andreas Politz <politza@fh-trier.de> writes:

> senny wrote:
>> Hey there
>>
>> I wrote a function to open any file in my project using ido. The
>> Problem is, that the project has a lot of files and a deeply nested
>> directory structure. When i then complete the file name in ido using
>> fuzzy-matching it is extremely slow and unusable. Is there a way to
>> increase the speed of ido? or should i take another approach to this
>> problem? I am currently using the method "ido-completing-read".
>>
>> regards,
>> Senny
>
>
> A while ago I wrote a mode called ido-hacks. As the name
> suggests it contains a couple of more or less dirty
> modifications to ido, among them :
>
> * faster display of prospect (streamlined one loop)
> * caching of already computed flex-matches
> * use ido for every completing-read call
>
> If you're interested I could mail/post it and you could maybe
> extract the optimization parts (It's one advice and one redefinition.)
> Btw, I am using it all the time.
>
> -ap

Could you please send it here? I think there are more people
interested in it!

Thanks in advance.
-- 
regards,
Seweryn

Re: How to speed up ido?

[William Xu]

Andreas Politz <politza@fh-trier.de> writes:

> * use ido for every completing-read call

How do you accomplish this? I have to redefine or advice many
functions to achieve that.

-- 
William

http://xwl.appspot.com

Re: How to speed up ido?

[senny]

On 17 Jan., 14:09, Andreas Politz <poli...@fh-trier.de> wrote:
> Andreas Politz wrote:
> > Seweryn Kokot wrote:
> >> Andreas Politz <poli...@fh-trier.de> writes:
>
> >>> senny wrote:
> >>>> Hey there
>
> >>>> I wrote a function to open any file in my project using ido. The
> >>>> Problem is, that the project has a lot of files and a deeply nested
> >>>> directory structure. When i then complete the file name in ido using
> >>>> fuzzy-matching it is extremely slow and unusable. Is there a way to
> >>>> increase the speed of ido? or should i take another approach to this
> >>>> problem? I am currently using the method "ido-completing-read".
>
> >>>> regards,
> >>>> Senny
>
> >>> A while ago I wrote a mode called ido-hacks. As the name
> >>> suggests it contains a couple of more or less dirty
> >>> modifications to ido, among them :
>
> >>> * faster display of prospect (streamlined one loop)
> >>> * caching of already computed flex-matches
> >>> * use ido for every completing-read call
>
> >>> If you're interested I could mail/post it and you could maybe
> >>> extract the optimization parts (It's one advice and one redefinition.)
> >>> Btw, I am using it all the time.
>
> >>> -ap
>
> >> Could you please send it here? I think there are more people
> >> interested in it!
>
> >> Thanks in advance.
>
> > I attached it to this mail. Compiling is suggested. Load and enable it
> > like this:
>
> > (ido-mode 1)
> > (ido-hacks-mode 1)
>
> > -ap
>
> I have send the wrong version....
>
> -ap
>
>  ido-hacks.el.gz
> 4KAnzeigenHerunterladen

Thanks a lot. Im gonna try your modifications as soon as possible.

Re: How to speed up ido?

[Seweryn Kokot]

senny <yves.senn@gmail.com> writes:

> On 17 Jan., 14:09, Andreas Politz <poli...@fh-trier.de> wrote:
>> Andreas Politz wrote:
>> > Seweryn Kokot wrote:
>> >> Andreas Politz <poli...@fh-trier.de> writes:
>>
>> >>> senny wrote:
>> >>>> Hey there
>>
>> >>>> I wrote a function to open any file in my project using ido. The
>> >>>> Problem is, that the project has a lot of files and a deeply nested
>> >>>> directory structure. When i then complete the file name in ido using
>> >>>> fuzzy-matching it is extremely slow and unusable. Is there a way to
>> >>>> increase the speed of ido? or should i take another approach to this
>> >>>> problem? I am currently using the method "ido-completing-read".
>>
>> >>>> regards,
>> >>>> Senny
>>
>> >>> A while ago I wrote a mode called ido-hacks. As the name
>> >>> suggests it contains a couple of more or less dirty
>> >>> modifications to ido, among them :
>>
>> >>> * faster display of prospect (streamlined one loop)
>> >>> * caching of already computed flex-matches
>> >>> * use ido for every completing-read call
>>
>> >>> If you're interested I could mail/post it and you could maybe
>> >>> extract the optimization parts (It's one advice and one redefinition.)
>> >>> Btw, I am using it all the time.
>>
>> >>> -ap
>>
>> >> Could you please send it here? I think there are more people
>> >> interested in it!
>>
>> >> Thanks in advance.
>>
>> > I attached it to this mail. Compiling is suggested. Load and enable it
>> > like this:
>>
>> > (ido-mode 1)
>> > (ido-hacks-mode 1)
>>
>> > -ap
>>
>> I have send the wrong version....
>>
>> -ap
>>
>>  ido-hacks.el.gz
>> 4KAnzeigenHerunterladen
>
> Thanks a lot. Im gonna try your modifications as soon as possible.

It seems that posts from Andreas didn't arrived to emacs help group

http://news.gmane.org/gmane.emacs.help
http://lists.gnu.org/archive/html/help-gnu-emacs/2009-01/threads.html

Any idea what is wrong?

-- 
regards,
Seweryn

Re: How to speed up ido?

[Andreas Politz]

Andreas Politz wrote:
> Andreas Politz wrote:
>> Seweryn Kokot wrote:
>>> Andreas Politz <politza@fh-trier.de> writes:
>>>
>>>> senny wrote:
>>>>> Hey there
>>>>>
>>>>> I wrote a function to open any file in my project using ido. The
>>>>> Problem is, that the project has a lot of files and a deeply nested
>>>>> directory structure. When i then complete the file name in ido using
>>>>> fuzzy-matching it is extremely slow and unusable. Is there a way to
>>>>> increase the speed of ido? or should i take another approach to this
>>>>> problem? I am currently using the method "ido-completing-read".
>>>>>
>>>>> regards,
>>>>> Senny
>>>>
>>>> A while ago I wrote a mode called ido-hacks. As the name
>>>> suggests it contains a couple of more or less dirty
>>>> modifications to ido, among them :
>>>>
>>>> * faster display of prospect (streamlined one loop)
>>>> * caching of already computed flex-matches
>>>> * use ido for every completing-read call
>>>>
>>>> If you're interested I could mail/post it and you could maybe
>>>> extract the optimization parts (It's one advice and one redefinition.)
>>>> Btw, I am using it all the time.
>>>>
>>>> -ap
>>>
>>> Could you please send it here? I think there are more people
>>> interested in it!
>>>
>>> Thanks in advance.
>>
>> I attached it to this mail. Compiling is suggested. Load and enable it 
>> like this:
>>
>> (ido-mode 1)
>> (ido-hacks-mode 1)
>>
>>
>> -ap
> 
> I have send the wrong version....
> 
> -ap

Since noone seems to be able to receive the attached file, I have put it
on the web :

http://www0.fh-trier.de/~politza/emacs/ido-hacks.el.gz

-ap

Re: How to speed up ido?

[David Combs]

Then why not add it to the e-wiki?

David