Emacs and C++ codesense

Emacs and C++ codesense

[spamfilteraccount]

I'm looking for solutions for working with C++ code in Emacs.

I'd like the following:

- Jumping to the definition of the symbol I'm standing on when I press
a key. I know tags is useful for that, but they don''t seem to work
properly in every case, only in most of the cases and I'd like a
proper solution.

- Show the documentation of the current symbol in a tooltip when
pressing a key. The tooltip is not a big deal. I can do it if I can
get the documentation somehow.

- Inteligent completion of the current symbol when I press a key.

I don't want ECB, because I don't want separate windows within the
frame showing information. I like to work in a single frame and get
information only if I press a key explicitly.

I suspect semantic can do some or all of the above. Is it true? Is
there a solution out of the box which works like described above or I
have to learn how the semantic api works and implement my own
solution? How good is semantic's C++ parsing/support?

Are there other alternatives?

xref seems nice, but I'd like to evaluate free software first.

I'm also checking out kdevelop which I hear has good support for C++,
but I really don't want to leave emacs (iswitchb!) if I can find a
satisfactory solution.

Thanks.

Re: Emacs and C++ codesense

[Eli Zaretskii]

> From: "spamfilteraccount@gmail.com" <spamfilteraccount@gmail.com>
> Date: 25 Apr 2007 07:56:38 -0700
> 
> - Jumping to the definition of the symbol I'm standing on when I press
> a key. I know tags is useful for that, but they don''t seem to work
> properly in every case, only in most of the cases and I'd like a
> proper solution.

Tags work for me; if they don't for you, please give an example.

There's also Ebrowse, which is more C++-aware.

> - Show the documentation of the current symbol in a tooltip when
> pressing a key. The tooltip is not a big deal. I can do it if I can
> get the documentation somehow.

Try "C-h S".

> - Inteligent completion of the current symbol when I press a key.

Is "M-/" intelligent enough for you?  If not, please define
``intelligent''.

Re: Emacs and C++ codesense

[Maciej Katafiasz]

Den Wed, 25 Apr 2007 07:56:38 -0700 skrev spamfilteraccount@gmail.com:

> I'm looking for solutions for working with C++ code in Emacs.
> 
> I'd like the following:
> 
> - Jumping to the definition of the symbol I'm standing on when I press
> a key. I know tags is useful for that, but they don''t seem to work
> properly in every case, only in most of the cases and I'd like a
> proper solution.

I think you'll have to stick with tags, just make sure you use
exuberant-ctags, it has much more fancy support for things. A solution
based on semantic is possible, but not available as of yet.

> - Show the documentation of the current symbol in a tooltip when
> pressing a key. The tooltip is not a big deal. I can do it if I can
> get the documentation somehow.

See semantic-idle-completion-mode, semantic-ia-show-doc, although that
depends on how much info you want. You do get the prototype though.

> - Inteligent completion of the current symbol when I press a key.
> 
> I don't want ECB, because I don't want separate windows within the
> frame showing information. I like to work in a single frame and get
> information only if I press a key explicitly.

That's good, since ECB doesn't concern itself with completion :).

> I suspect semantic can do some or all of the above. Is it true? Is
> there a solution out of the box which works like described above or I
> have to learn how the semantic api works and implement my own
> solution? How good is semantic's C++ parsing/support?

Semantic should fit the bill. There's semantic-ia-complete-symbol[-menu],
and also semantic-hover-completion.el, which I'm going to post in a
minute. I'm not exactly sure how much stress the C++ support can stand, as
I don't code C++, but basic testing says it works pretty well.

Cheers,
Maciej

Re: Emacs and C++ codesense

[spamfilteraccount]

On Apr 25, 8:24 pm, Eli Zaretskii <e...@gnu.org> wrote:
>
> > - Jumping to the definition of the symbol I'm standing on when I press
> > a key. I know tags is useful for that, but they don''t seem to work
> > properly in every case, only in most of the cases and I'd like a
> > proper solution.
>
> Tags work for me; if they don't for you, please give an example.

I used Emacs' etags first. Moving to exuberant ctags solved the
problem.

> > - Show the documentation of the current symbol in a tooltip when
> > pressing a key. The tooltip is not a big deal. I can do it if I can
> > get the documentation somehow.
>
> Try "C-h S".

describe-syntax? Are you sure you meant that?

I was thinking of something like this:

http://esense.sourceforge.net/shot2.jpg

> > - Inteligent completion of the current symbol when I press a key.
>
> Is "M-/" intelligent enough for you?  If not, please define
> ``intelligent''.

  instance.<pressing completion key>

Only those completions should be offered which make sense in the scope
of the object (data members, member functions, inherited member
functions, etc.)

Re: Emacs and C++ codesense

[Eli Zaretskii]

> From: "spamfilteraccount@gmail.com" <spamfilteraccount@gmail.com>
> Date: 26 Apr 2007 00:43:41 -0700
> 
> > Try "C-h S".
> 
> describe-syntax? Are you sure you meant that?

You didn't say which Emacs version you had, so I guessed, and my guess
was wrong.  I now think you use v21.x, where the command I meant is
bound to C-h TAB.

> I was thinking of something like this:
> 
> http://esense.sourceforge.net/shot2.jpg

If that is what you want, why not download esense and use it?

> > > - Inteligent completion of the current symbol when I press a key.
> >
> > Is "M-/" intelligent enough for you?  If not, please define
> > ``intelligent''.
> 
>   instance.<pressing completion key>
> 
> Only those completions should be offered which make sense in the scope
> of the object (data members, member functions, inherited member
> functions, etc.)

I think you want "C-c C-m TAB" in ebrowse.el.  Read all about that in
the Ebrowse manual (which comes with Emacs).

Re: Emacs and C++ codesense

[spamfilteraccount]

On Apr 26, 10:14 pm, Eli Zaretskii <e...@gnu.org> wrote:
>
> > I was thinking of something like this:
>
> >http://esense.sourceforge.net/shot2.jpg
>
> If that is what you want, why not download esense and use it?
>

Because it's for Erlang which is not exactly related to C++.

Re: Emacs and C++ codesense

[Hadron]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: "spamfilteraccount@gmail.com" <spamfilteraccount@gmail.com>
>> Date: 26 Apr 2007 00:43:41 -0700
>> 
>> > Try "C-h S".
>> 
>> describe-syntax? Are you sure you meant that?
>
> You didn't say which Emacs version you had, so I guessed, and my guess
> was wrong.  I now think you use v21.x, where the command I meant is
> bound to C-h TAB.

What does C-h TAB do? Can you please describe the function names as key
bindings change.

>
>> I was thinking of something like this:
>> 
>> http://esense.sourceforge.net/shot2.jpg
>
> If that is what you want, why not download esense and use it?

What has erlang got to do with C++?

>
>> > > - Inteligent completion of the current symbol when I press a key.
>> >
>> > Is "M-/" intelligent enough for you?  If not, please define
>> > ``intelligent''.
>> 
>>   instance.<pressing completion key>
>> 
>> Only those completions should be offered which make sense in the scope
>> of the object (data members, member functions, inherited member
>> functions, etc.)
>
> I think you want "C-c C-m TAB" in ebrowse.el.  Read all about that in
> the Ebrowse manual (which comes with Emacs).

Again - please list function names. It really would help.

Re: Emacs and C++ codesense

[Eli Zaretskii]

> From: Hadron <hadronquark@gmail.com>
> Date: Mon, 30 Apr 2007 01:16:30 +0200
> 
> >> > Try "C-h S".
> >> 
> >> describe-syntax? Are you sure you meant that?
> >
> > You didn't say which Emacs version you had, so I guessed, and my guess
> > was wrong.  I now think you use v21.x, where the command I meant is
> > bound to C-h TAB.
> 
> What does C-h TAB do? Can you please describe the function names as key
> bindings change.

The function is info-lookup-symbol.

> > I think you want "C-c C-m TAB" in ebrowse.el.  Read all about that in
> > the Ebrowse manual (which comes with Emacs).
> 
> Again - please list function names. It really would help.

It's ebrowse-tags-complete-symbol.

But please note that by insisting on function names like that you
abuse the help I give here to some extent.  The bindings I mentioned
are the default bindings, not my personal bindings.  Even though key
bindings may be different in your setup, you can always fire up
"emacs -q --no-site-file" to see what are the default bindings.  And
in the last case, where you need to load a library to have the
command, I named that library, and even pointed to the appropriate
manual.  It is not nice to require that I also write the function
names explicitly, as the information I gave originally already
provides you everything you need to find those commands by yourself.
This is not a hot help line, and you are not paying me for my time,
you know.  Please be ready to invest some of your own time and some
minimal effort into digesting the help you are given.

Re: Emacs and C++ codesense

[Hadron]

Eli Zaretskii <eliz@gnu.org> writes:

>> From: "spamfilteraccount@gmail.com" <spamfilteraccount@gmail.com>
>> Date: 25 Apr 2007 07:56:38 -0700
>> 
>> - Jumping to the definition of the symbol I'm standing on when I press
>> a key. I know tags is useful for that, but they don''t seem to work
>> properly in every case, only in most of the cases and I'd like a
>> proper solution.
>
> Tags work for me; if they don't for you, please give an example.
>
> There's also Ebrowse, which is more C++-aware.
>
>> - Show the documentation of the current symbol in a tooltip when
>> pressing a key. The tooltip is not a big deal. I can do it if I can
>> get the documentation somehow.
>
> Try "C-h S".
>
>> - Inteligent completion of the current symbol when I press a key.
>
> Is "M-/" intelligent enough for you?  If not, please define
> ``intelligent''.

What is "M-/"? Not all setups have the same key bindings.

Re: Emacs and C++ codesense

[Eli Zaretskii]

> From: Hadron <hadronquark@gmail.com>
> Date: Mon, 30 Apr 2007 01:13:35 +0200
> 
> What is "M-/"? Not all setups have the same key bindings.

M-/ is the default binding of dabbrev-expand.

Re: Emacs and C++ codesense

[Hadron]

"spamfilteraccount@gmail.com" <spamfilteraccount@gmail.com> writes:

> I'm looking for solutions for working with C++ code in Emacs.
>
> I'd like the following:
>
> - Jumping to the definition of the symbol I'm standing on when I press
> a key. I know tags is useful for that, but they don''t seem to work
> properly in every case, only in most of the cases and I'd like a
> proper solution.

cscope is the best for this IMO. Well, it is for C .....

>
> - Show the documentation of the current symbol in a tooltip when
> pressing a key. The tooltip is not a big deal. I can do it if I can
> get the documentation somehow.
>
> - Inteligent completion of the current symbol when I press a key.
>
> I don't want ECB, because I don't want separate windows within the
> frame showing information. I like to work in a single frame and get
> information only if I press a key explicitly.

yes. ECB is a bit in your face I agree.

>
> I suspect semantic can do some or all of the above. Is it true? Is
> there a solution out of the box which works like described above or I
> have to learn how the semantic api works and implement my own
> solution? How good is semantic's C++ parsing/support?
>
> Are there other alternatives?
>
> xref seems nice, but I'd like to evaluate free software first.
>
> I'm also checking out kdevelop which I hear has good support for C++,
> but I really don't want to leave emacs (iswitchb!) if I can find a
> satisfactory solution.

I used hippy-expand for the expansion! No tooltips, but cscope worked
for me. I never used it for c++, but I believe it works ok.

>
> Thanks.
>

-- 

Re: Emacs and C++ codesense

[Klaus Berndl]

On Mon, 30 Apr 2007, hadronquark@gmail.com wrote:

>  "spamfilteraccount@gmail.com" <spamfilteraccount@gmail.com> writes:
>  
> > I'm looking for solutions for working with C++ code in Emacs.
> >
> > I'd like the following:
> >
> > - Jumping to the definition of the symbol I'm standing on when I press
> > a key. I know tags is useful for that, but they don''t seem to work
> > properly in every case, only in most of the cases and I'd like a
> > proper solution.
>  
>  cscope is the best for this IMO. Well, it is for C .....
>  
> >
> > - Show the documentation of the current symbol in a tooltip when
> > pressing a key. The tooltip is not a big deal. I can do it if I can
> > get the documentation somehow.
> >
> > - Inteligent completion of the current symbol when I press a key.
> >
> > I don't want ECB, because I don't want separate windows within the
> > frame showing information. I like to work in a single frame and get
> > information only if I press a key explicitly.
>  
>  yes. ECB is a bit in your face I agree.

Thanks a lot ;-)

>  
> >
> > I suspect semantic can do some or all of the above. Is it true? Is
> > there a solution out of the box which works like described above or I
> > have to learn how the semantic api works and implement my own
> > solution? How good is semantic's C++ parsing/support?
> >
> > Are there other alternatives?
> >
> > xref seems nice, but I'd like to evaluate free software first.
> >
> > I'm also checking out kdevelop which I hear has good support for C++,
> > but I really don't want to leave emacs (iswitchb!) if I can find a
> > satisfactory solution.
>  
>  I used hippy-expand for the expansion! No tooltips, but cscope worked
>  for me. I never used it for c++, but I believe it works ok.
>  
> >
> > Thanks.
> >

-- 
Klaus Berndl			mailto: klaus.berndl@sdm.de
sd&m AG				http://www.sdm.de
software design & management	
Carl-Wery-Str. 42, 81739 Muenchen, Germany
Tel +49 89 63812-392, Fax -220

Re: Emacs and C++ codesense

[Hadron]

Klaus Berndl <klaus.berndl@sdm.de> writes:

> On Mon, 30 Apr 2007, hadronquark@gmail.com wrote:
>
>>  "spamfilteraccount@gmail.com" <spamfilteraccount@gmail.com> writes:
>>  
>> > I'm looking for solutions for working with C++ code in Emacs.
>> >
>> > I'd like the following:
>> >
>> > - Jumping to the definition of the symbol I'm standing on when I press
>> > a key. I know tags is useful for that, but they don''t seem to work
>> > properly in every case, only in most of the cases and I'd like a
>> > proper solution.
>>  
>>  cscope is the best for this IMO. Well, it is for C .....
>>  
>> >
>> > - Show the documentation of the current symbol in a tooltip when
>> > pressing a key. The tooltip is not a big deal. I can do it if I can
>> > get the documentation somehow.
>> >
>> > - Inteligent completion of the current symbol when I press a key.
>> >
>> > I don't want ECB, because I don't want separate windows within the
>> > frame showing information. I like to work in a single frame and get
>> > information only if I press a key explicitly.
>>  
>>  yes. ECB is a bit in your face I agree.
>
> Thanks a lot ;-)

No insult intended - but it takes a lot of configuring and learning to
get familiar with it. Even the very first design issue you hit - why
doesn't "switch to other buffer" include the ecb windows? After a while I
appreciated this touch. But not when I started :-; I find that the
overhead of the ascii character window splits takes too much real estate
from the editor - yes I know you can expand it .... Also, I never got
speedbar working properly with it.

In fact, I intend to be trying ecb out again soon - it will be
interesting to see how well the issues with winman integration are now.

It is still confusing, btw, on what ecb to use. The "best" way seems to
be from cedet, but cedet seems to be ancient. Could you maybe post a
"how to get up and running" update so someone can get the latest all
working and integrated in a few easy steps?

>
>>  
>> >
>> > I suspect semantic can do some or all of the above. Is it true? Is
>> > there a solution out of the box which works like described above or I
>> > have to learn how the semantic api works and implement my own
>> > solution? How good is semantic's C++ parsing/support?
>> >
>> > Are there other alternatives?
>> >
>> > xref seems nice, but I'd like to evaluate free software first.
>> >
>> > I'm also checking out kdevelop which I hear has good support for C++,
>> > but I really don't want to leave emacs (iswitchb!) if I can find a
>> > satisfactory solution.
>>  
>>  I used hippy-expand for the expansion! No tooltips, but cscope worked
>>  for me. I never used it for c++, but I believe it works ok.
>>  
>> >
>> > Thanks.
>> >

-- 

Re: Emacs and C++ codesense

[Maciej Katafiasz]

Den Thu, 03 May 2007 12:07:17 +0200 skrev Hadron:

> It is still confusing, btw, on what ecb to use. The "best" way seems to
> be from cedet, but cedet seems to be ancient. Could you maybe post a
> "how to get up and running" update so someone can get the latest all
> working and integrated in a few easy steps?

Huh? What do you mean? Go to ECB's website, everything is described there
in great detail. And there's no ECB included in CEDET, rather it's a
dependency of ECB.

Cheers,
Maciej

Re: Emacs and C++ codesense

[Hadron]

Maciej Katafiasz <mathrick@gmail.com> writes:

> Den Thu, 03 May 2007 12:07:17 +0200 skrev Hadron:
>
>> It is still confusing, btw, on what ecb to use. The "best" way seems to
>> be from cedet, but cedet seems to be ancient. Could you maybe post a
>> "how to get up and running" update so someone can get the latest all
>> working and integrated in a few easy steps?
>
> Huh? What do you mean? Go to ECB's website, everything is described there
> in great detail. And there's no ECB included in CEDET, rather it's a
> dependency of ECB.

Apologies. I was on crack when I posted that. I was referring more to
the links between the two. Or meant to. In addition my comments about
ecb were unfair as I was or meant to refer to the confusion with cedet
version more than ecb.

Actually, the whole bunch of cedet/winring/ecb does confuse me. I never
quite got the hang or the point of semantic either..... yes, I know its part of
cedet....

Re: Emacs and C++ codesense

[Maciej Katafiasz]

Den Fri, 04 May 2007 14:36:26 +0200 skrev Hadron:

> Actually, the whole bunch of cedet/winring/ecb does confuse me. I never
> quite got the hang or the point of semantic either..... yes, I know its
> part of cedet....

CEDET, as the name suggests, is a collection of various tools, connected
mostly by the fact they're developed together, and some depend on the
others. Semantic is the most prominent part of CEDET, which gives the
access to a complete language parser infrastructure, and subsequently to
its output, allowing for construction of language-sensitive tools (for
example, ones that know what variables are visible at any given point in
the source) without tying them to any particular language.

ECB is a collection of tools supposed to give better code browsing (hence
the name) tools, as opposed to the ones aimed at writing code.

Winring is not tied or related at all to either of the two, just that ECB
has explicit support for it because it does some heavy windows-management
magic.

Re: Emacs and C++ codesense

[Hadron]

Maciej Katafiasz <mathrick@gmail.com> writes:

> Den Fri, 04 May 2007 14:36:26 +0200 skrev Hadron:
>
>> Actually, the whole bunch of cedet/winring/ecb does confuse me. I never
>> quite got the hang or the point of semantic either..... yes, I know its
>> part of cedet....
>
> CEDET, as the name suggests, is a collection of various tools, connected
> mostly by the fact they're developed together, and some depend on the
> others. Semantic is the most prominent part of CEDET, which gives the
> access to a complete language parser infrastructure, and subsequently to
> its output, allowing for construction of language-sensitive tools (for
> example, ones that know what variables are visible at any given point in
> the source) without tying them to any particular language.

Yes - I read that. I just don't see how its useful in C development. I'm
not saying it is not useful. Just that I don't see what its for. Mea
culpa. The code navigation doesn't seem to do much and the decoration
modes are not particularly helpful. I guess it could do with some sort
of tutorial as I played with it a while and didn't really see what
benefits it gave. Maybe you could describe how you use it?

>
> ECB is a collection of tools supposed to give better code browsing (hence
> the name) tools, as opposed to the ones aimed at writing code.

Yes, I used it for a while but found the locals/files stuff to be of
limited use.

>
> Winring is not tied or related at all to either of the two, just that ECB
> has explicit support for it because it does some heavy windows-management
> magic.

I realise that, but its kind of a must have when switching between
things like ecb and gnus IMO.

-- 

Re: Emacs and C++ codesense

[Maciej Katafiasz]

Den Fri, 04 May 2007 17:15:39 +0200 skrev Hadron:

>> CEDET, as the name suggests, is a collection of various tools, connected
>> mostly by the fact they're developed together, and some depend on the
>> others. Semantic is the most prominent part of CEDET, which gives the
>> access to a complete language parser infrastructure, and subsequently to
>> its output, allowing for construction of language-sensitive tools (for
>> example, ones that know what variables are visible at any given point in
>> the source) without tying them to any particular language.
> 
> Yes - I read that. I just don't see how its useful in C development. I'm
> not saying it is not useful. Just that I don't see what its for. Mea
> culpa. The code navigation doesn't seem to do much and the decoration
> modes are not particularly helpful. I guess it could do with some sort
> of tutorial as I played with it a while and didn't really see what
> benefits it gave. Maybe you could describe how you use it?

http://mathrick.org/files/emacs-demo.gif

Notice how it very clearly has an idea what's going on in the C code and
reacts to that.

>>
>> ECB is a collection of tools supposed to give better code browsing (hence
>> the name) tools, as opposed to the ones aimed at writing code.
> 
> Yes, I used it for a while but found the locals/files stuff to be of
> limited use.

Locals are really useful for getting an overview what's there in the file,
and for jumps. Files stuff I don't use too much indeed. 
 
>> Winring is not tied or related at all to either of the two, just that
>> ECB has explicit support for it because it does some heavy
>> windows-management magic.
> 
> I realise that, but its kind of a must have when switching between
> things like ecb and gnus IMO.

Not using Gnus, and even if, I use separate frames for clearly separate
layouts. So far I didn't feel the need to optimise layout management
enough to take the time to evaluate and choose one of the window managers.

Cheers,
Maciej

Re: Emacs and C++ codesense

[Hadron]

Maciej Katafiasz <mathrick@gmail.com> writes:

> Den Fri, 04 May 2007 17:15:39 +0200 skrev Hadron:
>
>>> CEDET, as the name suggests, is a collection of various tools, connected
>>> mostly by the fact they're developed together, and some depend on the
>>> others. Semantic is the most prominent part of CEDET, which gives the
>>> access to a complete language parser infrastructure, and subsequently to
>>> its output, allowing for construction of language-sensitive tools (for
>>> example, ones that know what variables are visible at any given point in
>>> the source) without tying them to any particular language.
>> 
>> Yes - I read that. I just don't see how its useful in C development. I'm
>> not saying it is not useful. Just that I don't see what its for. Mea
>> culpa. The code navigation doesn't seem to do much and the decoration
>> modes are not particularly helpful. I guess it could do with some sort
>> of tutorial as I played with it a while and didn't really see what
>> benefits it gave. Maybe you could describe how you use it?
>
> http://mathrick.org/files/emacs-demo.gif

Page not found.

Re: Emacs and C++ codesense

[Patrick Drechsler]

Maciej Katafiasz <mathrick@gmail.com> writes:

> http://mathrick.org/files/emacs-demo.gif

Nice demo. How did you create this?

Cheers

Patrick
-- 
By working faithfully eight hours a day,
you may eventually get to be boss and work twelve.
		-- Robert Frost

Re: Emacs and C++ codesense

[Hadron]

Patrick Drechsler <patrick@pdrechsler.de> writes:

> Maciej Katafiasz <mathrick@gmail.com> writes:
>
>> http://mathrick.org/files/emacs-demo.gif
>
> Nice demo. How did you create this?
>
> Cheers
>
> Patrick

Hmm I can get the page now.

This is all good but how did you do these things? Do you have
accompanying instructions/tutorial?

Re: Emacs and C++ codesense

[Maciej Katafiasz]

Den Fri, 04 May 2007 18:24:32 +0200 skrev Hadron:

>>> http://mathrick.org/files/emacs-demo.gif
>>
>> Nice demo. How did you create this?
> 
> Hmm I can get the page now.
> 
> This is all good but how did you do these things? Do you have
> accompanying instructions/tutorial?

Not in any organised manner, no. It's just a demonstration of some simple
(more or less) functions of several packages I use, amongst them semantic
and related tools. The tooltip completion you can see I posted not so long
ago to emacs-sources, it's called semantic-hover-completion.el. Version
0.1 is very buggy, though, so it's not really very usable for anyone who
isn't me. Unfortunately, I won't have the time to fix that in the
immediate feature.

Cheers,
Maciej

Re: Emacs and C++ codesense

[Hadron]

Maciej Katafiasz <mathrick@gmail.com> writes:

> Den Fri, 04 May 2007 18:24:32 +0200 skrev Hadron:
>
>>>> http://mathrick.org/files/emacs-demo.gif
>>>
>>> Nice demo. How did you create this?
>> 
>> Hmm I can get the page now.
>> 
>> This is all good but how did you do these things? Do you have
>> accompanying instructions/tutorial?
>
> Not in any organised manner, no. It's just a demonstration of some simple
> (more or less) functions of several packages I use, amongst them semantic
> and related tools. The tooltip completion you can see I posted not so long
> ago to emacs-sources, it's called semantic-hover-completion.el. Version
> 0.1 is very buggy, though, so it's not really very usable for anyone who
> isn't me. Unfortunately, I won't have the time to fix that in the
> immediate feature.

ok, so it wasn't a demo of what semantic can do. It was a mish mash of
different things not likely to be reproduced by anyone but yourself :-;

Why did you do your own completion? What was wrong with intellisense?
Have you used intellisense?

Re: Emacs and C++ codesense

[Maciej Katafiasz]

Den Fri, 04 May 2007 23:04:41 +0200 skrev Hadron:

>>> This is all good but how did you do these things? Do you have
>>> accompanying instructions/tutorial?
>>
>> Not in any organised manner, no. It's just a demonstration of some simple
>> (more or less) functions of several packages I use, amongst them semantic
>> and related tools. The tooltip completion you can see I posted not so long
>> ago to emacs-sources, it's called semantic-hover-completion.el. Version
>> 0.1 is very buggy, though, so it's not really very usable for anyone who
>> isn't me. Unfortunately, I won't have the time to fix that in the
>> immediate feature.
> 
> ok, so it wasn't a demo of what semantic can do. It was a mish mash of
> different things not likely to be reproduced by anyone but yourself :-;

But of course, large parts of it were due to semantic usage. You asked how
it can fit C coding, this is my answer.

> Why did you do your own completion? What was wrong with intellisense?
> Have you used intellisense?

Uhh, this *is* intellisense. At least as long as you regard "intellisense"
as a generic description and not trademarked name by MSFT. I wrote it
because the existing completion popups for semantic were not close enonugh
to the original IntelliSense, or rather, diverged in ways that seriously
hindered their usefulness.

Cheers,
Maciej

Re: Emacs and C++ codesense

[Hadron]

Maciej Katafiasz <mathrick@gmail.com> writes:

> Den Fri, 04 May 2007 23:04:41 +0200 skrev Hadron:
>
>>>> This is all good but how did you do these things? Do you have
>>>> accompanying instructions/tutorial?
>>>
>>> Not in any organised manner, no. It's just a demonstration of some simple
>>> (more or less) functions of several packages I use, amongst them semantic
>>> and related tools. The tooltip completion you can see I posted not so long
>>> ago to emacs-sources, it's called semantic-hover-completion.el. Version
>>> 0.1 is very buggy, though, so it's not really very usable for anyone who
>>> isn't me. Unfortunately, I won't have the time to fix that in the
>>> immediate feature.
>> 
>> ok, so it wasn't a demo of what semantic can do. It was a mish mash of
>> different things not likely to be reproduced by anyone but yourself :-;
>
> But of course, large parts of it were due to semantic usage. You asked how
> it can fit C coding, this is my answer.

But which bits? I have no idea which bits were semantic and which others
were part of your "mish mash of other packages".

>
>> Why did you do your own completion? What was wrong with intellisense?
>> Have you used intellisense?
>
> Uhh, this *is* intellisense. At least as long as you regard
> "intellisense"

Well, when you say your wrote your own completion package, how is one
supposed to know the difference?

> as a generic description and not trademarked name by MSFT. I wrote it
> because the existing completion popups for semantic were not close enonugh
> to the original IntelliSense, or rather, diverged in ways that seriously
> hindered their usefulness.

I turn the pop ups off and use tab cycle completion - works quite
nicely.

But again, could U request some sort of explanation of the demo gif?
Just looking at it tells me nothing whatsoever about how you used
semantic to do a job of work.

>
> Cheers,
> Maciej
>
>
>

-- 

Re: Emacs and C++ codesense

[Maciej Katafiasz]

Den Sat, 05 May 2007 12:18:58 +0200 skrev Hadron:

>>> ok, so it wasn't a demo of what semantic can do. It was a mish mash of
>>> different things not likely to be reproduced by anyone but yourself :-;
>>
>> But of course, large parts of it were due to semantic usage. You asked how
>> it can fit C coding, this is my answer.
> 
> But which bits? I have no idea which bits were semantic and which others
> were part of your "mish mash of other packages".

Semantic shines through whenever you can see emacs visibly having an idea
about the code's meaning. In this demo, it means 1) the suggestion popus
2) the second time I use `fwritex' template, in which FILE *STREAM formal
parameter is replaced by a choice of all currently visible variables of
type FILE*

>>> Why did you do your own completion? What was wrong with intellisense?
>>> Have you used intellisense?
>>
>> Uhh, this *is* intellisense. At least as long as you regard
>> "intellisense"
> 
> Well, when you say your wrote your own completion package, how is one
> supposed to know the difference?

"intellisense" is either 1) IntelliSense, a trademarked name for MSVS
context-sensitive completion, in which case none of the above is
intellisense 2) a generic name for a class of similarly-working packages,
in which case both are implementations of intellisense. If you want to
know the difference, download the package and read the description 

>> as a generic description and not trademarked name by MSFT. I wrote it
>> because the existing completion popups for semantic were not close
>> enonugh to the original IntelliSense, or rather, diverged in ways that
>> seriously hindered their usefulness.
> 
> I turn the pop ups off and use tab cycle completion - works quite
> nicely.
> 
> But again, could U request some sort of explanation of the demo gif?
> Just looking at it tells me nothing whatsoever about how you used
> semantic to do a job of work.

Not really, I don't have the time for that now. I intend to turn it into a
properly integrated and easy to install and setup package, but it hasn't
happened yet.

Cheers,
Maciej

Re: Emacs and C++ codesense

[Hadron]

Maciej Katafiasz <mathrick@gmail.com> writes:

> Den Sat, 05 May 2007 12:18:58 +0200 skrev Hadron:
>
>>>> ok, so it wasn't a demo of what semantic can do. It was a mish mash of
>>>> different things not likely to be reproduced by anyone but yourself :-;
>>>
>>> But of course, large parts of it were due to semantic usage. You asked how
>>> it can fit C coding, this is my answer.
>> 
>> But which bits? I have no idea which bits were semantic and which others
>> were part of your "mish mash of other packages".
>
> Semantic shines through whenever you can see emacs visibly having an idea
> about the code's meaning. In this demo, it means 1) the suggestion popus
> 2) the second time I use `fwritex' template, in which FILE *STREAM formal
> parameter is replaced by a choice of all currently visible variables of
> type FILE*
>
>>>> Why did you do your own completion? What was wrong with intellisense?
>>>> Have you used intellisense?
>>>
>>> Uhh, this *is* intellisense. At least as long as you regard
>>> "intellisense"
>> 
>> Well, when you say your wrote your own completion package, how is one
>> supposed to know the difference?
>
> "intellisense" is either 1) IntelliSense, a trademarked name for MSVS

Why would I be talking about Intellisense from MS? What sense would that
make when discussing semantic in an emacs NG?

> context-sensitive completion, in which case none of the above is
> intellisense 2) a generic name for a class of similarly-working packages,
> in which case both are implementations of intellisense. If you want to
> know the difference, download the package and read the description 

Are really do not know why you appear to be so determined to make this
difficult. Do you really think I mean anything other than the
Intellisense which is part of the semantic package? And is called
"Intellisense" by cedet package itself? Here:

http://cedet.sourceforge.net/intellisense.shtml

>
>>> as a generic description and not trademarked name by MSFT. I wrote it
>>> because the existing completion popups for semantic were not close
>>> enonugh to the original IntelliSense, or rather, diverged in ways that
>>> seriously hindered their usefulness.
>> 
>> I turn the pop ups off and use tab cycle completion - works quite
>> nicely.
>> 
>> But again, could U request some sort of explanation of the demo gif?
>> Just looking at it tells me nothing whatsoever about how you used
>> semantic to do a job of work.
>
> Not really, I don't have the time for that now. I intend to turn it into a
> properly integrated and easy to install and setup package, but it hasn't
> happened yet.

Cheers.

Basically all you have said is "look at this gif which is made by a
mash of packages only I know about". How that is supposed to help in any
way understand semantic is beyond me. It does look impressive though and
my basic frustration is simply being unable to get it (semantic) to work
anything like you have demonstrated :-(

>
> Cheers,
> Maciej
>
>
>

--