CEDET

klaus.berndl@... wrote:
> Hi Timothee,
>
> I forward your email to the cedet-list as well, because you mentioned
> semantic too and you are right: probably for an end-user it isn't not always easy
> to know which part (ecb, semantic) does what.
>
> But in general there is a simple rule: ALL (!!) parsing is done by semantic,
> ECB is only the displayor for the parsing results.
>
> But i'm stll wondering what is so difficult for you to come up with a working
> combination of ECB/cedet(semantic)??
>
> I would understand if you mean such topics like intellisense or navigating large
> code-bases residing in a lot of different direktories because for these there are
> still suboptimal solutions....
>   
Yes, exactly that.

The C++ codebase I work with has around 1400 files in about a 100
directories. That's not a really big codebase either, seems to me a lot
of people have to deal with much larger stuff.

The features I am looking for are indeed intellisense and navigating
types. Pretty much all the things that are mentioned in the page:
http://www.emacswiki.org/cgi-bin/wiki/WhatAreTheNeedsOfProgrammers

Sometimes I feel that a lot the priorities of functionality work in IDE
tools ( such as ECB and several others ) do not match what as an
everyday user I consider most basic and essential. Visual Studio doesn't
provide much code browsing functionality, but at least it has the basics
working, which are autocomplete and easy navigation to
definition/declaration.

TTimo
> But again: Please give us a list of your problems which you want to be solved by such
> a emacswiki (which is a good idea IMO)???
>
> In general it should be very simple to setup a working ECB/semantic-environment
> and the docu of ECB and ist topics "Installation" and "First steps" and "FAQ" should
> all adress in an understandable manner - if not, please help us to enhance it.
>
> Ciao,
> Klaus
>
> Timothee Besset wrote:
>   
>> I appreciate Klaus's effort with ECB very much, and I've grown very
>> frustrated with my incapacity to do much of anything useful with the
>> software.  
>>
>> So I started putting a page together on the emacs wiki, hopefully
>> hoping to gather practical, pragmatic and to the point documentation
>> of what you can do with ECB and how to do it.  
>>
>> I hope some of you will want to comment and add content to that page
>> and that eventually some of semantic/ecb functionality's design can
>> be more user-driven.  
>>
>> TTimo
>>
>>
>>
>> -------------------------------------------------------
>> This SF.Net email is sponsored by xPML, a groundbreaking scripting
>> language that extends applications into web and mobile media. Attend
>> the live webcast and join the prime developer group breaking into
>> this new coding territory!
>> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
>> _______________________________________________   
>> Ecb-list mailing list
>> Ecb-list@...
>> https://lists.sourceforge.net/lists/listinfo/ecb-list
>>     
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by xPML, a groundbreaking scripting language
> that extends applications into web and mobile media. Attend the live webcast
> and join the prime developer group breaking into this new coding territory!
> http://sel.as-us.falkag.net/sel?cmd=k&kid0944&bid$1720&dat1642
> _______________________________________________
> Ecb-list mailing list
> Ecb-list@...
> https://lists.sourceforge.net/lists/listinfo/ecb-list
>

Hi Timo,

You are right that the intellisense stuff doesn't live up to the
expectations of anyone coming from Visual Studio or other environment,
but that doesn't mean it isn't a priority.

Good intellisense / tag jumping for a very large project basically
requires good parsing in a local context, that all the code is parsed,
and tags are stored in a fast database.  Currently the parsing is slow,
and the database really needs to be offloaded outside of Emacs.

Anyway, that is where the effort needs to go.  Solving those problems
will make solving intellisense specific problems worthwhile.

Eric

>>> Timothee Besset <ttimo@...> seems to think that:
>klaus.berndl@... wrote:
>> Hi Timothee,
>>
>> I forward your email to the cedet-list as well, because you mentioned
>> semantic too and you are right: probably for an end-user it isn't not always easy
>> to know which part (ecb, semantic) does what.
>>
>> But in general there is a simple rule: ALL (!!) parsing is done by semantic,
>> ECB is only the displayor for the parsing results.
>>
>> But i'm stll wondering what is so difficult for you to come up with a working
>> combination of ECB/cedet(semantic)??
>>
>> I would understand if you mean such topics like intellisense or navigating large
>> code-bases residing in a lot of different direktories because for these there are
>> still suboptimal solutions....
>>   
>Yes, exactly that.
>
>The C++ codebase I work with has around 1400 files in about a 100
>directories. That's not a really big codebase either, seems to me a lot
>of people have to deal with much larger stuff.
>
>The features I am looking for are indeed intellisense and navigating
>types. Pretty much all the things that are mentioned in the page:
>http://www.emacswiki.org/cgi-bin/wiki/WhatAreTheNeedsOfProgrammers
>
>Sometimes I feel that a lot the priorities of functionality work in IDE
>tools ( such as ECB and several others ) do not match what as an
>everyday user I consider most basic and essential. Visual Studio doesn't
>provide much code browsing functionality, but at least it has the basics
>working, which are autocomplete and easy navigation to
>definition/declaration.
>
>TTimo

-- 
          Eric Ludlam:                 zappo@..., eric@...
   Home: http://www.ludlam.net            Siege: http://www.siege-engine.com
Emacs: http://cedet.sourceforge.net               GNU: http://www.gnu.org

Hi,

>> I would understand if you mean such topics like intellisense or
>> navigating large code-bases residing in a lot of different
>> direktories=20
>> because for these there are still suboptimal solutions....
>>=20
> Yes, exactly that.
>=20
> The C++ codebase I work with has around 1400 files in about a 100
> directories. That's not a really big codebase either, seems to me a
> lot of people have to deal with much larger stuff. =20

surely

> The features I am looking for are indeed intellisense and navigating
> types. Pretty much all the things that are mentioned in the page:=20
> http://www.emacswiki.org/cgi-bin/wiki/WhatAreTheNeedsOfProgrammers

Yes, indeed, all of that is pretty much important for serious =
programming,
not only with C++... And the most stuff is currently not really well =
adressed
by Emacs - Emacs is a much better editor than all other IDEs contain but =
the
"central" goodies which makes coding really efficient and effective (as
good intellisense and refactoring and code navigation over file- and
Directory-boundaries to mention only the most important ones - IMHO) are =

Currently much better supported by IDES like MS Studio, IDEA (java),
Eclipse etc....

>=20
> Sometimes I feel that a lot the priorities of functionality work in
> IDE tools ( such as ECB and several others ) do not match what as an
> everyday user I consider most basic and essential. Visual Studio
> doesn't provide much code browsing functionality, but at least it has
> the basics working, which are autocomplete and easy navigation to
> definition/declaration.

No, i think exactly this stuff has highest priority at least for the=20
Cedet- and ECB-developers but currently tere is a lack of needed basic
functionality as really fast parsing (for huge code-bases) and a good
"database" which can store the parsed informations with fast access...

I think there are some promising approaches but currently we have to =
admit,
that there is no really ready-to-use-solution out.

Do you know Harmonia-project? See =
http://harmonia.cs.berkeley.edu/harmonia/index.html
IMHO this could be a good approach to reach the goals... Maybe ECB could =
integrate this
(harmonia has already a Xemacs integration) or maybe cedet/semantic =
could
take some benefit from this project - do not know...

Ciao,
Klaus


>=20
> TTimo
>> But again: Please give us a list of your problems which you want to
>> be=20
>> solved by such a emacswiki (which is a good idea IMO)???
>>=20
>> In general it should be very simple to setup a working
>> ECB/semantic-environment and the docu of ECB and ist topics
>> "Installation" and "First steps" and "FAQ" should all adress in an
>> understandable manner - if not, please help us to enhance it.=20
>>=20
>> Ciao,
>> Klaus
>>=20
>> Timothee Besset wrote:
>>=20
>>> I appreciate Klaus's effort with ECB very much, and I've grown very
>>> frustrated with my incapacity to do much of anything useful with
>>> the software.=20
>>>=20
>>> So I started putting a page together on the emacs wiki, hopefully
>>> hoping to gather practical, pragmatic and to the point documentation
>>> of what you can do with ECB and how to do it.
>>>=20
>>> I hope some of you will want to comment and add content to that page
>>> and that eventually some of semantic/ecb functionality's design can
>>> be more user-driven.=20
>>>=20
>>> TTimo
>>>=20
>>>=20
>>>=20
>>> -------------------------------------------------------
>>> This SF.Net email is sponsored by xPML, a groundbreaking scripting
>>> language that extends applications into web and mobile media. Attend
>>> the live webcast and join the prime developer group breaking into
>>> this new coding territory!
>>> =
http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D110944&bid=3D241720&dat=3D=
121642
>>> _______________________________________________=20
>>> Ecb-list mailing list
>>> Ecb-list@...
>>> https://lists.sourceforge.net/lists/listinfo/ecb-list
>>>=20
>>=20
>>=20
>> -------------------------------------------------------
>> This SF.Net email is sponsored by xPML, a groundbreaking scripting
>> language that extends applications into web and mobile media. Attend
>> the live webcast and join the prime developer group breaking into
>> this new coding territory!
>> =
http://sel.as-us.falkag.net/sel?cmd=3Dk&kid=110944&bid$1720&dat=121642
>> _______________________________________________=20
>> Ecb-list mailing list
>> Ecb-list@...
>> https://lists.sourceforge.net/lists/listinfo/ecb-list

>>> <klaus.berndl@...> seems to think that:
>Hi,
  [ ... ]
>Do you know Harmonia-project? See http://harmonia.cs.berkeley.edu/harmonia/index.html
>IMHO this could be a good approach to reach the goals... Maybe ECB could integrate this
>(harmonia has already a Xemacs integration) or maybe cedet/semantic could
>take some benefit from this project - do not know...
>
>Ciao,
>Klaus
  [ ... ]

I had investigated this tool briefly last time it came up.  They have
taken their tool, and implemented the things we'd like to do.  Chances
are their tool could be hooked into the semantic framework but I
haven't had time to investigate.

Eric

-- 
          Eric Ludlam:                 zappo@..., eric@...
   Home: http://www.ludlam.net            Siege: http://www.siege-engine.com
Emacs: http://cedet.sourceforge.net               GNU: http://www.gnu.org

klaus.berndl@... wrote:
> [ ..]
>> Sometimes I feel that a lot the priorities of functionality work in
>> IDE tools ( such as ECB and several others ) do not match what as an
>> everyday user I consider most basic and essential. Visual Studio
>> doesn't provide much code browsing functionality, but at least it has
>> the basics working, which are autocomplete and easy navigation to
>> definition/declaration.
>>     
>
> No, i think exactly this stuff has highest priority at least for the 
> Cedet- and ECB-developers but currently tere is a lack of needed basic
> functionality as really fast parsing (for huge code-bases) and a good
> "database" which can store the parsed informations with fast access...
>
> I think there are some promising approaches but currently we have to admit,
> that there is no really ready-to-use-solution out.
>   
Well .. my problem may really be not knowing what can actually be done
by semantic/ECB at this point. I came to it expect the same basic level
of functionality. Didn't see anything in the docs that would explicitely
say otherwise .. so I assumed it was not documented, whereas it's
actually not possible?

I think I eventually want my wiki page to list topics, and cover if they
can be done with semantic/ECB, or if you need to use other tools for
that, or if there is currently no efficient solution.
> Do you know Harmonia-project? See http://harmonia.cs.berkeley.edu/harmonia/index.html
> IMHO this could be a good approach to reach the goals... Maybe ECB could integrate this
> (harmonia has already a Xemacs integration) or maybe cedet/semantic could
> take some benefit from this project - do not know...
>
>   
Harmonia looks very promising indeed. The current release doesn't
support C++ however, so I'm on a wait and see till C++ is there to try
it out.

TTimo
> Ciao,
> Klaus
>
>
>   
>> TTimo
>>     
>>> But again: Please give us a list of your problems which you want to
>>> be 
>>> solved by such a emacswiki (which is a good idea IMO)???
>>>
>>> In general it should be very simple to setup a working
>>> ECB/semantic-environment and the docu of ECB and ist topics
>>> "Installation" and "First steps" and "FAQ" should all adress in an
>>> understandable manner - if not, please help us to enhance it. 
>>>
>>> Ciao,
>>> Klaus
>>>
>>> Timothee Besset wrote:
>>>
>>>       
>>>> I appreciate Klaus's effort with ECB very much, and I've grown very
>>>> frustrated with my incapacity to do much of anything useful with
>>>> the software. 
>>>>
>>>> So I started putting a page together on the emacs wiki, hopefully
>>>> hoping to gather practical, pragmatic and to the point documentation
>>>> of what you can do with ECB and how to do it.
>>>>
>>>> I hope some of you will want to comment and add content to that page
>>>> and that eventually some of semantic/ecb functionality's design can
>>>> be more user-driven. 
>>>>
>>>> TTimo
>>>>
>>>>
>>>>
>>>> -------------------------------------------------------
>>>> This SF.Net email is sponsored by xPML, a groundbreaking scripting
>>>> language that extends applications into web and mobile media. Attend
>>>> the live webcast and join the prime developer group breaking into
>>>> this new coding territory!
>>>> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
>>>> _______________________________________________ 
>>>> Ecb-list mailing list
>>>> Ecb-list@...
>>>> https://lists.sourceforge.net/lists/listinfo/ecb-list
>>>>
>>>>         
>>> -------------------------------------------------------
>>> This SF.Net email is sponsored by xPML, a groundbreaking scripting
>>> language that extends applications into web and mobile media. Attend
>>> the live webcast and join the prime developer group breaking into
>>> this new coding territory!
>>> http://sel.as-us.falkag.net/sel?cmd=k&kid0944&bid$1720&dat1642
>>> _______________________________________________ 
>>> Ecb-list mailing list
>>> Ecb-list@...
>>> https://lists.sourceforge.net/lists/listinfo/ecb-list
>>>       
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by xPML, a groundbreaking scripting language
> that extends applications into web and mobile media. Attend the live webcast
> and join the prime developer group breaking into this new coding territory!
> http://sel.as-us.falkag.net/sel?cmd=k&kid0944&bid$1720&dat1642
> _______________________________________________
> cedet-semantic mailing list
> cedet-semantic@...
> https://lists.sourceforge.net/lists/listinfo/cedet-semantic
>

Timothee Besset wrote:
> Harmonia looks very promising indeed. The current release doesn't
> support C++ however, so I'm on a wait and see till C++ is there to try
> it out.

I use ebrowse (which comes with Emacs) to generate a Tags-DB for all my 
source files, and use its functionality to jump to method 
definitions/declarations.  Ebrowse is C++ only, but does its job very 
well.  It parses several hundred source files per second, and works also 
on quite big projects.  I use it in a project consisting of 3500+ files 
in 500+ directories.

If you cannot get ebrowse up and running, I will post some more details.

Have fun,
Markus

Markus Gritsch <gritsch@...> writes:

> Timothee Besset wrote:
>> Harmonia looks very promising indeed. The current release doesn't
>> support C++ however, so I'm on a wait and see till C++ is there to try
>> it out.
>
> I use ebrowse (which comes with Emacs) to generate a Tags-DB for all
> my source files, and use its functionality to jump to method
> definitions/declarations.  Ebrowse is C++ only, but does its job very
> well.  It parses several hundred source files per second, and works
> also on quite big projects.  I use it in a project consisting of 3500+
> files in 500+ directories.

What would be neat is if cedet backends were implemented for tools
like Ebrowse. Then one would have a consistent emacs interface to use
for all code.

For instance Ocaml has a feature to compile so as to create
annotiation files that can be used for finding the type at point in
emacs tuareg mode. It would be neat if this also was made into a cedet
backend.

Also it would be nice if cedet was clever enough to use for instance
ebrowse to find tags when the source file hasnt been touched, and its
own wisent parser otherwise.



>
> If you cannot get ebrowse up and running, I will post some more details.
>
> Have fun,
> Markus
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by xPML, a groundbreaking scripting language
> that extends applications into web and mobile media. Attend the live webcast
> and join the prime developer group breaking into this new coding territory!
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642

-- 
Joakim Verona
 http://www.verona.se

>>> Joakim Verona <joakim@...> seems to think that:
>Markus Gritsch <gritsch@...> writes:
>
>> Timothee Besset wrote:
>>> Harmonia looks very promising indeed. The current release doesn't
>>> support C++ however, so I'm on a wait and see till C++ is there to try
>>> it out.
>>
>> I use ebrowse (which comes with Emacs) to generate a Tags-DB for all
>> my source files, and use its functionality to jump to method
>> definitions/declarations.  Ebrowse is C++ only, but does its job very
>> well.  It parses several hundred source files per second, and works
>> also on quite big projects.  I use it in a project consisting of 3500+
>> files in 500+ directories.
>
>What would be neat is if cedet backends were implemented for tools
>like Ebrowse. Then one would have a consistent emacs interface to use
>for all code.
>
>For instance Ocaml has a feature to compile so as to create
>annotiation files that can be used for finding the type at point in
>emacs tuareg mode. It would be neat if this also was made into a cedet
>backend.
>
>Also it would be nice if cedet was clever enough to use for instance
>ebrowse to find tags when the source file hasnt been touched, and its
>own wisent parser otherwise.

Hi,

  The semantic parser and the semanticdb storage system were both
designed so they could be replaced.  If you have your own parser, all
you need to do is write a transformation function to convert one data
type into semantic tags.

  If you have your own tag table, you can write a semanticdb object
and you just need to implement a small number of search routines, plus
something to convert the result into a semantic tag.

  In both cases there are examples for doing this.  As Klaus said,
someone just needs to take the time to do it.  I had looked into the
EBrowse structure, hoping to use the ebrowse lisp code, and I was
befuddled by what it was trying to do.  I have half of this system
written and it needs someone to take it the rest of the way and
implement the ebrowse piece into the semanticdb skeleton.

  The 'clever' part of knowing when to use which database is already
in the system, and used with Emacs Lisp (which has a working alternate
semanticdb back end.)

Eric

-- 
          Eric Ludlam:                 zappo@..., eric@...
   Home: http://www.ludlam.net            Siege: http://www.siege-engine.com
Emacs: http://cedet.sourceforge.net               GNU: http://www.gnu.org