Emacs Ada code navigation tools

Emacs Ada code navigation tools

[Mats Weber]

In the good old days, I was using DEC Ada with the LSE editor and the compiler
generated analysis output that told the editor where the declaration of each
occurrence of an identifier was. I just had to press ctrl-D on any identifier,
and it automatically took me to the declaration. There were many other useful
functions such as list all uses of a given declaration, etc.

Now I am using Emacs (which I find better than LSE in some respects, and worse
in others, but let's not discuss that).

Is there an equivalent tool for Emacs/Xemacs/GNAT (maybe compiler independent,
but then probably overloading will not be treated correctly) ?

Re: Emacs Ada code navigation tools

[Markus Kuhn]

Mats Weber wrote:
> In the good old days, I was using DEC Ada with the LSE editor and the compiler
> generated analysis output that told the editor where the declaration of each
> occurrence of an identifier was. I just had to press ctrl-D on any identifier,
> and it automatically took me to the declaration. There were many other useful
> functions such as list all uses of a given declaration, etc.

Such features are rumoured to be better supported by GNAT 3.11p, which
I hope will be out in just a few weeks. The ALI files have been
extended to contain the cross-referencing information necessary
to support for declaration finding without the use of an extra
cross-referencing tool as it is necessary at the moment for the
current Emacs Ada mode (see documentation).

Markus

-- 
Markus G. Kuhn, Security Group, Computer Lab, Cambridge University, UK
email: mkuhn at acm.org,  home page: <http://www.cl.cam.ac.uk/~mgk25/>

Re: Emacs Ada code navigation tools

[dennison]

In article <35F6F05E.9A8F2DD0@elca-matrix.ch>,
  Mats.Weber@elca-matrix.ch wrote:
> In the good old days, I was using DEC Ada with the LSE editor and the compiler
> generated analysis output that told the editor where the declaration of each
> occurrence of an identifier was. I just had to press ctrl-D on any identifier,
> and it automatically took me to the declaration. There were many other useful
> functions such as list all uses of a given declaration, etc.

As I remember it, that required using a third tool (in addition to LSE and
DecAda). As it has been about 8 years, I don't remember its three-letter
acronym anymore.

What I do remember is that it was so slow it was next to useless. It would
take several minutes to find a declaration. In that amount of time you could
easily find it yourself with "search". Plus it took its information from the
symbol table, so if you significantly changed your source since the last
compile it wouldn't work right anymore. I was the only developer out of 50+
developers on the project who used it more than a day, and I gave up on it
after a week. Even DTM was more useful!

Not that I wouldn't like to see a *good* implementation of that
functionality...

--
T.E.D.

-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/rg_mkgrp.xp   Create Your Own Free Member Forum

Re: Emacs Ada code navigation tools

[Mats Weber]

dennison@telepath.com wrote:

> As I remember it, that required using a third tool (in addition to LSE and
> DecAda). As it has been about 8 years, I don't remember its three-letter
> acronym anymore.

It is SCA (for Source Code Analyzer).

> What I do remember is that it was so slow it was next to useless. It would
> take several minutes to find a declaration. In that amount of time you could
> easily find it yourself with "search". Plus it took its information from the
> symbol table, so if you significantly changed your source since the last
> compile it wouldn't work right anymore. I was the only developer out of 50+
> developers on the project who used it more than a day, and I gave up on it
> after a week. Even DTM was more useful!

For me, it was more like 3 seconds to get to the declaration. The reason I
stopped using it is that LSE became an unstable mess when they merged VMS/UNIX
implementations, and you had to throw away your whole environment at every new version.

I found that tool very useful when putting new people on a project, or when a
programmer has over-used use clauses, a case where grep/search does not help much.

Anyway, I didn't start this thread to talk about the old days, but about what
can be done _now_ with Emacs :-)

Re: Emacs Ada code navigation tools

[dewarr]

In article <6t8n57$st7$1@nnrp1.dejanews.com>,
  dennison@telepath.com wrote:
> In article <35F6F05E.9A8F2DD0@elca-matrix.ch>,
>   Mats.Weber@elca-matrix.ch wrote:
> > In the good old days, I was using DEC Ada with the LSE editor and the
compiler
> > generated analysis output that told the editor where the declaration of
each
> > occurrence of an identifier was. I just had to press ctrl-D on any
identifier,
> > and it automatically took me to the declaration. There were many other
useful
> > functions such as list all uses of a given declaration, etc.
>
> As I remember it, that required using a third tool (in addition to LSE and
> DecAda). As it has been about 8 years, I don't remember its three-letter
> acronym anymore.
>
> What I do remember is that it was so slow it was next to useless. It would
> take several minutes to find a declaration. In that amount of time you could
> easily find it yourself with "search". Plus it took its information from the
> symbol table, so if you significantly changed your source since the last
> compile it wouldn't work right anymore. I was the only developer out of 50+
> developers on the project who used it more than a day, and I gave up on it
> after a week. Even DTM was more useful!
>
> Not that I wouldn't like to see a *good* implementation of that
> functionality...
>
> --
> T.E.D.
>
> -----== Posted via Deja News, The Leader in Internet Discussion ==-----
> http://www.dejanews.com/rg_mkgrp.xp   Create Your Own Free Member Forum
>


As Markus hinted, this functionality is indeed present in
the latest version of GNAT, without having to use any
additional tools. Version 3.11b of GNAT generates cross-
reference information by default that is used by EMACS to
provide this functionality.

We will be announcing the availability of GNAT 3.11b very
shortly with details on this forum and elsewhere (the beta
release of 3.11b has already been made available some time
ago). A corresponding public release (3.11p) will occur some
time in the future.

-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/rg_mkgrp.xp   Create Your Own Free Member Forum

Re: Emacs Ada code navigation tools

[dennison]

In article <6t8pri$d4t$1@fleetstreet.Austria.EU.net>,
  "Werner Pachler" <werner.pachler@austria.ncr.com> wrote:
> i have been using DEC/VAX with VMS, DEC Ada, LSE and VAX/VMS debugger
> Especially after using the other Compilers, debuggers, etc. i thanked god
> that he gave
> the developers of DEC the power to develop these tools (including VMS).

I can't disagree. The DEC development environment was definitely the best
overall development environment it has ever been my pleasure to use.

> >As I remember it, that required using a third tool (in addition to LSE and
> >DecAda). As it has been about 8 years, I don't remember its three-letter
> >acronym anymore.
> >
> I can't image what three-letter tool you mean.

I think it was something like "DEC SCA" (source code analizer?).

> The proplem you had was that your development environment had to less power
> to be usefull.
> Buy a faster VAX or Alpha...

Well, at the time I was working on the fastest VMS systems available. We had
some highly unrealistic speed requirements, and money was no object. We had a
network of about 8 Vax Clusters each with two Vaxen and a shadow disk (to
make up the quoroum if one Vax died). Each Vax had two processors. The
developers each had their own DECstation, which was served off one of two of
the clusters. Because nature of the work I was doing, I usually had an entire
cluster to myself.

SCA was just  s l o w.

That was 9 years ago, so I'm sure things improved. But at the time this was
just about the fastest (DEC) setup around.

--
T.E.D.

-----== Posted via Deja News, The Leader in Internet Discussion ==-----
http://www.dejanews.com/rg_mkgrp.xp   Create Your Own Free Member Forum

Re: Emacs Ada code navigation tools

[Werner Pachler]

T.E.D.,
i have been using DEC/VAX with VMS, DEC Ada, LSE and VAX/VMS debugger
a couple of years and i did also use Ada Compilers, Editors Debuggers on
many
other platforms
(MS-DOS, Windows-95, Windows-NT, DEC Ultrix and SGI-IRIX).

Especially after using the other Compilers, debuggers, etc. i thanked god
that he gave
the developers of DEC the power to develop these tools (including VMS).

The big trouble came to me when i tried to port my code (Ada Code!!!!!) to
the plattforms described above.
(Compiler bugs, editors like MS-DOS edit, debuggers like dbx, ...).

>
>As I remember it, that required using a third tool (in addition to LSE and
>DecAda). As it has been about 8 years, I don't remember its three-letter
>acronym anymore.
>
I can't image what three-letter tool you mean.

>What I do remember is that it was so slow it was next to useless. It would
>take several minutes to find a declaration. In that amount of time you
could
>easily find it yourself with "search". Plus it took its information from
the
>symbol table, so if you significantly changed your source since the last
>compile it wouldn't work right anymore. I was the only developer out of 50+
>developers on the project who used it more than a day, and I gave up on it
>after a week. Even DTM was more useful!
>
>Not that I wouldn't like to see a *good* implementation of that
>functionality...


The proplem you had was that your development environment had to less power
to be usefull.
Buy a faster VAX or Alpha...

PS: A friend of mine, after changing from VAX/VMS to a unix plattform said:
"Every alpha release of DIGITAL software is better than the official release
of
software on my plattform!".

W.Pachler@mobilkom.at

Re: Emacs Ada code navigation tools

[Werner Pachler]

dennison@telepath.com schrieb in Nachricht
<6t94gi$gr0$1@nnrp1.dejanews.com>...
>...
>
>I think it was something like "DEC SCA" (source code analizer?).


Now i know what you meant:
Digital sold a toolset called "VAXSet" including
LSE  = Language Sensitive Editor
DTM = Dec Test Manager
SCA = Source Code Analyzer
MMS = Module Management System
PCA = Performance and Coverage Analyzer
CMS = Code Management System

>Well, at the time I was working on the fastest VMS systems available. We
had
>some highly unrealistic speed requirements, and money was no object. We had
a
>network of about 8 Vax Clusters each with two Vaxen and a shadow disk (to
>make up the quoroum if one Vax died). Each Vax had two processors. The
>developers each had their own DECstation, which was served off one of two
of
>the clusters. Because nature of the work I was doing, I usually had an
entire
>cluster to myself.
>
>SCA was just  s l o w.
>
>That was 9 years ago, so I'm sure things improved. But at the time this was
>just about the fastest (DEC) setup around.
>

I did use a VAXStation 3100 M76 (speed approximatly something like an
Pentium 70) a few
years ago in a cluster environment and reached times up to 30 seconds for
compiling
1000 lines of Ada Code.
You're right, SCA was horrible slow, but as i can remember SCA checked
things like
"Does actual number and types of parameters match with defined subprogs,..."
and we all know that
you do not need this tool for developing in Ada.


>developers each had their own DECstation, which was served off one of two
of
                                                        ^^^^^^^^^^^^
PS: As i can remember, the computer DIGITAL sold 9 years ago distinguished
between
VAXStaion ... a Workstation with a VAX processor and
DECStation ... a Workstation with a MIPS processor and DEC Ultrix.