doxygen-develop — List for discussing new features, submitting patches, etc.

[Doxygen-develop] Aliases to create tables

[<Eck...@t-...>]

Hello everybody.


I often use the following examples to create tables:

ALIASES     = "TabBegin = <TABLE><TR><TD>" \
                         "TabEnd   = </TD></TR></Table>" \
                         "I        = </TD><TD>" \
                         
"I________________________________________________________________________________ 
= </TD></TR><TR><TD>" \
                         
"I_______________________________________________________________________________ 
= </TD></TR><TR><TD>" \
                         
"I______________________________________________________________________________ 
= </TD></TR><TR><TD>" \
                            .
                            .
                            .
                               the number of "_" changes from 80 to 1
                            .
                            .
                            .
                         "I_ = </TD></TR><TR><TD>"

This tags produce html-tables in the documentation


This is a source example where the new commands are used:
 @TabBegin
               Title 1         @I            Title 2         @I      Title 3
  @I______________________________________________
       row 1 column 1  @I   row 1 column 2   @I   row 1 column 3
  @I______________________________________________
       row 2 column 1  @I   row 2 column 2   @I   row 2 column 3
  @I______________________________________________
       row 3 column 1  @I   row 3 column 2   @I   row 3 column 3
  @I______________________________________________
       row 4 column 1  @I   row 4 column 2   @I   row 4 column 3
@TabEnd

Since more the one @I___-commands are defined it is possible to use 
short lines for small tables and long lines for tables with more columns.

May be it is a good idea to change the design of the aliases a little 
bit. For example it may be good to insert tag to write the headline in bold.

ALIASES     = "TabBegin = <TABLE><TR><TD><B>" \
                         "Ib        = </B></TD><TD><B>" \
                         "IbEnd        = </B>\

This is a source example where the new commands are used:

 @TabBegin
               Title 1         @Ib           Title 2        @Ib      
Title 3     @IbEnd
  @I______________________________________________
       row 1 column 1  @I   row 1 column 2   @I   row 1 column 3
  @I______________________________________________
       row 2 column 1  @I   row 2 column 2   @I   row 2 column 3
  @I______________________________________________
       row 3 column 1  @I   row 3 column 2   @I   row 3 column 3
  @I______________________________________________
       row 4 column 1  @I   row 4 column 2   @I   row 4 column 3
@TabEnd


Is there a special alias-archive with other examples like this?



Kind Regards,
                        Eckard Klotz.





>   

Re: [Doxygen-develop] PATCH: Add tags for tabular lists (ala \param, \retval, etc)

[mwoehlke <mwo...@ti...>]

mwoehlke wrote:
> The attached patch adds \li2 and \li3 tags to Doxygen (diff'd against 
> 1.4.7). \li2 is identical to \param except that it does not cause a 
> section to be created, and so can be used with custom sections (\par) or 
> in the detailed description. \li3 uses the first (typically unused) 
> column to make three-column lists (\li2 does 2-column lists, obviously); 
> it treats '-' as an empty cell.
> 
> Needed improvements:
> - Support the new tags in other output formats (shouldn't be difficult, 
> but I don't feel familiar enough with the code - or LaTeX for that 
> matter - to want to mess with it).
> 
> Suggested improvements:
> - Allow mixing \li2 and \li3 in the same table/list (and then remove 
> treating '-' as empty).
> - Add \li1?
> 
> I'm also open to suggestions on how to make the new code less ugly (for 
> instance, not needing the second params list would be nice).

Any comments?

-- 
Matthew
Download. Untar. Configure. Make. Install. Lather. Rinse. Repeat.

Re: [Doxygen-develop] Support for C++ Template Classes

[Jens M. <ju...@ma...>]

Am 27.09.2006 um 22:14 schrieb Ethan Tira-Thompson:

>> We use a lot of template base classes that are specialized for  
>> different types (similar to vector<int>). To simplify things we  
>> use a lot of typedefs (like typdef vector<int> ivec). But given  
>> that they're just typedefs they don't show up in the class list or  
>> the inhertiance hierarchy, even though for all practical reasons  
>> they're classes.
>
> I just want to add my vote for this as well -- in our case, we have  
> a templated class:
> template<class T, const char* mcName=defName, const char*  
> mcDesc=defDesc, bool completes=true> class MCNode;
>
> And then for the "simple" cases we only need to pass template  
> parameters:
> typedef  
> MCNode<TailWagMC,defTailWagNodeName,defTailWagNodeDesc,false>  
> TailWagNode;
> typedef  
> MCNode<HeadPointerMC,defHeadPointerNodeName,defHeadPointerNodeDesc>  
> HeadPointerNode;
> ...
>
> And for "harder" issues we can uses subclasses instead:
> class LedNode : public MCNode<LedMC> { /*...*/ }
>
> So the trick is, LedNode shows up in the class listing as expected,  
> but TailWagNode and HeadPointerNode don't.  Technically, this is  
> accurate since they're really just typedefs, but from a user  
> standpoint, there's no difference -- the user is going to look in  
> the class listing to see what's available and not know the 'type'  
> exists.
>
> Perhaps there could be a keyword to treat a typedef as a subclass  
> instead of "just" a typedef?
>
>> Any ideas how to make them show up as classes in an easy way?  
>> Right now the place to find them is in the namespace list, and  
>> that's not very intuitive.
>
> Our workaround is to modify the code to use a "real" subclass  
> instead, but that's less than ideal, as it causes additional  
> overhead and requires the subclass to replicate all of the  
> constructors of the base class to forward them through (maintenance  
> issue).

You could, of course, make them subclasses for doxygen only, e.g.

//! nodes for the tail-wag thingy
#if __DOXYGEN__
	class TailWagNode : public  
MCNode<TailWagMC,defTailWagNodeName,defTailWagNodeDesc,false> {};
#else
	typedef  
MCNode<TailWagMC,defTailWagNodeName,defTailWagNodeDesc,false>  
TailWagNode;
#endif

This is still tedious (and some additional support in doxygen for  
this kind of problem would surely be great), but at least you don't  
have the additional overhead in your application and only minimally  
more maintenance issues.

</jum>

Re: [Doxygen-develop] Support for C++ Template Classes

[Ethan Tira-T. <ej...@an...>]

> We use a lot of template base classes that are specialized for  
> different types (similar to vector<int>). To simplify things we use  
> a lot of typedefs (like typdef vector<int> ivec). But given that  
> they're just typedefs they don't show up in the class list or the  
> inhertiance hierarchy, even though for all practical reasons  
> they're classes.

I just want to add my vote for this as well -- in our case, we have a  
templated class:
template<class T, const char* mcName=defName, const char*  
mcDesc=defDesc, bool completes=true> class MCNode;

And then for the "simple" cases we only need to pass template  
parameters:
typedef MCNode<TailWagMC,defTailWagNodeName,defTailWagNodeDesc,false>  
TailWagNode;
typedef  
MCNode<HeadPointerMC,defHeadPointerNodeName,defHeadPointerNodeDesc>  
HeadPointerNode;
...

And for "harder" issues we can uses subclasses instead:
class LedNode : public MCNode<LedMC> { /*...*/ }

So the trick is, LedNode shows up in the class listing as expected,  
but TailWagNode and HeadPointerNode don't.  Technically, this is  
accurate since they're really just typedefs, but from a user  
standpoint, there's no difference -- the user is going to look in the  
class listing to see what's available and not know the 'type' exists.

Perhaps there could be a keyword to treat a typedef as a subclass  
instead of "just" a typedef?

> Any ideas how to make them show up as classes in an easy way? Right  
> now the place to find them is in the namespace list, and that's not  
> very intuitive.

Our workaround is to modify the code to use a "real" subclass  
instead, but that's less than ideal, as it causes additional overhead  
and requires the subclass to replicate all of the constructors of the  
base class to forward them through (maintenance issue).

-ethan

Re: [Doxygen-develop] Custom commands with arguments?

[Iain B. <ia...@pc...>]

Hi,

On Tue, 2006-09-26 at 22:43 -0500, Dirk Reiners wrote:
> 	Hi everybody,
> 
> we've been thinking about integrating doxygen better with Trac, and one 
> solution would require custom commands with arguments. This is listed as 
> #28 on http://www.stack.nl/~dimitri/doxygen/todo.html, so apparently 
> somebody else would like it, too.
> 
> Is anybody working on that, and if not, can anybody give me a hint on 
> where to start and how to get it done?

I don't know how well it will fit your outcome, but I made a perl script
that you use as a pre-filter to the source code.  It looks for custom
commands that you put in a command file, and replaces the
doxygen-looking alias with what you specify from the file.

It sounds the same as the way doxygen does it except that you can put
latex or html before _and_after_ the doxygen tag.  eg:

=== doxygen way: ===

source file:
/**
 * @cmd world
 **/

doxygen.cfg:
ALIAS="cmd=<b>hello</b>"

output
<b>hello</b> world

=== my way: ===

source file:
/**
 * @cmd hello world
 **/

filter.txt:
'cmd', '<b>', '</b>'

output:
<b>hello world</b>

You can see the subtle difference - effectively "hello world" is an
argument to the command "cmd".

In reality, this is _very_ powerful.  I've created some very nice
looking pdf's using this filter, however it may not work so well for
you, because I customised it just how I wanted it, without thought for
other uses.  I've attached it anyway.  YMMV!

You use it like this:
doxygen.cfg:
INPUT_FILTER = "perl -w filter.pl filter.txt"

filter.txt:
'tag', 'html open', 'html close', 'latex open', 'latex close'

an argument is ended by either the start of a new alias (ie. '@') or the
end of the comment block (**/) you may want to change that to */

ie:

/**
 * @cmd1 argument to command 1
 * @cmd2 argument to command 2.  This
 * one wraps around.
 * @cmd3 arguement to command 3.
 **/
<c++ item to be documented>

The filter also keeps line numbering the same, so that you can still use
doxygen line references.

HTH!

-- 
Iain Buchanan <iain at pcorp dot com dot au>

Cats, no less liquid than their shadows, offer no angles to the wind.

[Doxygen-develop] Custom commands with arguments?

[Dirk R. <re...@ca...>]

	Hi everybody,

we've been thinking about integrating doxygen better with Trac, and one 
solution would require custom commands with arguments. This is listed as 
#28 on http://www.stack.nl/~dimitri/doxygen/todo.html, so apparently 
somebody else would like it, too.

Is anybody working on that, and if not, can anybody give me a hint on 
where to start and how to get it done?

Thanks

	Dirk

[Doxygen-develop] Support for C++ Template Classes

[Dirk R. <re...@ca...>]

	Hi everybody,

we're using templates pretty heavily, and some of their methods don't go 
well with doxygen.

We use a lot of template base classes that are specialized for different 
types (similar to vector<int>). To simplify things we use a lot of 
typedefs (like typdef vector<int> ivec). But given that they're just 
typedefs they don't show up in the class list or the inhertiance 
hierarchy, even though for all practical reasons they're classes. Any 
ideas how to make them show up as classes in an easy way? Right now the 
place to find them is in the namespace list, and that's not very intuitive.

The second problem, and that is hairier, is using mixins or traits. The 
basic idea is that classes are constructed as templates and are derived 
from one of their template arguments, or one of the arguments internal 
types. I can understand that this would be very hard to do right, but I 
would like to take a look at it anyway, and maybe kludge it. Can anybody 
give me a hint where and how doxygen stores template classes?

Thanks

	Dirk

[Doxygen-develop] Parallelisation?

[Dirk R. <re...@ca...>]

	Hi everybody,

like some other people, I'm having problems with the runtime of our 
doxygen project. It's fairly big and takes a very long time to run. One 
obvious soultuino would be to run multiple scanners/generators in 
parallel, as they are pretty independent.

I tried looking into it, but the classes didn't seem to be very 
parallelization-friendly (explicitly forbidden copy constructors and 
assignments, lots of global variables). Has anybody thoguht about it? 
Any hints on the best way of doing it?

Thanks

	Dirk

Re: [Doxygen-develop] STL classes dummy header

[Dimitri v. H. <do...@gm...>]

Hi Martin,

The FAQ is somewhat outdated at this point. Recent versions
of doxygen have a config option BUILTIN_STL_SUPPORT.
This does not provide documentation for the STL classes however, so your
file still has an advantage.

Regards,
  Dimitri

On 9/22/06, Martin Sherburn <em...@no...> wrote:
>
> In response to the FAQ, question #15. Why are dependencies via STL classes
> not shown in the dot graphs?
> >> I'm still looking for someone who can provide me with definitions for
> all (relevant) STL classes.
>
> I've made a header file which covers the main STL classes (vector, deque,
> list, slist, set, map and pair).
>
> Hopefully it will be usefull to others as well... There are still some
> classes missing (multiset, multimap, hash_set, hash_map, hash_multiset,
> hash_multimap). I don't think hash needs to be done since in the
> documentation it says "The hash<T> template is only defined for template
> arguments of type char*, const char*, crope, wrope, and the built-in
> integral types."
>
> If somebody wants to take some time to add the missing classes, what i did
> was just go to:
> http://www.sgi.com/tech/stl/table_of_contents.html
>
> there is a list of classes there... and if you click on each one there is
> a description for the class. I've just cut and pasted that into the header
> file.
>
> One other thing, I didn't put the classes in the std namespace because in
> my project "using namespace std;" is in my precompiled header file and
> doxygen misses it. Leaving it as it is will cause problems for people using
> the stl classes like this "std::vector<MyClass>", rather than "using
> namespace std; vector<MyClass>". I would suggest if this is ever included in
> with doxygen that the classes should be in the correct namespace. I'll just
> have to go and add "using namespace std;" at the top of all my .h files.
>
> SpaceDude.
>
>
> This message has been checked for viruses but the contents of an
> attachment
> may still contain software viruses, which could damage your computer
> system:
> you are advised to perform your own checks. Email communications with the
> University of Nottingham may be monitored as permitted by UK legislation.
>
>
>
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share
> your
> opinions on IT & business topics through brief surveys -- and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>
> _______________________________________________
> Doxygen-develop mailing list
> Dox...@li...
> https://lists.sourceforge.net/lists/listinfo/doxygen-develop
>
>
>
>

[Doxygen-develop] STL classes dummy header

[Martin S. <em...@no...>]

In response to the FAQ, question #15. Why are dependencies via STL classes =
not shown in the dot graphs?=0A=
>> I'm still looking for someone who can provide me with definitions for al=
l (relevant) STL classes.=0A=
=0A=
I've made a header file which covers the main STL classes (vector, deque, l=
ist, slist, set, map and pair).=0A=
=0A=
Hopefully it will be usefull to others as well... There are still some clas=
ses missing (multiset, multimap, hash_set, hash_map, hash_multiset, hash_mu=
ltimap). I don't think hash needs to be done since in the documentation it =
says "The hash<T> template is only defined for template arguments of type c=
har*, const char*, crope, wrope, and the built-in integral types."=0A=
=0A=
If somebody wants to take some time to add the missing classes, what i did =
was just go to:=0A=
http://www.sgi.com/tech/stl/table_of_contents.html=20=0A=
=0A=
there is a list of classes there... and if you click on each one there is a=
 description for the class. I've just cut and pasted that into the header f=
ile.=0A=
=0A=
One other thing, I didn't put the classes in the std namespace because in m=
y project "using namespace std;" is in my precompiled header file and doxyg=
en misses it. Leaving it as it is will cause problems for people using the =
stl classes like this "std::vector<MyClass>", rather than "using namespace =
std; vector<MyClass>". I would suggest if this is ever included in with dox=
ygen that the classes should be in the correct namespace. I'll just have to=
 go and add "using namespace std;" at the top of all my .h files.=0A=
=0A=
SpaceDude.=0A=
=0A=
=0A=
This message has been checked for viruses but the contents of an attachment=
=0A=
may still contain software viruses, which could damage your computer system:=
=0A=
you are advised to perform your own checks. Email communications with the=
=0A=
University of Nottingham may be monitored as permitted by UK legislation.=
=0A=
=0A=

[Doxygen-develop] PATCH: Add tags for tabular lists (ala \param, \retval, etc)

[mwoehlke <mwo...@ti...>]

The attached patch adds \li2 and \li3 tags to Doxygen (diff'd against 
1.4.7). \li2 is identical to \param except that it does not cause a 
section to be created, and so can be used with custom sections (\par) or 
in the detailed description. \li3 uses the first (typically unused) 
column to make three-column lists (\li2 does 2-column lists, obviously); 
it treats '-' as an empty cell.

Needed improvements:
- Support the new tags in other output formats (shouldn't be difficult, 
but I don't feel familiar enough with the code - or LaTeX for that 
matter - to want to mess with it).

Suggested improvements:
- Allow mixing \li2 and \li3 in the same table/list (and then remove 
treating '-' as empty).
- Add \li1?

I'm also open to suggestions on how to make the new code less ugly (for 
instance, not needing the second params list would be nice).

-- 
Matthew
Will your shell have salvation? Only if it's Bourne Again.

Re: [Doxygen-develop] What to use for an empty word?

[mwoehlke <mwo...@ti...>]

mwoehlke wrote:
> mwoehlke wrote:
>> I am writing code for a new tag, \li3, that is used to make generic 
>> lists with the style of \param, \retval, etc. It should take two words 
>> followed by a paragraph, however the first word should be optional. 
>> Therefore I want to be able to specify a null 'placeholder'. I am 
>> thinking about using '-' but thought I would see if there are any 
>> suggestions. (If there is already something that can be used as an empty 
>> placeholder, I have not found it.)
> 
> Ah, never mind, I was looking for ','...

Ack, scratch that, ',' doesn't work (the lexer skips empty tokens, so I 
can't detect the ',' or that there was an empty word given). So unless 
someone has a better idea, I will use '-'.

-- 
Matthew
73% of all statistics are made up on the spot.

Re: [Doxygen-develop] What to use for an empty word?

[mwoehlke <mwo...@ti...>]

mwoehlke wrote:
> I am writing code for a new tag, \li3, that is used to make generic 
> lists with the style of \param, \retval, etc. It should take two words 
> followed by a paragraph, however the first word should be optional. 
> Therefore I want to be able to specify a null 'placeholder'. I am 
> thinking about using '-' but thought I would see if there are any 
> suggestions. (If there is already something that can be used as an empty 
> placeholder, I have not found it.)

Ah, never mind, I was looking for ','...

-- 
Matthew
73% of all statistics are made up on the spot.

[Doxygen-develop] What to use for an empty word?

[mwoehlke <mwo...@ti...>]

I am writing code for a new tag, \li3, that is used to make generic 
lists with the style of \param, \retval, etc. It should take two words 
followed by a paragraph, however the first word should be optional. 
Therefore I want to be able to specify a null 'placeholder'. I am 
thinking about using '-' but thought I would see if there are any 
suggestions. (If there is already something that can be used as an empty 
placeholder, I have not found it.)

Example (different implementation of \param):
\par Parameters:
\li3 <tt>in</tt> foo A silly argument

The above is equivalent to:
\param[in] foo A silly argument

-- 
Matthew
73% of all statistics are made up on the spot.

[Doxygen-develop] Feature request

[Uri M. <ur...@4r...>]

I have a feature request and am not sure if this is the right place. The 
tracker on the source forge homepage doesn't seem to be very active. 
Anyway, I would like to use C#'s XML-style comments in my C++ project 
but support for a few tags is missing. In particular, I would like the 
tags "<todo>" and "<deprecated>". Are there plans to support these and 
other Doxygen tags in the C# XML style? I looked at modifying the source 
but this change doesn't seem to be straightforward. Can someone provide 
me with some hints as to what changes need to be made to support these 
tags? Thanks.

Uri

[Doxygen-develop] Fortran 9x

[Oleg B. <og...@gm...>]

Hi,

For the last 3 days I've been adding Fortran9x support to doxygen and
allready have some doc generated. But suddenly saw message with the
same effort.

What is the state of that code? I could drop my code and help writing
or testing it.

Oleg Batrashev

Re: [Doxygen-develop] [PATCH] IGNORE_PREFIX for functions

[Kevin M. <ke...@pl...>]

Eivind Magnus Hvidevold wrote:
> This patch includes resorting of the lists. It is against CVS from 
> yesterday.

I have imported the patch into the following bug report:

http://bugzilla.gnome.org/show_bug.cgi?id=333154

You should "CC" yourself onto the bug report (registration required) to
get e-mails about its status.

- KJM

-- 
This message has been digitally signed with a GlobalSign-issued certificate.
For more information, please visit GlobalSign's web site at:
http://www.globalsign.net/

Re: [Doxygen-develop] [PATCH] IGNORE_PREFIX for functions

[Eivind M. H. <em...@ar...>]

On Fri, 07 Jul 2006 15:05:57 +0200, Eivind Magnus Hvidevold  
<em...@ar...> wrote:

>
> Hi!
>
> I'm using Doxygen on an SDK where all C functions have the same prefix. I
> used the IGNORE_PREFIX setting to get them indexed properly by first
> letter after the prefix, but they all kept going under the first letter  
> of the prefix. So I searched the mailing lists for IGNORE_PREFIX and  
> found
> that it just works for classes. Here is a patch which makes it work also
> for functions. I'd be happy if this or something similar could be merged.
> Do you think there should there be a separate config setting? I was
> thinking of something like IGNORE_PREFIX_FUNCTIONS with YES|NO (using the
> list from IGNORE_PREFIX) or containing another ignore list. I don't need
> that extra setting but maybe others do?
The patch was incomplete. I forgot to resort the list with the prefix  
ignored,
so if there were functions with different prefix, the alphabetical index
would restart on 'a' for each prefix.

This patch includes resorting of the lists. It is against CVS from  
yesterday.

-- 
EMH

[Doxygen-develop] Fortran90 parser

[A.Visser <A.V...@fz...>]

Hi!

I'm currently writing a Fortran90 parser (subset, not all inherited f77 
botch) for doxygen. (I supposed that level 10 at doxygen wish list and 
-Fortran- means, that it will be put back the next decades :-))

I've got a few question:
- I've yet only read scanner.l and found nothing about the function 
references. How are they generated? Just equal names?
- How should registerParser be used and how to register several endings? 
At the moment, I've added one instances of the parser for each ending 
(f90 and F90).

Greetings,
Anke Visser

[Doxygen-develop] [PATCH] IGNORE_PREFIX for functions

[Eivind M. H. <em...@ar...>]

Hi!

I'm using Doxygen on an SDK where all C functions have the same prefix. I
used the IGNORE_PREFIX setting to get them indexed properly by first
letter after the prefix, but they all kept going under the first letter of
the prefix. So I searched the mailing lists for IGNORE_PREFIX and found
that it just works for classes. Here is a patch which makes it work also
for functions. I'd be happy if this or something similar could be merged.
Do you think there should there be a separate config setting? I was
thinking of something like IGNORE_PREFIX_FUNCTIONS with YES|NO (using the
list from IGNORE_PREFIX) or containing another ignore list. I don't need
that extra setting but maybe others do?

-- 
EMH

[Doxygen-develop] Imprecise warning

[Eivind M. H. <em...@ar...>]

doxygen.cpp, line 665 outputs a warning message:

text.sprintf("Warning: the name `%s' supplied as "
                   "the second argument in the \\class statement ",
                   root->includeFile.data()
                  );

This may happen for the \struct command also, so the warning should be  
"the second argument in the \\class or \\struct statement".

-- 
EMH

[Doxygen-develop] Moritz new release , new project-path, first linux-trial

[<Eck...@t-...>]

Hello Everybody.

There is a new release of Moritz the nassi-shneiderman generator for 
Doxygen.
The sourceforge-team has changed the project path so you can find Moritz 
under "www.sourceforge.net/projects/moritz" now.
I tried to create a linux-version also. But I'm a linux newbe and I 
don't really know if it works on all linux-distributions. I have an old 
Suse 9.0 distribution.

I know there is still a lot to do and sometimes it's a little bit 
difficult to decide what's the most important thing to do next. If you 
are missing important things like the documentation of special things or 
additional features don't hesitate to ask for. If you have questions or 
proposals or if you want to criticize something you can also use the 
forum of Moritz provided by sourceforge.
But since I'm on a business-trip next week I'm not able to answer before 
next weekend, sorry.

Regards,
               Eckard Klotz.

[Doxygen-develop] Building libdoxygen dylib?

[B.S. <br...@ma...>]

Has anyone tried building libdoxygen as a dylib rather than as a  
static library? I'm on Mac OS X and would like to experiment with  
embedding doxygen in an application.

Thanks in advance,
Brent

Re: [Doxygen-develop] Morirz First Releas on sourceforge.net

[<Eck...@t-...>]

Hello Mr. Dodge.

Thank you for your effort to make further tests.

Since I'm not a very good c++ programmer too, I've learned a lot from 
your mail.

1. I followed your proposal to use the standard definition  EXIT_FAILURE 
instead of -1 and hope that the compiler knows the right translation for 
the target-system.
2. I never saw the warning-message before you found, so I checked the 
default compiler-options an found that some warning-classes were not 
active. After changing that If got a lot of additional warnings, more 
than you told me. My philosophy is to try to solve the problems behind 
all warnings. On one hand it standardises my style of coding and on the 
other side I hope that those who designed this messages know more about 
the possible results than I.

I changed this thinks in the code and tested it. The compiler gives no 
warnings and the program works like before. But the binary is a little 
bit smaller because I set the methods of the abstract classes to 0. This 
means the compiler makes no object-code for them.

3. You proposed to deactivate the pause-commands. I thought about this 
too. But I think I have to recognize two scenarios.

    * A. Between normal operation it is not useful that the program
      needs an input of the user to continue its operation. In this
      situation it would be better to write information into a log-file.
    * B. But while changing the configuration of Moritz it may be a good
      opportunity to stop the whole generation-process  if an error
      occurs. For example to make the whole doc of the Moritz-source
      takes several minutes. Imagine how long it takes if you make a new
      set-up with out a break after an error.

I think I will implement an a special output-class that manages the 
communication with the user. The user can configure this outputs via 
config-file, so that errors, warnings or messages causes a brake or not, 
will written to a log or not and so on. Perhaps I will also add an 
possibility to configure the output-messages them self so that the user 
can translate them in his language.
What do you think?

And now lets talk about the crash

I have also used Moritz to build the source-documentation you will find 
on sourceforge (by the way it is possible to open chm-files on Linux or 
should I add the html-version of the docs?) . I had no problems to 
generate the diagrams.

So first I thought this is a special Linux-problem ore a change of the 
xml-format of the doxygen-output so that Moritz is not able to work with 
older versions. I used doxygen since version 1.4.
But then I remember my big failure to upload to many sources. So I 
followed exactly your way and include all sources you have inclusive 
those witch are not used for Moritz it self . After that I saw what you 
mean. Moritz crashes with a typical windows-message that tells you a lot 
but nothing about the real problem.
Now I started to reduce the files to find out which file causes the crash:

    * a. If you take only the files I've listed as necessary  for Moritz
      you will get what you want without a crash or an error-message.
    * b. If you include the file HTML.cpp in the sub-directory
      tools_wx/cpp you will get an error-message of Moritz, but this is
      a "normal" operation of the program and after pressing ENTER the
      rest runs fine. But in this special case behind this behaviour I
      found a bug of Moritz. The analysed function has a legal c++
      format that Moritz doesn't know yet. The problem is the const
      behind the parameter-block of the function head. This is a bug and
      I have to change it.
    * c. If you include the file cfgText.h in the sub-directory config/h
      Moritz crashes. This is caused by a conflict between 2 files with
      the same name and nearly the same content because there is also a
      file named cfgText.h  in the sub-directory xmlblock/h. I think
      this causes  inconsistent  information which Moritz can not solve.
      On the one hand there is a bug in Moritz because mistakes like
      these  should not lead in to a crash. Moritz has to recognize them
      and has to react with an error-message. On the other side this is
      a not usual operation because I think in a normal project for one
      target-binary you will not have 2 files with the same name and the
      same content which also means 2 classes with the same name, or not?

What do you think?

I hope this will help you. I have to make some other changes in the next 
days so it may take some days until the next release.

At the end I have a favour to ask you.
You wrote that you have changed the dos-batch in to a bash-script for 
Linux. I want to create a Linux-version too. But I'm an absolute new-be 
in Linux. I have an older Suse-distribution (version 9.0). I tested to 
compile Moritz on Linux sometimes between the development and it works. 
I bought a book about bash-scripts but didn't start to learn more about 
writing them yet. So if your script works it would be a good example for 
me.

Regards,
               Eckard Klotz.

Re: [Doxygen-develop] Morirz First Releas on sourceforge.net

[Dave D. <do...@do...>]

On Wed, May 31, 2006 at 06:25:05PM +0200, Eck...@t-... wrote:
> To build Moritz just the following files has to be used:
[...]

Thanks for the information.  I was able to create a simple Makefile
and get it compiled on a Linux machine.  I also made some small
changes to the code, such as changing the failure code in main() to
return EXIT_FAILURE instead of -1, and to not run the PAUSE command.
g++ still shows one warning, but I'm not really a C++ programmer so I
don't know if anything needs to be done about it:

  src/xmlvisitor/cpp/xmlvst_gennsd.cpp: In copy constructor `XmlVisitor_GeneratorFuncNSD::XmlVisitor_GeneratorFuncNSD(const XmlVisitor_GeneratorFuncNSD&)':
  src/xmlvisitor/cpp/xmlvst_gennsd.cpp:79: warning: base class `class XmlVisitor' should be explicitly initialized in the copy constructor

It sort of works on Linux, but not completely:

By converting the .bat file to a bash script and using forward slashes
in the config file paths, I was able to process the files in
MoritzDistribution_2006_05_29 to produce HTML output.  That seemed to
be okay.

Then I changed Doxyfile_xml to point to the sourcecode for Moritz
itself, but it crashes while trying to process the XML that doxygen
created.  What's happening is that in the
XmlVisitor_FunctionBodyAnalyse::analyse function, in the section
marked "collect_BodyLines", it's entering the loop with this data:

  Function->id=class_xml_block___d___config_1048b2c620927a4b2e3841b2100cd28fa
  Function->location->bodystart=22
  Function->location->bodyend=24
  Codeline->lineno=25

Since Codeline is not in the function body, the loop never ends.
Codeline just keeps increasing until it gets to 54 and hits a null
pointer somewhere.  I tried changing the logic to test that
CodeLine < bodyend, which got it to run further, but eventually it
gets to this point:

  function 68 of 395: getCommentText of compound ToolGenNDS_AnlsStrct
  function 69 of 395: ToolGenNDS_AnlsStrct_CmdBf of compound ToolGenNDS_AnlsStrct_CmdBf
  function 70 of 395: ~ToolGenNDS_AnlsStrct_CmdBf of compound ToolGenNDS_AnlsStrct_CmdBf

and then gets stuck.  I let it run for a few minutes but it's just
using 99% of the CPU to do memory mapping and unmapping operations
over and over again for some reason.  I also see some more Pause
commands over in xmlvst_gennsd.cpp that would need to be removed.

                                                  -Dave Dodge