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

[Doxygen-develop] Link fault in "List of all members" (patch)

[Johan E. <Joh...@ua...>]

Hi,
   I've found a bug in the "List of all members" section, the generated
link for e.g. a method is wrong when the member is part of a group. The
link links to the corresponding class' file instead of the group file.

I've attached a patch for this......

//Johan

[Doxygen-develop] Suggestion for a @sideeffect tag

[Peter N. L (QRA) <Pet...@er...>]

Hello,

Here comes a newbie request.
I hope that I have not missed anything in the documentation.
I tried to find out how to make a custom tag, but I didn't find any info about that.

In any case, I think that there should exist a tag @sideeffect.
It should work in the same way as
@pre, @post and @invariant and the purpose would be to document what side-effect
a function call has.

Best regards,
Peter Nordlund

Re: [Doxygen-develop] Bug? Redundant generation of inner class docs

[Dimitri v. H. <di...@st...>]

On Thu, Oct 18, 2001 at 02:45:50PM -0300, Brad Spencer wrote:
> After running doxygen-1.2.11.1 over my (large) source tree, I found
> that it was taking a _long_ time to finish writing out the namespace
> docs.  After a bit of poking around and debugging, I narrowed it down
> to line 5816 of doxygen.cpp, inside generateNamespaceDocs.
> 
> It looks like the loop construct there was copied from
> generateClassDocs.  Anyway, it ends up calling
> ClassDef::writeDocumentationForInnerClasses for each class in the
> dictionary over again for every namespace!  As far as I can tell, this
> results in these inner classes' documentation being regenerated
> identically over and over again.  I'm providing a patch to remove this
> call since it has definately already been done by the time we get to
> generateNamespaceDocs.  I haven't seen any trouble with the output so
> far.

This is wrong indeed. I'll use a different fix to make sure 
nested compounds are always written to a tag file after their parent.
But thanks for pointing out the problem.

Regards,
  Dimitri

[Doxygen-develop] Bug? Redundant generation of inner class docs

[Brad S. <sp...@in...>]

After running doxygen-1.2.11.1 over my (large) source tree, I found
that it was taking a _long_ time to finish writing out the namespace
docs.  After a bit of poking around and debugging, I narrowed it down
to line 5816 of doxygen.cpp, inside generateNamespaceDocs.

It looks like the loop construct there was copied from
generateClassDocs.  Anyway, it ends up calling
ClassDef::writeDocumentationForInnerClasses for each class in the
dictionary over again for every namespace!  As far as I can tell, this
results in these inner classes' documentation being regenerated
identically over and over again.  I'm providing a patch to remove this
call since it has definately already been done by the time we get to
generateNamespaceDocs.  I haven't seen any trouble with the output so
far.

I've attached the brief patch.
	
-- 
-----------------------------------------------------------------
Brad Spencer - sp...@in... - "It's quite nice..."
  Systems Architect | InfoInterActive Corp. | An AOL Company

[Doxygen-develop] alpahbetical index for c based projects

[Stefan K. <ko...@im...>]

Hi,

I would like to specify in the Doxyfile what to add to the ALHABETICAL_INDEX.
I currently work on some c-based libraries. Personally I would like to see *everything* there (no just structs, but functions, structs, defines, enums,...)

Stefan
-- 
       \|/
      <@ @> Stefan Kost  private                   business
+-oOO-(_)-OOo------------------------------------------------------------- - - -  -   -
|        __    Address  Zwenkauer Str. 24         HTWK Leipzig, Fb IMN, Postfach 300066
|       ///             04277 Leipzig             04277 Leipzig
|  __  ///              Germany                   Germany
|  \\\///      Phone    +49341 3910483            +49341 3076101
|   \__/       EMail    st...@gm...           ko...@im...
|              WWW      http://www.sonicpulse.de  http://www.imn.htwk-leipzig.de/~kost/about.html
===-=-=--=---=---------------------------------- - - -  -    -

[Doxygen-develop] Slovene translation - polished version

[Matjaz O. <mat...@zr...>]

Hi !

Petr suggested some changes in translator_si.h. The file is fixed now.
Changes are:
- no translator.h include
- obsolete methods removed

Regards

Matjaz

[Doxygen-develop] Doxygen: translator.pl update

[Prikryl,Petr <PRI...@sk...>]

Hi Dimitri and others,

Here is the updated translator.pl

# 2001/10/17
#  - Minor update of GetInfoFrom() to ignore spaces between the
#    method identifier and the opening parenthesis to match better
#    the method prototype with the one in the translator.h.
 <<translator.pl>> 
See you,
  Petr
-- 
Petr Prikryl, SKIL, spol. s r.o., pri...@sk...

[Doxygen-develop] Slovene translation

[Matjaz O. <mat...@zr...>]

Hi all !

The atached file is a Slovene translation of Doxygen.

Regards

Matjaz 

[Doxygen-develop] Chinese translation update

[Wei L. <li...@as...>]

Hi,

I have finally updated Chinese translation for doxygen-1.2.11. 
Here is two patch file included as an attachement.

Charlie Liu

[Doxygen-develop] Chinese translation update

[Wei L. <li...@as...>]

Hi,

I have finally updated Chinese translation for doxygen-1.2.11. 
Here is two patch file included as an attachement.

Charlie Liu

[Doxygen-develop] Slovak translation update

[Stanislav K. <sk...@po...>]

I have finally updated slovak translation for doxygen-1.2.11. Here is=20
updated version of translator_sk.h included as an attachement.

Stanislav Kudlac
=0A____________________________________=0AP. S. V. P. U.=0Ahttp://www.pobox=
.sk/=0A=0A

[Doxygen-develop] index out of range error

[Berno L. <ber...@ep...>]

Hello,

has anyone succeeded in feeding the Qt-Lib into doxygen? I tried it with the 
qt-copy of KDE and the current cvs checkout of Doxygen and get many errors 
like "QGArray::at: Absolute index -1 out of range".

The file of Qt that causes the error is qmotifdnd_x11.cpp.

The error seems to occur in the method "getResolvedClass" of utils.cpp in the 
while-statement after one of the comments "// strip * and & from n" (ca. line 
476).

After uncommenting the printf lines, the last words of doxygen are:
    getResolvedClass(DndTargetsTableRec,DndTargetsTableEntry)
      typedef subst=`*'
    QGArray::at: Absolute index -1 out of range
    QGArray::at: Absolute index -2 out of range
    QGArray::at: Absolute index -3 out of range
    ...

Checking whether ip is greater/equal 0 supresses the error but I don't know 
whether that makes sense:
while (ip>=0 && (subst.at(ip)=='*' || subst.at(ip)=='&' ||
                 subst.at(ip)==' ')) ip--;

---
Gruß

Berno

[Doxygen-develop] parser bug?

[<tw...@im...>]

Hallo,

I write a documentation, an will NOT, that a name
ist linked, then I must put an % in front of the
word (I have seen that in the FAQ) like:

/**
*
...

* this document the table %keyaccount witch is used for
* ..
*/

then it is correctly not linked. 

Now I will make the word bold for better understanding like

/**
*
...

* this document the table \b %keyaccount witch is used for
* ..
*/

Now the Word ist not linked an NOT bold. The follow word (witch) is
now bold. 
Other cobinations with \b and % may have more bugs!

Version of doxygen: 1.2.10-20010909

Thanks

Waffel

[Doxygen-develop] (no subject)

[Jose L. Z. <jlz...@ca...>]

-- 
José Luis Zabalza
jlz...@ca...
Linux User 172551

[Doxygen-develop] latex item generation bug (1.2.10)

[Xander B. <xb...@fo...>]

Hi,

I think I found a bug in the latex generation part. The offending piece of 
source:

<DL>
<DT>AD1</DT>  HCI Detailed Design (EDO-HCI-NT-003), July 4th,2 001.
<DT>AD2</DT>  EuroSim GUI Framework Design (EDO_GUI_NT_001), April 26st, 2001.
</DL>

This results in:
 \begin{description}
 \item[AD1 HCI Detailed Design (EDO-HCI-NT-003), July 4th, 2001. \item[AD2 
Euro\-Sim GUI Framework Design (EDO\_\-GUI\_\-NT\_\-001), April 26st, 2001. 
\end{description}

The closing "]" bracket is missing. LaTeX of course does not like this very 
much and reports:

Runaway argument?
AD1 HCI Detailed Design (EDO-HCI-NT-003), Issue 0 Rev C, July 4th, 20\ETC.
! Paragraph ended before \@item was complete.
<to be read again>
                   \par
l.24
 
?

Or am I missing something here?
	sincerely,

Xander Burgerhout
Fokker Space BV Leiden
The Netherlands

[Doxygen-develop] some ideas for future doxygen versions

[Stefan K. <ko...@im...>]

hi hi,

at first a big thanks to all who made doxygen to what it already is.
I am currently using doxygen for a pure c as well as for a java project.
I would like to see the following things in new versions and if it it okay I'll do it myself and sent in diffs.
1. can all lists (except the alphabetical index) have delimiters for each letter.
2. can  these delimiters not look like :
     "--a--"
    I would suggest :
    [a][b][c]...m...[z]
    with all letters in [] beeing links to the local achor and the letter without the [] beeing the letter
    where we currently are. The 'm' in this example should llok like the letters in the alpahbetical index (inverse).

Further there is some idea, which definitely should be discussed first. Lets call it doc-templates. You know in OO
development you usually end up with lots of set/get methods. It's usually a pain in the ass to document them.
It would be great if I could tell doxy that everytime it hits a method beginning with e.g. "set" and having
just one parameter it sould take the comment from a certain template and insert the field from the source-code there.

Another issue I am curently thinking about is EJB. The java project is using EJB and this means I have lots
of classes with have simmiliar names, but not simmiliar enough that doxygen could easily find out that one
defines the interface to another.

Ciao
   Stefan
-- 
       \|/
      <@ @> Stefan Kost  private                   business
+-oOO-(_)-OOo------------------------------------------------------------- - - -  -   -
|        __    Address  Zwenkauer Str. 24         HTWK Leipzig, Fb IMN, Postfach 300066
|       ///             04277 Leipzig             04277 Leipzig
|  __  ///              Germany                   Germany
|  \\\///      Phone    +49341 3910483            +49341 3076101
|   \__/       EMail    st...@gm...           ko...@im...
|              WWW      http://www.sonicpulse.de  http://www.imn.htwk-leipzig.de/~kost/about.html
===-=-=--=---=---------------------------------- - - -  -    -

[Doxygen-develop] Macros II

[Stefan M. <me...@sk...>]

Hi,

I've advanced somewhat in my quest to make the stuff work.

I set the macro expansion

PREDEFINED             = "IWMETHODIMP(x)=x __stdcall " \
                         "PURE= = 0"

and

EXPAND_AS_DEFINED      = "IWMETHODIMP(x)=x __stdcall " \
                         "PURE= = 0"

and enabled all macro expansion like this :

ENABLE_PREPROCESSING   = YES
MACRO_EXPANSION        = YES
EXPAND_ONLY_PREDEF     = YES

I now get this error message :

D:/Projekte/IW/Src/baselib/Random/Random.h:48 Warning: no matching class
member
found for
  IWMETHODIMP IW::CRandomGenerator::SetSeed(void)

since SetSeed is defined as :

	IWMETHODIMP(void) SetSeed( ulong _uSeed );

I suppose that the macro is either

- not expanded as should

or

- I made a fault within the PREDEFINED declaration.

If anyone had a suggestion what I could do, I appreciate.

Kind regards,
Stefan Maton
-----------------------------
Programmer Resources Site
http://www.sun-a-moon.org
-----------------------------

[Doxygen-develop] Macros

[Stefan M. <me...@sk...>]

Hi,

I have following class definition :


// ----------------------------------------------------------
// Class definition.
class CRandomGenerator : public IRandomGenerator
{

public:
[...snipped code...]

	// ------------------------
	// Seed setting.

	IWMETHODIMP(void)SetSeed( ulong _uSeed );
	IWMETHODIMP(ulong)GetSeed();

[...snipped code...]
};

IWMETHODIMP is defined within another header that is not
included in the randomgenerator header in this way :

#define IWMETHODCALLTYPE __stdcall
#define IWMETHODIMP(type)		type IWMETHODCALLTYPE

Unfortunately, doxygen doesn't seem to be able to resolve following
comment :

/*! \fn IWMETHODIMP(void) CRandomGenerator::SetSeed( ulong _uSeed );
 *  \brief Set the seed of the random number generator.
 *  \param _uSeed The seed to use.
 */

I also tried to put the comment just above the SetSeed function declaration
in this way :

/*! \fn IWMETHODIMP(void) SetSeed( ulong _uSeed );
 *  \brief Set the seed of the random number generator.
 *  \param _uSeed The seed to use.
 */

Which didn't work, too...

The warning I get is :
D:/Projekte/IW/Src/baselib/Random/Random.h:38 Warning: member SetSeed of
class C
RandomGenerator cannot be found

What can I do ?

Kind regards,
Stefan Maton
-----------------------------
Programmer Resources Site
http://www.sun-a-moon.org
-----------------------------

[Doxygen-develop] Picky \example

[Fabian C. <Cen...@in...>]

Hi

I have a comment block to a class and try to link to a different file by
\example. But the example paragraph in the class documentation only
shows up if the name of this class is mentioned inside the example file.
In the example file list it is always there, no matter what file it is, but
not in the class documentation. Is this intentional? And if yes, why?

Thanks

bye  Fabi

[Doxygen-develop] latex error (1.2.9.1)

[<nb...@fr...>]

I get this:

In annotated.tex:
[...]
\item\contentsline{section}{{\bf Demod\-Stats} }{\pageref{structDemodStats}}{}
\item\contentsline{section}{{\bf Error}  (\begin{Desc}
\item[{\bf Id: }]\par
 Error.H,v 1.1 2001/04/25 19:41:53 nbecker Exp \end{Desc}
)}{\pageref{structError}}{}
\item\contentsline{section}{{\bf F} }{\pageref{structF}}{}
Runaway argument?
{{\bf Error} (\begin {Desc} \item [{\bf Id: }]
! Paragraph ended before \@dottedtocline was complete.
<to be read again> 
                   \par 
l.26 \item[{\bf Id: }]\par

Here is Error.H:
                          
#ifndef Error_H
#define Error_H

//! $Id: Error.H,v 1.1 2001/04/25 19:41:53 nbecker Exp $

struct Error {
  static void Err (const char* msg, ...);
  static void Warn (const char* msg, ...);
};

#endif

[Doxygen-develop] Index item name

[Fabian C. <Cen...@in...>]

Hi

I have a module defined like this:

/*! \defgroup a_b a b   */

The page itself shows as title "a b" as well as the item in the tree view.
But if I generate HTML-help and compile it with Help Workshop, the item
in the index is a_b. Is this an error? Shouldn't the index base on the given
names instead of the internal identifiers?

That wouldn't matter if I could have a space in the group name like:

/*! \defgroup "a b" a b   */

or something like that, but I know it's not possible now.

Second is documentation (A really good one I have to say): The chapter
history is kinda out of date (1.2.0 :) Maybe it would be better to remove this
section from the pdf and deliver it as single file. Doesn't matter what you
decide I just thought I mention it.

Thanks.

bye  Fabi

[Doxygen-develop] Re: Predefined problem

[Fabian C. <Cen...@in...>]

> > Hi
> >
> > I want to do something like this: In the config file I place:
> >
> > PREDEFINED   =  "CLASS_MEMBER(c, m)=c_m"
> >
> > so that c and m are expanded and it results in class_member. The 
> problem now
> > is, that they only get expanded if the c and m are single words. If I 
> place
> > them like above it is treated as "c_m" without expansion. Also #c #m or 
> stuff
> > like that didn't help. Is this possible at all?
>
>Yes, but use the correct C syntax for concatenation:
>
>PREDEFINED   =  "CLASS_MEMBER(c,m)=c##_##m"
>
>Look it up in your C manual :-)

I did and I also found these commands but the examples weren't that clear
on how to use. Thanks a lot for your help, I like Doxygen every day more :)

bye  Fabi

PS: The mail archive seems somehow down as there aren't any postings this
month (including mine) which is hard to believe...

Re: [Doxygen-develop] Predefined problem

[Dimitri v. H. <di...@st...>]

On Tue, Aug 07, 2001 at 03:36:39PM +0200, Fabian Cenedese wrote:
> Hi
> 
> I want to do something like this: In the config file I place:
> 
> PREDEFINED   =  "CLASS_MEMBER(c, m)=c_m"
> 
> so that c and m are expanded and it results in class_member. The problem now
> is, that they only get expanded if the c and m are single words. If I place 
> them
> like above it is treated as "c_m" without expansion. Also #c #m or stuff 
> like that
> didn't help. Is this possible at all?

Yes, but use the correct C syntax for concatenation:

PREDEFINED   =  "CLASS_MEMBER(c,m)=c##_##m"

Look it up in your C manual :-)

Regards,
  Dimitri

[Doxygen-develop] Predefined problem

[Fabian C. <Cen...@in...>]

Hi

I want to do something like this: In the config file I place:

PREDEFINED   =  "CLASS_MEMBER(c, m)=c_m"

so that c and m are expanded and it results in class_member. The problem now
is, that they only get expanded if the c and m are single words. If I place 
them
like above it is treated as "c_m" without expansion. Also #c #m or stuff 
like that
didn't help. Is this possible at all?

Thanks.

bye  Fabi

[Doxygen-develop] XML DTD for DOXYGEN

[Michael L. <li...@ea...>]

Hi,

I checked the new XML output possibility of Doxygen. I'm missing following
items:

- The DTD contains no documentation items, this means: It's not possible to
create a documentation from XML output. You can only create a list of
founded items.
- The DTD also contains no items for the line and file where the item was
found

The XML output shall contain enough information to reproduce the same HTML
output as Doxygen output. That's why it is necessary to store also the
documentation and line information in the XML output.

When do you will insert this feature and update your DTD ? It's also very
helpful (and the fastest way) if you update the DTD first.

Thanks for reply,
Michael