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

[Doxygen-develop] Private and Undocumented Members in UML

[Olaf K. <o_k...@us...>]

Dear Doxygen Developers,

Please find a patch in the attachments that fixes the display of
private and undocumented members in the UML diagrams created with dot.

The issue has been raised before on the doxygen-users list back at
2009-01-15 13:45. No action was taken at that time, however (see e.g.
http://old.nabble.com/Hide-private-members-in-UML-td21477924.html ). As
of version 1.7.1 the same issue annoys me enough to send in this
patch. :)

The patch simply avoids printing the "unlinkable" members, if
HIDE_UNDOC_MEMBERS has been specified and does not print any private
members, unless EXTRACT_PRIVATE is set to true. I've tested the four
obvious combinations and it seems to work fine for all of them.

Best regards,
    Olaf

[Doxygen-develop] Automatic image file dependency generation patch

[Robert J. E. <rob...@gm...>]

This patch adds the HTML_IMAGE_DEPENDENCY_FILE_NAME and
LATEX_IMAGE_DEPENDENCY_FILE_NAME tags.  For each of these tags that
are defined, warnings about missing image files will be supressed (for
the related output type --- i.e. LaTeX or HTML), and a list of used
image files will be saved as prerequisite to a relevant make rule in a
make file named by the value of that tag.

For usage examples, see the included doxygen comments in the patch.

Please note that we included changes made to the file
optionsconfig.cpp although it is automatically generated from the
config.xml file.  We did this since that file is under revision
control (you might want to change that).

Regards Robert Jørgensgaard Engdahl and Anders Franz Terkelsen

p.s. We sent this patch previously, but since the sender address was
not on the mailing list the mail did not get through. I hope that we
are not offending anyone by sending this patch again.

[Doxygen-develop] inline template documentation

[Andrey A. <asa...@gm...>]

hello.

How hard would it to implement in line template documentation:

template<
  class F ///< documentation
>

can you please tell me where to start, I not familiar with internals.
Thank you

[Doxygen-develop] Adding file name to caller graph nodes

[Jan E. <jan...@we...>]

Hi,

I'm mostly interested in call- and caller-graphs generated by doxygen. I
would like to add the filename of a function (i.e. where it is defined)
to its node in the graph.
Is there an easy way of doing this?
If not, where in the source code should I look to find my way?

Thanks for any efforts
Jan

Re: [Doxygen-develop] Doxygen messes up with PRINTF_ATTRIBUTES macro

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

Hi Andreas,


On 7 mei 2010, at 21:37, Andreas Schneider wrote:

> On Friday 07 May 2010 20:48:26 you wrote:
>> Hi Andreas,
>>
>> On 7 mei 2010, at 10:32, Andreas Schneider wrote:
>>> Hi,
>>>
>>> I've created API documentation for talloc.samba.org. Doxygen has a
>>> problem
>>> correctly creating documentation for a function which has a printf
>>> attribute
>>> checking at the end using a macro.
>>>
>>> #define PRINTF_ATTRIBUTE(a1, a2) __attribute__ ((format (__printf__,
>>> a1, a2)))
>>> #else
>>> #define PRINTF_ATTRIBUTE(a1, a2)
>>> #endif
>>> #endif
>>>
>>> void *talloc_init(const char *fmt, ...) PRINTF_ATTRIBUTE(1,2);
>>>
>>> It adds the return value of this function to the output of the next.
>>>
>>> void * 	talloc_init (const char *fmt,...) PRINTF_ATTRIBUTE(1
>>>
>>> 	Create a new top level talloc context.
>>>
>>> void *int 	talloc_free (void *ptr)
>>>
>>> 	Free a chunk of talloc memory.
>>>
>>> See http://talloc.samba.org/talloc/doc/html/group__talloc.html
>>>
>>> I can work around the problem using #ifdef DOXYGEN but I think it
>>> should be
>>> fixed in doxygen.
>>
>> You can configure doxygen's preprocessor such that no #ifdef's are
>> needed to
>> parse the code properly.
>
> Do you have an example how to do it?
>

Given this:

--------------------------------------------------------------
#ifdef SOMETHING
#define PRINTF_ATTRIBUTE(a1, a2) __attribute__ ((format (__printf__,  
a1, a2)))
#else
#define PRINTF_ATTRIBUTE(a1, a2)
#endif

/// Blah
void *talloc_init(const char *fmt, ...) PRINTF_ATTRIBUTE(1,2);
--------------------------------------------------------------

This should work:

ENABLE_PREPROCESSING   = YES
MACRO_EXPANSION        = YES
EXPAND_ONLY_PREDEF     = NO
PREDEFINED             =

But you can also use this:

ENABLE_PREPROCESSING   = YES
MACRO_EXPANSION        = YES
EXPAND_ONLY_PREDEF     = YES
PREDEFINED             = PRINTF_ATTRIBUTE(x,y)=

See http://www.stack.nl/~dimitri/doxygen/preprocessing.html for more  
info.
Use doxygen with the -d Preprocessor option to see the result after  
preprocessing.

Regards,
   Dimitri

Re: [Doxygen-develop] Doxygen messes up with PRINTF_ATTRIBUTES macro

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

Hi Andreas,

On 7 mei 2010, at 10:32, Andreas Schneider wrote:

> Hi,
>
> I've created API documentation for talloc.samba.org. Doxygen has a  
> problem
> correctly creating documentation for a function which has a printf  
> attribute
> checking at the end using a macro.
>
> #define PRINTF_ATTRIBUTE(a1, a2) __attribute__ ((format (__printf__,  
> a1, a2)))
> #else
> #define PRINTF_ATTRIBUTE(a1, a2)
> #endif
> #endif
>
> void *talloc_init(const char *fmt, ...) PRINTF_ATTRIBUTE(1,2);
>
> It adds the return value of this function to the output of the next.
>
> void * 	talloc_init (const char *fmt,...) PRINTF_ATTRIBUTE(1
> 	Create a new top level talloc context.
>
> void *int 	talloc_free (void *ptr)
> 	Free a chunk of talloc memory.
>
>
> See http://talloc.samba.org/talloc/doc/html/group__talloc.html
>
> I can work around the problem using #ifdef DOXYGEN but I think it  
> should be
> fixed in doxygen.

You can configure doxygen's preprocessor such that no #ifdef's are  
needed to
parse the code properly.

Regards,
   Dimitri

[Doxygen-develop] Doxygen messes up with PRINTF_ATTRIBUTES macro

[Andreas S. <ma...@cy...>]

Hi,

I've created API documentation for talloc.samba.org. Doxygen has a problem 
correctly creating documentation for a function which has a printf attribute 
checking at the end using a macro.

#define PRINTF_ATTRIBUTE(a1, a2) __attribute__ ((format (__printf__, a1, a2)))                                                                               
#else                                                                                                                                                        
#define PRINTF_ATTRIBUTE(a1, a2)                                                                                                                             
#endif                                                                                                                                                       
#endif

void *talloc_init(const char *fmt, ...) PRINTF_ATTRIBUTE(1,2); 

It adds the return value of this function to the output of the next.

void * 	talloc_init (const char *fmt,...) PRINTF_ATTRIBUTE(1
 	Create a new top level talloc context.

void *int 	talloc_free (void *ptr)
 	Free a chunk of talloc memory. 


See http://talloc.samba.org/talloc/doc/html/group__talloc.html

I can work around the problem using #ifdef DOXYGEN but I think it should be 
fixed in doxygen.


Cheers,


	-- andreas

-- 
cybernetic synapses - http://www.cynapses.org/

[Doxygen-develop] [PATCH] Fix anchors for docset generation and provide publisher information

[Guido T. <gt...@od...>]

Hi,

here's a tiny patch against the doxygen trunk that fixes XCode docset generation.  It generates <Anchor> tags in Nodes.xml instead of using url#anchor syntax (which docsetutil does not support).  I've also added DOCSET_PUBLISHER_ID and DOCSET_PUBLISHER_NAME as config options, because Apple suggests to always provide these in Info.plist (otherwise the docset will be filed under "Unknown publisher"):
http://developer.apple.com/mac/library/documentation/DeveloperTools/Conceptual/Documentation_Sets/030-Configuring_Documentation_Sets/configuring_docsets.html#//apple_ref/doc/uid/TP40005266-CH11-SW22

Best regards,
	Guido

[Doxygen-develop] @package declaration in PHP class declaration

[Michael C. <mic...@gm...>]

Hello,

I'd like to document a PHP homemade framework. I wrote a DocBlock at the top
of each class, like this:


-------------------------------------------------------------------------------------
/*
 * Brief description of my class
 *
 * Detailed description
 *
 * @author Michael Courtier<mic...@gm...>
 * @package html
 *
 */
class HTML_Element {
   ...
}
-------------------------------------------------------------------------------------


I generate the documentation in XML.

in index.xml, the main file generated which sums up the whole framework, I
can see my "html" package in this xml document :
      <compound refid="d5/d71/namespacehtml"
kind="namespace"><name>html</name></compound>

I assume the @package attribute is in the right DocBlock. But now, in my
detailed xml file (the one for the class HTML_Element), the brief
description and detailed description nodes are empty !!

if I delete the @package attribute, then it works, and I have the brief and
detailed description in my xml file. Am I documenting this class correctly ?

Sincerely

Michael Courtier

Re: [Doxygen-develop] Bug 613616 - PREPROCESS is broken with complex code

[Kirby Z. <kir...@gm...>]

It seems works well.

-----Original Message-----
From: Dimitri Van Heesch [mailto:do...@gm...] 
Sent: Thursday, March 25, 2010 4:53 AM
To: Kirby Zhou
Cc: dox...@li...
Subject: Re: [Doxygen-develop] Bug 613616 - PREPROCESS is broken with
complex code

Hi Kirby,

On 23 mrt 2010, at 15:26, Kirby Zhou wrote:

> https://bugzilla.gnome.org/show_bug.cgi?id=613616
>
>
> I have tested following old versions of doxygen, all of them have  
> the same bug.
>
> doxygen-1.4.7
> doxygen-1.6.1
> doxygen-1.6.2
>
> Bug seems stay here:
>
> pre.l
>  188 static bool macroIsAccessible(Define *def)
>  201     bool b = g_inputFileDef->includes(def- 
> >fileDef,&includedFiles);
>
> g_inputFileDef donot know files that not directly included by itself  
> and being processed in prior files.
> For example:
> If A includes B, B includes C, C includes D, while A is processed  
> first and then B, The macros defined in D will be hidden to B.
>
> Anyone can correct that bug?

Can you check whether or not the bug is fixed with the update I just  
committed to the subversion repository?

Regards,
   Dimitri

Re: [Doxygen-develop] Bug 613616 - PREPROCESS is broken with complex code

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

Hi Kirby,

On 23 mrt 2010, at 15:26, Kirby Zhou wrote:

> https://bugzilla.gnome.org/show_bug.cgi?id=613616
>
>
> I have tested following old versions of doxygen, all of them have  
> the same bug.
>
> doxygen-1.4.7
> doxygen-1.6.1
> doxygen-1.6.2
>
> Bug seems stay here:
>
> pre.l
>  188 static bool macroIsAccessible(Define *def)
>  201     bool b = g_inputFileDef->includes(def- 
> >fileDef,&includedFiles);
>
> g_inputFileDef donot know files that not directly included by itself  
> and being processed in prior files.
> For example:
> If A includes B, B includes C, C includes D, while A is processed  
> first and then B, The macros defined in D will be hidden to B.
>
> Anyone can correct that bug?

Can you check whether or not the bug is fixed with the update I just  
committed to the subversion repository?

Regards,
   Dimitri

[Doxygen-develop] Bug 613616 - PREPROCESS is broken with complex code

[Kirby Z. <kir...@gm...>]

https://bugzilla.gnome.org/show_bug.cgi?id=613616

 

 

I have tested following old versions of doxygen, all of them have the same
bug.

 

doxygen-1.4.7

doxygen-1.6.1

doxygen-1.6.2

 

Bug seems stay here:

 

pre.l

 188 static bool macroIsAccessible(Define *def)

 201     bool b = g_inputFileDef->includes(def->fileDef,&includedFiles);

 

g_inputFileDef donot know files that not directly included by itself and
being processed in prior files.

For example:

If A includes B, B includes C, C includes D, while A is processed first and
then B, The macros defined in D will be hidden to B.

 

Anyone can correct that bug?

 

 

[Doxygen-develop] Bug 398942 - Doesn't detect this.* references correctly

[Antonio T. <ter...@so...>]

Hello,

I've just attached a patch that fixes this bug, and found that the bug
is not a duplicate of the bug it is marked as duplicate of.

https://bugzilla.gnome.org/show_bug.cgi?id=398942

It would be nice if someone review it and eventually commit to the
repository.

-- 
Antonio Terceiro <ter...@so...>
http://softwarelivre.org/terceiro

[Doxygen-develop] Feature enhancement to image tag

[Neil T. <nei...@ta...>]

Could the image tag options provide access to the html <IMG> parameters 
that define the size of the displayed image i.e HEIGHT and WIDTH

If the image being displayed is large then it falls off the edge of the 
screen.

I'd like to display smaller ones that fit on the screen by defining the 
size.

If I could then click on these smaller images to get a full scale one 
then even better.

Thanks

=======================================================================
This email, including any attachments, is only for the intended
addressee.  It is subject to copyright, is confidential and may be
the subject of legal or other privilege, none of which is waived or
lost by reason of this transmission.
If the receiver is not the intended addressee, please accept our
apologies, notify us by return, delete all copies and perform no
other act on the email.
Unfortunately, we cannot warrant that the email has not been
 altered or corrupted during transmission.
=======================================================================

[Doxygen-develop] Patch for generating correct call/caller graphs and function references if conditional compilation is used

[Martin G. <mge...@gm...>]

Hello everyone!

I recently found out that when generating call / caller graphs even
function calls which would be removed by the preprocessor (due to
conditional compilation) get included in the call graph. In some cases
this may not be desired. Consider the following example:

test.cpp:

#include <iostream>
using namespace std;

#define FUNC1

void func1(int test)
{
        cout << "func1: Hallo" << endl;
}
void func2(int test)
{
        cout << "func2: Hallo" << endl;
}

int main() {

#ifdef FUNC1
        func1(0);
#else
        func2(0);
#endif

        return 0;
}


Doxygen produces something like this for the main function when
generating xml output:

<references refid="test_8cpp_1a5d7f916959f4f32d1f8d38816c20fbef"
compoundref="test_8cpp" startline="7" endline="10">func1</references>
<references refid="test_8cpp_1af40a1280e9eecc3ea5aa6e37a4b41f65"
compoundref="test_8cpp" startline="11" endline="14">func2</references>

But the second function (func2) gets never called.

The patch modifies the preprocessor in "pre.l" so that all line numbers
which would be removed by the preprocessor are stored in a dictionary.
This information is then used by "code.l" to check if function gets
actually called. This information could also be used for other purposes
in future. E. g. different background color for lines excluded by the
preprocessor like eclipse does.

With this patch I tried the least invasive approach. Only a small
modification to "prel.l" and "code.l" was necessary. The only thing
missing at the moment is a configuration option to turn this patch off,
because sometimes it could be useful to see all functions that could be
called.
Please test this patch and let me know if it breaks some things or
doesn't work for you as expected.  If this patch will be included in
doxygen I'll also provide the missing configuration option.

The patch and the test case can be found at:

http://mmg.ath.cx:8081/cond_compilation_patch.zip

Regards,
Martin Gerlach.
 

[Doxygen-develop] Patch for Ruby parsing

[Vincent F. <vin...@9o...>]

  Hello dear doxygen users (and developers) !

  I'm a fervent doxygen user, but until recently I found that it was
missing something for me: support for parsing Ruby files. So I spent
some time dwelling in the internals of Doxygen, I'm probably far from
actually understanding anything at all, but I've implemented something
that works.

  If there are Ruby users around, I'd be grateful if you could try it
out and flame me because it doesn't work for your code...

  Patches are available at

http://www.normalesup.org/~fourmond/doxygen-ruby/

  Get the latest one (for now -1.patch.gz) and follow the instructions
(you need to be able build doxygen from source).

  My hope is that this patch will receive sufficient testing to be good
enough to make it into the official distribution.

  Side notes: I've tried to keep the compatibility with the main tool
used for documenting Ruby code, rdoc. This makes doxygen behave
more-or-less like rdoc does, but not like doxygen does on, say C++ code.
This can be disable by setting the RUBY_RDOC_COMPATIBILITY configuration
variable to false.

  Cheers, and happy testing !

	Vincent

PS: please don't CC me, I'm subscribed to both lists.

-- 
Vincent Fourmond, Doctor in Physics
http://vince-debian.blogspot.com/

[Doxygen-develop] Problem in class collaboration diagram when using template members

[Keshetti M. <kes...@gm...>]

I have recently observed that Doxygen is not showing class collaboration
diagram properly in some situations like the flowing.
suppose there is a class A which contains vector of objects of class B.

class B {
int data;
};

class A {
int data;
std::vector<B> b;
};

'b' is not shown in the collaboration diagram of class A in this
situation.

I have searched bugzilla but didn't find any match. Is this known behavior ?

-- 
Keshetti Mahesh

[Doxygen-develop] Method Internals

[joaomm <joa...@gm...>]

Hi,

I'm developing a tool that reports some information collected by Doxygen.
It's called Doxyparse (
http://gitorious.org/doxygen/mainline/commits/doxyparse).
It's used to calculate metrics such as number of functions, number of public
functions, coupling, depth of inheritance tree and number of childs.

Recently, I'm trying to collect information about method internals such as
Cyclomatic Complexity. Is that possible?

Having the MemberDef or the Definition of the method, can I at least get its
source code?

Thanks,

[Doxygen-develop] Plurals

[Jason M. <ko...@gm...>]

Doxygen's autolinks are great. However, if I have an object called 
"SomeObject", it would be very useful if I could talk about 
"SomeObjects" and have that properly link without having to explicitly 
reference it or other kinds of things.

[Doxygen-develop] Japanese translation update

[goyoki <go...@gm...>]

Hi,
translator_ja attached.(Minor fix and update for 1.6.0)

---
Hiroki Iseri

[Doxygen-develop] [PATCH] - Alias definition and arguments can now contain "{" and "}"

[Stefan O. <st...@ob...>]

Its not possible
   to pass a "{" as an argument to an alias or
   to put a "{" into a alias definition.

eg:
Doxyfile:
	ALIASES                = "MACRO_ARGV1{1}=\1"
	ALIASES               += "MACRO_ARGV2{1}=\{\1"
	ALIASES               += "MACRO_ARGV3{1}=\{\1\}"
	ALIASES               += "MACRO_ARGV4{1}=\1\}"

example page:
/*! @page page_argv ARGV Macros

Test of ARGV macros with { and }\n
@MACRO_ARGV1{ABCD\{ab\}cd}\n
@MACRO_ARGV2{ABCD\{ab\}cd}\n
@MACRO_ARGV3{ABCD\{ab\}cd}\n
@MACRO_ARGV4{ABCD\{ab\}cd}\n
*/

After applying the attached patch escaping of brackets works.
Use "\{" and "\}".



Best regards

Stefan

[Doxygen-develop] [PATCH] - LATEX - Parameternames and Parametertypes generated in explicit layoutstyle(environment) "DoxyParamCaption"

[Stefan O. <st...@ob...>]

This patch allows you to overwrite the layout of sections of modules within
the LATEX output.


The goal is to create a tabular look-like parameterlist - like in HTML.


eg:
  1.2.3 int foo(int param1, int param2)

can be converted into
  1.2.3 int foo(
                 mytype param1,
                 int param2
                )

The result is a list which uses the DoxyParamCaption environment.
example:
\subsubsection[{foo}]{\setlength{\rightskip}{0pt plus 5cm}{int} foo (
\begin{DoxyParamCaption}
\item[{mytype}}]{ param1, }
\item[{int}}]{ param2}
\end{DoxyParamCaption}
)}

For compatibliy reasons the default "DoxyParamCaption" environment in
doxygen.sty will produce the same style the output had before.

You can overwrite the DoxyParamCaption (use \renewenvironment) to generate
your own layout.


Best regards

Stefan

[Doxygen-develop] PHP5.3 support how to

[Andrew M. <an...@mi...>]

Hello all,
I'm trying to  add the php 5.3 namespace support in doxygen and thought I 
would just hit the mailing list to make sure that I am understanding correctly 
what I need to do. 

From what I understand it uses flex / yacc (bison?) to parse the source code in 
scanner.l ? This generates some C code that which executes a function when it 
encounters one of the tokens. 

I  can see the token declaration  that are in scanner.l which i need to create 
however I am having some difficulty finding the CopyPHPString() which I think is 
the function i need to modify in order for it to handle the namespace 
declaration. 

Would appreciate any advice / direction / input. I haven't worked with flex / 
yacc nor really done that much C before so apologies if what I am asking seems 
obvious. 

thanks in advance
Andrew

[Doxygen-develop] [PATCH] Formula transparency patch.

[E. R. <er...@ca...>]

I've attached a patch to render formulas slightly differently. The current 
method of rendering is to render the formula to a white background, then output 
the exact colour as opaque, except for pure white which is set to fully 
transparent.

The new method (attached) renders all non white pixels as pure black, but sets 
various levels of transparency to achieve a smooth effect. This ensures that 
the equations render correctly if the web browser background is not set to 
white.

I haven't quite figured out how to set it up so that it can be switched on with 
a configuration option.

To illustrate the method with a rather exxtreme example, I've composited a 
sample formula against a dark blue background, and attached the images.

Regards

-Ed


-- 
(You can't go wrong with psycho-rats.)(http://mi.eng.cam.ac.uk/~er258)

/d{def}def/f{/Times s selectfont}d/s{11}d/r{roll}d f 2/m{moveto}d -1
r 230 350 m 0 1 179{ 1 index show 88 rotate 4 mul 0 rmoveto}for/s 12
     d f pop 235 420 translate 0 0 moveto 1 2 scale show showpage

Re: [Doxygen-develop] Aliases

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

On Fri, 2009-07-31 at 10:41 +0100, John Tapsell wrote:
> Hey all,
> 
>   At the moment you can't use cond inside an alias, like:
> 
> ALIASES = foodev="\cond FooDev" \
>           endfoodev="\endcond"
> 
> (http://bugzilla.gnome.org/show_bug.cgi?id=485773)
> 
> Does anyone know of a work around?

as a workaround you could use the INPUT_FILTER option to pass your
source through a small (your choice of scripting language) script which
does the conditional checks for you... might be a bit of work though.

> Can anyone give a hint how I could fix this bug in doxygen?
> 
> John Tapsell

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

You may be right, I may be crazy,
But it just may be a lunatic you're looking for!
		-- Billy Joel