doxygen-users — List for users of doxygen

[Doxygen-users] Doxy for symbols not in the code

[Barnes, P. D. <bar...@ll...>]

Hello Folks,

We want to document symbols which aren’t declared in the code, such as environment variables and preprocessor symbols declared by the build system.

There doesn’t seem to be an easy way to do this.

For example, @var doesn’t generate output if the symbol doesn’t appear somewhere in the code.

We could define nuisance string variables for the names of the environment variables, but this wouldn’t work for preprocessor defines.

I can fake something that looks okay, using html tags, but there doesn’t seem to be a way to make the symbol name appear in search results, so that it’s easy to find.

Thanks for your help,
Peter
_____________________________________________________________
Dr. Peter D. Barnes, Jr. CASC Division, B451 R2035
Lawrence Livermore National Laboratory Computation Directorate
7000 East Avenue, L-561 email:  pdb...@ll...<mailto:pdb...@ll...>
P. O. Box 808 Voice:  (925) 422-3384
Livermore, California 94550

Re: [Doxygen-users] source file consisting of a single include statement

[Richard D. <Ri...@Da...>]

On 4/30/20 4:39 PM, Alan Frankel wrote:
>
> I have a source file, foo.cpp, that consists of a single include
> statement for a single header file, foo.hpp. The header file contains
> all the class and function definitions. The source file is compiled
> into a shared library. By giving preprocessor symbols that appear in
> the header file different values when the source file is compiled,
> different versions of the same shared library can be produced with
> their functionality identical except where the preprocessor symbol
> plays a role. 
>
> The problem is that when I try to produce Doxygen documentation for
> the source file, it's virtually empty. Basically, the only thing it
> does is to generate a file dependency graph for the header file,
> showing  the header files included by the header file. This
> demonstrates that Doxygen knows what those header files are. However,
> it's not listing any of the classes or functions contained in them.
> What should I be doing differently?
>
> Thanks,
>
> Alan
>
Is the header file provided as an input in the INPUT specification to
Doxygen? You only get documentation for the contents of files that are
inputs to doxygen, not files that they include.

-- 
Richard Damon

[Doxygen-users] source file consisting of a single include statement

[Alan F. <afr...@ma...>]

I have a source file, foo.cpp, that consists of a single include statement for a single header file, foo.hpp. The header file contains all the class and function definitions. The source file is compiled into a shared library. By giving preprocessor symbols that appear in the header file different values when the source file is compiled, different versions of the same shared library can be produced with their functionality identical except where the preprocessor symbol plays a role.

The problem is that when I try to produce Doxygen documentation for the source file, it's virtually empty. Basically, the only thing it does is to generate a file dependency graph for the header file, showing  the header files included by the header file. This demonstrates that Doxygen knows what those header files are. However, it's not listing any of the classes or functions contained in them. What should I be doing differently?

Thanks,
Alan

Re: [Doxygen-users] Double underscores are parsed

[cloun <clo...@gm...>]

This solves, thanks.

On 22/04/2020 13:30, Stefan Pendl wrote:
> Am 22.04.2020 um 00:09 schrieb cloun:
>> MARKDOWN_SUPPORT is 1.
>>
>
> Could you try setting it to zero (0)?
>
>

Re: [Doxygen-users] Double underscores are parsed

[Stefan P. <ste...@gm...>]

Am 22.04.2020 um 00:09 schrieb cloun:
> MARKDOWN_SUPPORT is 1.
> 

Could you try setting it to zero (0)?


-- 
*Stefan P.*

Top-posting:
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?

Re: [Doxygen-users] Double underscores are parsed

[cloun <clo...@gm...>]

MARKDOWN_SUPPORT is 1.

On 21/04/2020 21:43, Stefan Pendl wrote:
> Am 20.04.2020 um 22:49 schrieb cloun:
>> Hi. I am using doxygen 1.8.18 and I have such a fragment:
>>
>> @param    caller        supposed to be __func__
>>
>> __func__ is a builtin macro and I expect it to be printed verbatim. 
>> Nevertheless, it is translated as `func' in bold into the HTML 
>> generated documentation, without surrounding underscores. What do I 
>> do wrong?
>
> Do you have mark-down syntax enabled?
>

Re: [Doxygen-users] Double underscores are parsed

[Stefan P. <ste...@gm...>]

Am 20.04.2020 um 22:49 schrieb cloun:
> Hi. I am using doxygen 1.8.18 and I have such a fragment:
> 
> @param    caller        supposed to be __func__
> 
> __func__ is a builtin macro and I expect it to be printed verbatim. 
> Nevertheless, it is translated as `func' in bold into the HTML generated 
> documentation, without surrounding underscores. What do I do wrong?

Do you have mark-down syntax enabled?

-- 
*Stefan P.*

Top-posting:
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?

[Doxygen-users] Double underscores are parsed

[cloun <clo...@gm...>]

Hi. I am using doxygen 1.8.18 and I have such a fragment:

@param    caller        supposed to be __func__

__func__ is a builtin macro and I expect it to be printed verbatim. 
Nevertheless, it is translated as `func' in bold into the HTML generated 
documentation, without surrounding underscores. What do I do wrong?

[Doxygen-users] doxygen 1.8.17 performance regression

[Jasper V. <JVe...@om...>]

Hi,

Is there any known performance regression in generating the documentation between version 1.8.16 and 1.8.17?
Generating the html documentation on our C++ code base in 1.8.16 takes 5h. On 1.8.17 it takes 22h.
We’ve also tried the newly released 1.8.18 with the same result as 1.8.17.
We did not change any configuration options between these versions and I could not find any open performance issues on github.

Jasper


Jasper Verelst
Software Engineer

[phonenumber] +32 3 350 17 27

[mobilenumber] +32 498 01 65 96

jve...@om...
[https://www.omp.com]<https://www.omp.com>




IMPORTANT NOTICE
The information in this e-mail and any attachments is intended for the addressee only. The contents do not represent the opinion of OM Partners n.v. or any of its affiliates except to the extent that it relates to their official business. If you receive this in error, please contact the sender and delete the material from any computer.

Re: [Doxygen-users] Trigger inclusion of <defname> tags in xml output, when different than <declname>

[<sou...@ne...>]

So, after doing some digging, it seems like what controls whether or not
the `<defname>` is included is this line:

https://github.com/doxygen/doxygen/blob/341f860ab7ae3e5255891b0984bec8b36e7557ef/src/doxygen.cpp#L4985


...or more specifically, the `!root->doc.isEmpty()` bit, which sets the
`forceNameOverwrite` arg for `mergeArguments`.

Given that the `MemberDef` object already supports separate declaration
args and definition args (via `declArgumentList` and `argumentList`,
respectively), and that at this point it the code the `declArgumentList` is
already set, it seems odd that setting of `argumentList` is controlled by
whether or not the current root has a doc.  Ideally, I would think we'd
always put the declaration arguments in the `declArgumentList`, and the
definition arguments in `argumentList` - ie, have `forceNameOverwrite` set
to `!root->proto`.

At the very least, if we are going to try to make some fancier logic /
decisions at this point, we should take the `brief` into account as well,
and not just the `doc`.  For instance, it seems particularly odd to not
even include the definition args if the definition has a brief (though no
detail doc), and the declaration has no documentation whatsoever.

[Doxygen-users] Trigger inclusion of <defname> tags in xml output, when different than <declname>

[<sou...@ne...>]

Why is it that my xml output no longer includes <defname> tags, when the
declared and defined names for parameters differ?

For more detail - say I have a method whose parameter names differ between
the declaration and definition:

// sepDeclAndDef.h

class SepDeclAndDef {
    int myFunc(int declaredName);
};

// sepDeclAndDef.cpp

#include "sepDeclAndDef.h"

int SepDeclAndDef::myFunc(int definedName)
{
    return 7;
}

I have some older doxygen output that, in situations like this, would
include a <defname> tag in the xml output:
// class_sep_decl_and_def.xml (old)

...

      <memberdef kind="function" id=
"class_sep_decl_and_def_1aeb6a0f5c0c1e410803fb46a0f48eec36" prot="private"
static="no" const="no" explicit="no" inline="no" virt="non-virtual">
        <type>int</type>
        <definition>int SepDeclAndDef::myFunc</definition>
        <argsstring>(int declaredName)</argsstring>
        <name>myFunc</name>
        <param>
          <type>int</type>
          <declname>declaredName</declname>
          <defname>definedName</defname>
        </param>

However, whenever I try and regenerate the xml now, it never includes the
<defname>:
// class_sep_decl_and_def.xml (old)

...

      <memberdef kind="function" id=
"class_sep_decl_and_def_1aeb6a0f5c0c1e410803fb46a0f48eec36" prot="private"
static="no" const="no" explicit="no" inline="no" virt="non-virtual">
        <type>int</type>
        <definition>int SepDeclAndDef::myFunc</definition>
        <argsstring>(int declaredName)</argsstring>
        <name>myFunc</name>
        <param>
          <type>int</type>
          <declname>declaredName</declname>
        </param>

The older xml was generated using doxygen-1.8.11... however, I've tried
re-generating the xmls using 1.8.11 and current (1.8.18), and I can't
reproduce the <defname> inclusions.

I suspect that there's some configuration change that triggered this, but I
can't find what it is.

(Also, I can't compare the settings directly, because the source code base,
including the doxygen configuration, is not something I have access to - I
am writing a tool / plugin for a product, and the makers of that product
are good enough to supply me with the doxygen-generated xmls, from which I
parse some necessary information.  I was hoping to give them easy
instructions on how to get it to output the <defname> tags again, since I
rely on them, but I can't figure out how to make them myself.)

Looking at the source, ie:

https://github.com/doxygen/doxygen/blob/341f860ab7ae3e5255891b0984bec8b36e7557ef/src/xmlgen.cpp#L869
        if (defArg && !defArg->name.isEmpty() && defArg->name!=a.name)
        {
          t << "          <defname>";
          writeXMLString(t,defArg->name);
          t << "</defname>" << endl;
        }

...it would appear that the intent is to include the <defname> if it's
different - but I haven't been able to figure out why this isn't happening.

Any insight would be appreciated. Thanks!

- Paul

PS - Apologies if this ends up cross-posted on the mailing list and
discussion forum. I first tried posting to the forum, but the post did not
seem to go through... but it's possible it's just very delayed!

<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=icon>
Virus-free.
www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=link>
<#DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

[Doxygen-users] help

[Avina, M. <Mig...@AV...>]

Miguel Aviña Pantoja

mig...@av...<mailto:mig...@av...>

Embedded Software Engineer

GCET - Proof of Concept Team



[http://digital.avnet.com/signature/images/Avnet_logo_tagline_rgb.png]



Av. Camino al Iteso 8900 Edif. 1b

45080 San Pedro Tlaquepaque, Jal

O +52.33.8355158

M +52.33.16991715

avnet.com<http://www.avnet.com/>

Re: [Doxygen-users] member name appearing in Modules/Groups

[Mark <dox...@er...>]

> On Mar 20, 2020, at 17:17, Mark <dox...@er...> wrote:
> 
> That is for a function “classA_testcase” which has \memberof classA in the doxygen comment and is in some group created with \defgroup, if I navigate to the function via the Data Structures or Files entries of the left-hand pane’s tree, the name is displayed as classA_testcase. If I navigate via Modules it is displayed as classA::classA_testcase.

The classA:: prefix only appears in the Functions list at the top of a module’s page. Further down the page in the documentation for each of the functions, there is no classA:: prefix. Is this the expected behavior? Or should the prefix never appear or should it appear everywhere?

Regards

    -Mark

[Doxygen-users] member name appearing in Modules/Groups

[Mark <dox...@er...>]

Hi,

I have a c API which I am documenting with Doxygen. The API has an object oriented design and I name the functions as class_funcname. I have been using `\memberof class` in the documentation for a while. Usually this has no effect on how the function names are displayed in the documentation but recently I’ve noticed that in the context of Modules/Groups I see class::class_funcname.

That is for a function “classA_testcase” which has \memberof classA in the doxygen comment and is in some group created with \defgroup, if I navigate to the function via the Data Structures or Files entries of the left-hand pane’s tree, the name is displayed as classA_testcase. If I navigate via Modules it is displayed as classA::classA_testcase.

I’ve been using \memberof and \defgroup for a long time and have only just noticed this. I have recently updated to Doxygen 1.8.17. I wonder is this something that recently changed. Is it a bug? The intent of \memberof is not clear from the documentation which gives little information beyond it being useful when documenting c apis. Maybe I should just delete all the \memberof commands.

Regards

    -Mark

[Doxygen-users] How to make a doc containing other docs?

[Mark <dox...@er...>]

I have a package that includes a library and a set of tools. I am generating man pages as well as html. Because the tools man pages go in man 1 while the library pages go in man 3 I have to have separate projects, i.e., generate separate document packages for each. The only way to change the man page target is in the doxy configuration file (I use doxy as shorthand). I want to create package documentation which will have some of its own pages and refer to the lib and tools documents. I want the tree/index in the left pane of the package documentation to show the package documentation’s own pages and also include “lib” and “tools” entries. These entries should be the same as the trees of the individual document packages. Clicking on the reveal arrow of e.g. lib should expand it show the same entries as the tree/index of the lib documentation package and clicking on one of the entries should take you to that page.

I have been trying to use tags but I can’t for the life of me figure out how to refer to any of those tags in the package documentation. Furthermore there are no tags for the main pages with the tree/index.

Is it possible to accomplish this in Doxygen?

Regards

    -Mark

Re: [Doxygen-users] multiple use of page label warning with only 1 use

[Mark <dox...@er...>]

> On Mar 17, 2020, at 16:10, Mark <dox...@er...> wrote:
> 
> I am getting the following warning:
> 
>     multiple use of page label 'license', (other occurrence: <license>, line: 1)
> 

I figured it out. The file was input to two projects one of which was exporting tags read by the other. Removing it from the input to the second fixed the issue.

Regards

    -Mark

[Doxygen-users] multiple use of page label warning with only 1 use

[Mark <dox...@er...>]

I am getting the following warning:

    multiple use of page label 'license', (other occurrence: <license>, line: 1)

in a Markdown file with the following heading:

License and Attribution Notices           {#license}
=======================

-----------------

The warning appears on the last line shown - the dashed line.

No other files contain '@page license’ and no other .md files use the label ‘license’ so this is looking like a bug to me. Any else seen this or have any idea what is going on or any idea how to find out what file <license> refers to?

Regards

    -Mark

[Doxygen-users] Install Doxygen without CMake

[Stefan H. <de...@st...>]

Hey guys,

I'm on a shared webserver where CMake isn't available. Is there a way to install Doxygen without CMake?
Thanks and best regards
Stefan

[Doxygen-users] "List of all members" appears twice

[Olivier C. <Oli...@ce...>]

Dear All,

In my doc, at the top right of each page, the “List of all members” link
appears twice. Do you know how that’s possible ?

See:
https://root.cern/doc/master/classTGraph.html

Cheers,
O.Couet

Re: [Doxygen-users] Copybrief fails when used in markdown table (v1.8.17)

[DonM <dm4...@in...>]

Sorry, I didn't mean to take this discussion off-list.

Thanks, I get it now. The fact that both error messages have that "</td>" tag does make you wonder if the two issues are related. But I agree it's worth reporting on the bug tracker.


On 1/27/2020 11:13 PM, Dirk Dickmanns wrote:
> |Adding "&nbsp;" or whitespace and some additional text correctly removed the appended ||"|</td>" from the target name in the error message, but still the (now correct) \copybrief target is not found.||
> ||
> ||
> ||You are right, its not related to cross-references and yes, my problem started in 1.8.16.||
> ||
> ||
> ||Sounds like I should report it as bug...||
> ||
> ||
> ||Thanks,||
> ||
> ||
> ||Dirk
> ||
>
> On 1/27/20 3:24 PM, DonM wrote:
>> Not sure what you mean. Both this thread and the thread you started (\copybrief for \file not working after 1.8.15?) discuss the failure of copybrief to copy the brief text. It has nothing to do with cross-references.
>>
>> Adding `"&nbsp;"` does correct the problem of copybrief not copying text in a markdown table. I did not test it in a file description.
>>
>> I believe the two threads report different problems. Failure in a markdown table occurs starting with version 1.8.17. It works correctly in 1.8.16. Your problem of copybrief in a file description fails in 1.8.16.
>>
>> On 1/26/2020 10:31 PM, Dirk Dickmanns wrote:
>>>
>>> Thanks for the hint, DonM, I see the "|</td>" at the end of the refs as well.|
>>>
>>> |Adding "&nbsp;" or whitespace and some other text corrected this, but to no avail, the refs remain broken.|
>>>
>>> |It happens both inside and outside of tables.
>>> |
>>>
>>> Dirk
>>>

Re: [Doxygen-users] Copybrief fails when used in markdown table (v1.8.17)

[Dirk D. <dir...@on...>]

Thanks for the hint, DonM, I see the "|</td>" at the end of the refs as 
well.|

|Adding "&nbsp;" or whitespace and some other text corrected this, but 
to no avail, the refs remain broken.|

|It happens both inside and outside of tables.
|

Dirk

[Doxygen-users] Copybrief fails when used in markdown table (v1.8.17)

[DonM <dm4...@in...>]

In doxygen version 1.8.17, the copybrief command fails when used inside a markdown table.
This error does not occur with version 1.8.16.
I'm running on Windows 10.

For example, if your markdown file contains:

|Function | Use ---------------------|--------------------------- ::TestFunction2() | \copybrief TestFunction2() |

Doxygen reports:
|warning: @copybrief or @copydoc target 'TestFunction2()</td>' not found.|

Notice the |</td>| that has been incorrectly appended to the function name in that message.

This error does not occur if the referenced function has no parameters. You can work around the problem by appending |&nbsp;| after the copybrief reference. Another work around is to explicitly specify the parameters such as:

|\copybrief TestFunction2(int)|

I have reported this as issue #7452 <https://github.com/doxygen/doxygen/issues/7542> with a small example case.

DonM

[Doxygen-users] \copybrief for \file not working after 1.8.15?

[Dirk D. <dir...@on...>]

Hello,

I am not sure whether I completely misunderstood \brief for files...

Up to 1.8.15, I could use \copybrief for files like in the following 
minimal example test.h:

---

/** \file

     \brief File brief test

     blabla.

     \copybrief test.h */

---

doxygen -g Doxyfile

doxygen Doxyfile

produced something like the following as html:

---

<My Project header>

test.h File Reference

File brief test. More...

Go to the source code of this file.


    Detailed Description

File brief test.

blabla.

File brief test.

---

Later versions give a warning:

.../test.h:4: warning: @copybrief or @copydoc target 'test.h' not found

and omit the last line (the second "File brief test.") in the output.

Something wrong with my example?  Or a bug?  It is on a 5.4.14-arch1-1 
x86_64 GNU/Linux.

Kind regards,

Dirk

[Doxygen-users] TYPDEF_HIDES_STRUCT does no longer work for enums in IDL files in Doxygen 1.8.17

[Andreas T. <an...@st...>]

Hello World,

As the subject says: TYPDEF_HIDES_STRUCT does no longer work for enums 
in IDL files in Doxygen 1.8.17. It did work in 1.8.16.

Attached is a small example file which gives a warning ("Documentation 
for undefined enum 'digitsEnumeration' found") when the doxygen ist 
running with TYPDEF_HIDES_STRUCT = YES

Best regards
	Andreas
-- 
       ("`-''-/").___..--''"`-._
        `o_ o  )   `-.  (     ).`-.__.`)
        (_Y_.)'  ._   )  `._ `. ``-..-'
      _..`--'_..-_/  /--'_.' .'
     (il).-''  (li).'  ((!.-'

Andreas Tscharner   an...@st...   ICQ-No. 14356454

[Doxygen-users] possible to create a table of contents with outline-style numbering?

[Phillip F. <phi...@gm...>]

I would like to create a document that has a table of contents like the
following:

1. module xyz
   1.1 function a
   1.2 function b

2. module www
   2.1 function c
   2.2 function d

Is it possible to auto-generate a document with a table of contents in the
above style using Sphinx?

Phillip