LAPACK Archives

[Lapack] Doxygen and Lapack

Dear Albert,
I updated my own version of Doxygen to 1.7.5. It looks great.
The Doxygen on the LAPACK website will be updated with our next LAPACK version 
that will be released in November.
Attached is a piece of code that is integrating Doxygen.
The HTML output is wonderful but I do see the Parameter description in the 
generated manpages.
Could you please let me know if I am doing something wrong or is this feature 
not supported.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: dgesv.f90
Type: application/octet-stream
Size: 3583 bytes
Desc: not available
Url :
-------------- next part --------------

On Aug 15, 2011, at 5:30 PM, Albert wrote:

Dear Julie,

Glad to hear that you are going to integrate Doxygen more and more. I
think Dimitri van Heesch will also be glad to hear this.
I  never used the man pages possibility so far, I can imagine that
there might be some problems there. I normally only use only the HTML
version. In case you see any "weird" behavior don't hesitate to ask
questions in the Doxygen forum or file a bug report (in case of
Fortran related problem please mention the word Fortran in the header
of the bug. I'm trying to fix some bugs related to Fortran and scan
the bugs for this word in the header of the bug reports).

Glad to hear that you wil make the complex usage more uniform. Is
there a special reason for using the Complex*16 and not the
complex(kind=kind(1.d0)) or Integer, parameter :: dp = kind(1.d0) and
than using complex(kind=dp) ?
In the mean time I filed a bug report for the "Double Complex" and
also submitted a patch for it (bug 656560).

Best Regards,


On Mon, Aug 15, 2011 at 15:35, julie langou <julie@Domain.Removed> wrote:
Dear Albert,
thank you for the info. We will upgrade our Doxygen software.

The LAPACK project is just starting to use Doxygen and we are going very 
slowly towards a full integration.
We are indeed planning to use Doxygen capabilities more and more.
We have some prototypes using Doxygen format to get a proper LAPACK 
My main concern at the moment is the generation of man pages, it does not 
work as expected.

Concerning complex variables, we are going to uniform our double complex 
variables to use only COMPLEX*16.

indeed our SVN repository was down but it should be up and running by now.

Thanks for everything.
Julie Langou

On Aug 15, 2011, at 7:51 AM, Albert wrote:

Dear Team,

I was pleasantly surprised when I saw that the documentation for
Lapack was available in Doxygen format.
When browsing through the documentation I saw that you were using
version 1.7.3 of Doxygen. Yesterday version 1.7.5 was released and I
think there were quite a few improvements from which the Lapack
documentation can benefit as well,  just to mention SVG support for
graphs, bug fixes in respect to Fortran.

I noticed that you use in your code "Double Complex" as well as
"Complex*16", currently Doxygen does not support "Double Complex" as
this is no standard Fortran. I think it would be better to use a
construct like Complex(kind=kind(1.d0)), I'd like to hear your opinion
about this matter, is there a chance that this will be changed. In the
mean time I've made a note about the usage of "Double Complex" and
will file a bug report for it with Doxygen and fix it (hopefully it
will be in Doxygen with the next release).

I also see that Lapack (and Blas) have a lot of useful documentation
that is not in Doxygen style and thus is not shown, are there any
plans to convert this documentation into Doxygen style ?

I tried to access the repository (svn co but I got the error
svn: OPTIONS of '':
could not connect to server (

I also tried to access the link:
LAPACK for Windows

Please see:

But this link was not reachable for me (The server@Domain.Removed
is taking too long to respond.).

Best Regards,

Lapack mailing list

<Prev in Thread] Current Thread [Next in Thread>

For additional information you may use the LAPACK/ScaLAPACK Forum.
Or one of the mailing lists, or