Builidng deal.ii with Visual Studio: current situation

249 views
Skip to first unread message

volkovych

unread,
Sep 29, 2015, 10:08:57 AM9/29/15
to deal.II User Group
Hi everyone,

I'm very impressed with abilities and documentation of deal.ii. Thanks you all for this!
For some reasons I need to use it under Visual Studio. I see that some efforts in this direction were done in the past but right now it is not straightforward for me if anyone succeeded in building deal.ii with VS.

Can you please update me on this? Then I will know how to proceed.

Thanks!

Regards,
Sergiy.

Krzysztof Bzowski

unread,
Sep 30, 2015, 8:28:56 AM9/30/15
to dea...@googlegroups.com
Check this post
http://comments.gmane.org/gmane.comp.science.mathematics.dealii/6636

As far as I know, VS studio is still unsupported. Best alternative for
now is to use VisualGDB as a tunnel to Linux host or virtual machine.

Best regards,
Krzysztof

volkovych

unread,
Sep 30, 2015, 8:56:21 AM9/30/15
to deal.II User Group
Yes, I read this post as well as others devoted to this topic.
I think the most relevant is this one:
https://groups.google.com/forum/#!searchin/dealii-developers/visual$20studio/dealii-developers/AxIqVM5Yw8A/sJuV4ig3FEUJ

But finally it is still not clear what was the output of those efforts.
Does anyone succeeded in this?

Thanks.

Bruno Turcksin

unread,
Sep 30, 2015, 10:22:06 AM9/30/15
to deal.II User Group
Sergiy,

I think that if you use the development version of deal.II, you should be able to compile deal.II with VS 13 and some tutorials should work. I think that windows support is still a work-in-progress.

Best,

Bruno

volkovych .

unread,
Sep 30, 2015, 10:42:05 AM9/30/15
to dea...@googlegroups.com
Thanks, I will try. I'm particularly interested in modelling elasticity problem in 3D. Will this one work? E.g. step 8 of tutorial.


--
The deal.II project is located at http://www.dealii.org/
For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en
---
You received this message because you are subscribed to a topic in the Google Groups "deal.II User Group" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/dealii/st2aE85mU20/unsubscribe.
To unsubscribe from this group and all its topics, send an email to dealii+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.



--
З повагою та найкращими побажаннями,
Сергій Волкович

Bruno Turcksin

unread,
Sep 30, 2015, 10:49:30 AM9/30/15
to dea...@googlegroups.com

On 09/30/2015 09:42 AM, volkovych . wrote:
Thanks, I will try. I'm particularly interested in modelling elasticity problem in 3D. Will this one work? E.g. step 8 of tutorial.

I don't know which steps work and which don't but I would think that step 8 works. I advise you to turn off all the third-party libraries. Also you may have a problem with the latest version of cmake on windows so you may need to use an older version.

Wolfgang Bangerth

unread,
Oct 13, 2015, 9:12:21 AM10/13/15
to dea...@googlegroups.com
On 09/30/2015 09:22 AM, Bruno Turcksin wrote:
>
> I think that if you use the development version of deal.II, you should be able
> to compile deal.II with VS 13 and some tutorials should work. I think that
> windows support is still a work-in-progress.

Though we should be close. We would certainly be very interested in hearing
any feedback and suggestions to get us the rest of the way!

Best
W.

--
------------------------------------------------------------------------
Wolfgang Bangerth email: bang...@math.tamu.edu
www: http://www.math.tamu.edu/~bangerth/

Timo Heister

unread,
Oct 13, 2015, 10:50:29 AM10/13/15
to dea...@googlegroups.com
I don't know if you have seen, but I managed to submit test results
from windows: http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13350

That branch still has a few small changes compared to master, but with
them deal.II builds correctly. Some of the scripts for tests have to
be replaced (that is why they are currently failing).
> --
> The deal.II project is located at http://www.dealii.org/
> For mailing list/forum options, see
> https://groups.google.com/d/forum/dealii?hl=en
> --- You received this message because you are subscribed to the Google
> Groups "deal.II User Group" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to dealii+un...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.



--
Timo Heister
http://www.math.clemson.edu/~heister/

Krzysztof Bzowski

unread,
Oct 25, 2015, 8:08:48 AM10/25/15
to dea...@googlegroups.com
Hi,
I cannot find this branch on github. Is it located here:
https://bitbucket.org/tjhei/dealii/src/
?

I was trying to build this branch on Windows using Visual Studio 2015.
Unfortunately it still produces some errors.

http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13852

Best regards,
Krzysztof

Wolfgang Bangerth

unread,
Oct 25, 2015, 1:10:20 PM10/25/15
to dea...@googlegroups.com

Krzysztof,

> I was trying to build this branch on Windows using Visual Studio 2015.
> Unfortunately it still produces some errors.
>
> http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13852

I'm trying to match these error messages to the current source base, but the
lines don't seem to match up any more. What revision did you try to test here?

Krzysztof Bzowski

unread,
Oct 25, 2015, 2:25:41 PM10/25/15
to dea...@googlegroups.com
On 2015-10-25 18:10, Wolfgang Bangerth wrote:
> I'm trying to match these error messages to the current source base, but
> the lines don't seem to match up any more. What revision did you try to
> test here?
>

As I mentioned I took it from: https://bitbucket.org/tjhei/dealii/src/
I thought this is a branch suitable to build with Visual Studio.
Am I correct?

Best,
Krzysztof

Wolfgang Bangerth

unread,
Oct 25, 2015, 2:35:19 PM10/25/15
to dea...@googlegroups.com
On 10/25/2015 01:25 PM, Krzysztof Bzowski wrote:
>
> As I mentioned I took it from: https://bitbucket.org/tjhei/dealii/src/
> I thought this is a branch suitable to build with Visual Studio.
> Am I correct?

I think Timo simply put it there because it was easier to set up some testing
script. I don't think there were any actual changes on there, but we've made
some changes in the meantime on mainline. Can you try again with the version
from github.com/dealii/dealii and post the output like you did for the version
above?

Thanks

Krzysztof Bzowski

unread,
Oct 25, 2015, 2:57:07 PM10/25/15
to dea...@googlegroups.com

> I think Timo simply put it there because it was easier to set up some
> testing script. I don't think there were any actual changes on there,
> but we've made some changes in the meantime on mainline. Can you try
> again with the version from github.com/dealii/dealii and post the output
> like you did for the version above?
>
> Thanks
> W.
>

I tried current master branch from github.
http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13870

Best,
K.

Timo Heister

unread,
Oct 25, 2015, 7:56:40 PM10/25/15
to dea...@googlegroups.com
> I think Timo simply put it there because it was easier to set up some
> testing script. I don't think there were any actual changes on there

I did, but all code changes are already on mainline. There are still
some changes I made trying to get the unit tests to run, but this is
work in progress.

Speaking of mainline, it looks like some other stuff broke between then and now.

Wolfgang Bangerth

unread,
Oct 25, 2015, 8:22:50 PM10/25/15
to dea...@googlegroups.com

> I tried current master branch from github.
> http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13870

That's a few errors and warnings less, but ultimately it is still baffling how
amazingly poor a C++ compiler Microsoft manages to put out :-(

There is clearly nothing wrong with the code itself, to the best of my
knowledge. The best we can do is trying to work around whatever MSVC happens
to be unhappy about, but that's difficult to do unless one has access to the
compiler. I don't understand how the first error shown happens -- if you
compile on the command line (or from the Visual Studio environment), is this
the first error line you get?

Best

Timo Heister

unread,
Oct 25, 2015, 9:31:04 PM10/25/15
to dea...@googlegroups.com
> There is clearly nothing wrong with the code itself, to the best of my
> knowledge.

the library is now compiling again for me with #1789:
http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13878
also see https://ci.appveyor.com/project/tjhei/dealii-9v5eu/build/1.0.39-appveyor

One problem (not visible above because I haven't gotten the
tests/steps to show up) is that step-1 is no longer working. I opened
an issue: https://github.com/dealii/dealii/issues/1790

Krzysztof Bzowski

unread,
Oct 26, 2015, 3:19:36 PM10/26/15
to dea...@googlegroups.com
On 2015-10-26 02:30, Timo Heister wrote:
>> There is clearly nothing wrong with the code itself, to the best of my
>> knowledge.
>
> the library is now compiling again for me with #1789:
> http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13878
> also see https://ci.appveyor.com/project/tjhei/dealii-9v5eu/build/1.0.39-appveyor
>

Any idea why I cannot build it using VS 2015?
http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13899

Best,
K.

Wolfgang Bangerth

unread,
Oct 26, 2015, 3:40:18 PM10/26/15
to dea...@googlegroups.com
> Any idea why I cannot build it using VS 2015?
> http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13899

Good question -- I think it is more compiler bugs :-(

I think that the first error you see (about the problems in tria.h) should be
fixed by this patch
https://github.com/dealii/dealii/pull/1791
which has just been merged into the current github sources.

I do not understand the issues with SparseMatrix. Like I said, it would help
to see what the *first error* is you get from this file. I suspect that there
is more than is logged on the website above.

Krzysztof Bzowski

unread,
Oct 26, 2015, 4:28:24 PM10/26/15
to dea...@googlegroups.com
On 2015-10-26 20:40, Wolfgang Bangerth wrote:
>> Any idea why I cannot build it using VS 2015?
>> http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13899
>
> Good question -- I think it is more compiler bugs :-(
>
> I think that the first error you see (about the problems in tria.h)
> should be fixed by this patch
> https://github.com/dealii/dealii/pull/1791
> which has just been merged into the current github sources.
>

Latest revision:
http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13901

Supporting VS will be a huge pain if newer versions of VS cause more
problems than older ones.

Best,
K.

Wolfgang Bangerth

unread,
Oct 26, 2015, 6:55:27 PM10/26/15
to dea...@googlegroups.com
On 10/26/2015 03:28 PM, Krzysztof Bzowski wrote:
>>
>>
>
> Latest revision:
> http://cdash.kyomu.43-1.org/buildSummary.php?buildid=13901
>
> Supporting VS will be a huge pain if newer versions of VS cause more problems
> than older ones.

Yes. That is correct. It has been my frustration with the MS compilers for a
decade already.

The logs you show some other error where we build the expand_instantiations
script with the wrong flags. I've opened an issue for this at github.
Reply all
Reply to author
Forward
0 new messages