dynagraph2-devel — Dynagraph developers' list

Re: [dynagraph2-devel] graph resolution

[Gordon W. <go...@wo...>]

Hi Erik -

> Thanks again for writing Dynagraph, Gordon.  I like incrface; every
> bug I've found so far turned out to be an oversight on my part.

I'm glad you're enjoying Dynagraph.  Stephen North designed the  
incrface language; as you can see, it's a variant of the dot language  
graphviz uses.  In fact I believe the languages can be unified in the  
future.

> Edge pos precision can be increased!  It's controlled by the graph's
> resolution.

Oh, for some reason before I thought you were talking about the  
floating-point precision.  There is always some confusion about what  
these attributes mean, and I will probably want to rename them to make  
them clearer in later versions.

Resolution controls the minimum increment used for x and y  
coordinates.  The floating point coordinates are divided by these  
values in order to produce integers which are used internally, and  
then resolution is multiplied back in on output.

Unfortunately, the resolution attribute has to do with bitmap  
resolution in graphviz, which results in some incompatibilities.   
Really this attribute should be called minimal-increment or something.

> Errors such as "(exception) Assertion: cx>=px; .\Check.cpp, 107" only
> crop up when the resolution value is set too high.

Yes, invalid input should generate useful errors rather than hitting  
assertions deep in the sanity-check code.

> I could use a few hints regarding compiling graphviz and dynagraph for
> windows, although the binary will probably suffice for win32.  The
> unices have been straightforward.

Oh, has http://www.dynagraph.org/building-dgwin.php slipped out of  
date?  Please post any problems you're having.  Really dynagraph  
should have its own build process on windows apart from Dynagraph for  
Windows; the reason I still build dynagraph as part of dgwin is the  
dependencies on graphviz, which are currently handled by injecting  
visual studio projects into the graphviz source directory.  This could  
probably be cleaner, but on the other hand, the graphviz dependencies  
will also go away in due time.

Thanks for the feedback,
Gordon

[dynagraph2-devel] graph resolution

[Erik H. <ice...@gm...>]

Thanks again for writing Dynagraph, Gordon.  I like incrface; every  
bug I've found so far turned out to be an oversight on my part.  For  
example:

Edge pos precision can be increased!  It's controlled by the graph's  
resolution.
Errors such as "(exception) Assertion: cx>=px; .\Check.cpp, 107" only  
crop up when the resolution value is set too high.

I could use a few hints regarding compiling graphviz and dynagraph for  
windows, although the binary will probably suffice for win32.  The  
unices have been straightforward.

Cheers,
Erik

Re: [dynagraph2-devel] 1.1.6 branch: dgassert.h missing in SVN?

[Gordon W. <go...@wo...>]

Thanks Erik,

The graphsearch directory is extinct and I just removed it.  This was  
an old experiment with large graph exploration.  Dynasty (www.dyna.org/Dynasty) 
  does large graph exploration much better and should be released  
later this year.

Gordon

On Jul 9, 2008, at 9:53 PM, Erik Hvatum wrote:

> This might be a legit missing file; hopefully it contains the
> NodeNotFound definition.
>
> -Erik
>
> -------------------------------------------------------------------------
> Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW!
> Studies have shown that voting for your favorite open source project,
> along with a healthy diet, reduces your potential for chronic lameness
> and boredom. Vote Now at http://www.sourceforge.net/community/cca08
> _______________________________________________
> dynagraph2-devel mailing list
> dyn...@li...
> https://lists.sourceforge.net/lists/listinfo/dynagraph2-devel

[dynagraph2-devel] 1.1.6 branch: dgassert.h missing in SVN?

[Erik H. <ice...@gm...>]

This might be a legit missing file; hopefully it contains the  
NodeNotFound definition.

-Erik

Re: [dynagraph2-devel] 1.1.6 branch

[Gordon W. <go...@wo...>]

> I've started my latest windows efforts with this branch & ran into a  
> couple missing files:
> * dynagraph.8.00.sln (easy to recreate, although doing so loses any  
> project vs project dependancy info that may be included in your copy)

Yeah, I use the solution file from dgwin - nothing particularly  
interesting about the dependencies but it would probably be a good  
idea to include an sln in dynagraph.  The dgwin source also includes  
makefiles for the graphviz libraries used

> * dynadag/constraintgraph.cpp

Hum, you shouldn't need this, it was templatized and changed to a pure  
header in 1.1.6

> I'll try harvesting that cpp file from the 1.1.7 or the 1.1.5  
> branch, and likewise for any others that I run into.  Please advise  
> if I really ought to stay on trunk or wait for commits.

The way I am managing this, trunk always contains a super-stable  
version, and I work in the branches.  So 1.1.6 is the best place to  
work, yes, thanks.

Gordon

[dynagraph2-devel] 1.6.1 branch

[Erik <ice...@gm...>]

The 1.1.6 branch in your repository contains the most recent  
commits.  It has VS8 vcproj files, which are convenient as I  
(guiltily) do most of my development on Windows with occasional  
testing on OS X and linux.

I've started my latest windows efforts with this branch & ran into a  
couple missing files:
* dynagraph.8.00.sln (easy to recreate, although doing so loses any  
project vs project dependancy info that may be included in your copy)
* dynadag/constraintgraph.cpp

I'll try harvesting that cpp file from the 1.1.7 or the 1.1.5 branch,  
and likewise for any others that I run into.  Please advise if I  
really ought to stay on trunk or wait for commits.

Thanks,
Erik

[dynagraph2-devel] interested in patches?

[Erik <ice...@gm...>]

Hello Gordon et all,

If there's interest, I'd be happy to submit patches to:

*clean up makefile include and lib path variables so that the  
settings in Config.mk stick and are not overridden by each makefile
*print edge spline control point coordinates with more precision

In the mean time, I'll start version tracking my modified Dynagraph  
in RectaGraphTrace's sourceforge SVN repository.

-Erik

[dynagraph2-devel] hello; new releases

[Gordon W. <go...@wo...>]

Hello to those who have joined these new mailing lists!

I'm announcing two releases from yesterday:
1.1.5.9 is a stable release with improved build systems.  Using gmake, 
it automatically generates dependencies for quicker building on *nix.
1.1.6.0 is an extremely experimental version which delivers parallel 
edges (a.k.a. multiedges).  However, the drawing is not perfect: if the 
edges are short they will tend to pile up.  And this version is less 
stable (some of the more complicated tests crash).  I am presently 
working on making it more stable and then I will work on the drawing itself.

In the future, I will make release announcements to the users and 
announce mailing lists only, assuming that techies will subscribe to 
more than one list.  Not every release merits an announcement.  The 
users list should be used for questions of general interest, and the 
devel list for more in-depth discussion of ways to improve Dynagraph.

Note on the development process: All Dynagraph development occurs in 
subversion branches; the trunk always contains some old, stable version 
(currently 1.1.5.9).  It may be possible to get newer code directly from 
the repository, but it is not guaranteed to compile on all platforms, as 
the releases should.  For example, the 1.1.7 branch contains 
reconfiguration code that will allow switching layout engines on the 
fly, but it does not yet compile or work on any platform!

Cheers,
Gordon