Subject: Re: [boost] [BGL] StoerâWagner min-cut algorithm
From: Jeremiah Willcock (jewillco_at_[hidden])
Date: 2010-07-19 23:58:20
On Fri, 16 Jul 2010, Daniel Trebbien wrote:
> I uploaded a new archive, version 0.5.3, to:
> The main differences are:
> 1. The DOT files have been edited to make use of the `pos` attribute
> and other, layout-specific attributes.
> 2. The `make_min_cut_images.sh` script has been updated. A new Windows
> version, `make_min_cut_images.bat`, has been added.
I just need some kind of script for my records. I'll change the shell
script to output PNG files, though, and re-run it (or you can).
> 3. `graph_theory_review.html` and `stoer_wagner_min_cut.html` have
> been updated ever-so-slightly as a result of the graph images changing
> 4. I have (hopefully) fixed the issues with line endings. All line
> endings of text files in the archive should be Windows-style except
> for the `make_min_cut_images.sh` script, which is Unix-style.
I'll have to convert them on extracting the zip file so I can mark them as
text (with native line endings on checkout) in SVN. Having everything
Unix-style would be easier for me, but it's just one flag to do the
conversion automatically on all text files in the zip archive.
BTW, I added a read_dimacs_min_cut() function to read_dimacs.hpp; see if
that works for your test cases. I haven't tested it, but it should use
the same code as read_dimacs_max_flow() but with the source and sink
features turned off (and so removing the "n" line requirements).
I've been kind of busy recently, so I'll look through your files in more
detail soon. You might want to update your test code and upload a new
version, though, to use the new reading function.
-- Jeremiah Willcock
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk