Revise the documentation of kruskal()

[Peter Kovacs]

The attached patch file contains doc improvements for kruskal(), but there are two more things that should be revised.

1. If the input graph is not connected (or the input digraph is not weakly connected), a spanning forest is calculated instead of a spanning tree. Am I right? I think, there should be a \note in the doc about this important case.
2. The \warning seems to be meaningless. Perhaps a word is missing after "If Kruskal runs on an".

[Alpar Juttner]

Replying to kpeter:

The attached patch file contains doc improvements for kruskal(), but there are two more things that should be revised.

1. If the input graph is not connected (or the input digraph is not weakly connected), a spanning forest is calculated instead of a spanning tree. Am I right? I think, there should be a \note in the doc about this important case.

Yes, there should be.

2. The \warning seems to be meaningless. Perhaps a word is missing after "If Kruskal runs on an".

Or more words. Or even the whole warning should be rephrased, because reading this warning, I'm pretty much confused what I'm warned about. For example, what does "Arc type" mean.

I was digging in the svn history and found that this comment is due to deba, -r3261. At that time "Arc type" was "Edge type" which is a bit better, but still confusing. Before that version, there was another \warning in the code saying:

/// \warning If kruskal runs on an
/// \ref lemon::concepts::UGraph "undirected graph", be sure that the
/// map storing the tree is also undirected
/// (e.g. ListUGraph::UEdgeMap<bool>, otherwise the values of the
/// half of the edges will not be set.
///

It may help fixing this warning.

[Peter Kovacs]

Replying to alpar:

Or even the whole warning should be rephrased, because reading this warning, I'm pretty much confused what I'm warned about.

As I understand this warning should be only about using Edge for Graph types and Arc for Digraph types. However it seems to be clear in the current doc. What about removing thew warning?

[Peter Kovacs]

The attached patch is now in the main branch, but with different hash id: [74cd0c58b348]. Could you tell me why?

[Alpar Juttner]

Replying to kpeter:

The attached patch is now in the main branch, but with different hash id: [74cd0c58b348]. Could you tell me why?

Use 'hg diff' and you will see. (You wanted to change something (well, something unimportant) in your patch, but you made a mistake when you did it. I fix that in [74cd0c58b348]. Of course, I should have put a note here. I'm sorry)

[Alpar Juttner]

As I understand this warning should be only about using Edge for Graph types and Arc for Digraph types.

Probably you mean EdgeMap}} and {{{ArcMap.

However it seems to be clear in the current doc. What about removing thew warning?

OK from my side.

[Peter Kovacs]

Replying to alpar:

Use 'hg diff' and you will see. (You wanted to change something (well, something unimportant) in your patch, but you made a mistake when you did it. I fix that in [74cd0c58b348]. Of course, I should have put a note here. I'm sorry)

I see now. It's okay, thank you. I just missed your usual comment: "[...] went to the main branch".

[Peter Kovacs]

Replying to alpar:

As I understand this warning should be only about using Edge for Graph types and Arc for Digraph types.

Probably you mean EdgeMap and ArcMap.

Yes, or an iteraror of an STL container with Edge/Arc as its value_type.

[Alpar Juttner]

Peter, could you do the missing changes?

Thanks in advance.

[Peter Kovacs]

Replying to alpar:

Peter, could you do the missing changes?

See [6d7bfcf5b48e].

[Alpar Juttner]

Replying to kpeter:

Replying to alpar:

Peter, could you do the missing changes?

See [6d7bfcf5b48e].

Thanks.