Tamás Nepusz

I think the original intention for status handlers was to use them in place of progress handlers in cases when it is hard to add a percentage to the operation...

The original idea behind status handlers was something similar to the `tkigraph` use-case, i.e. let GUIs wrapping igraph be notified about what igraph is doing at the moment if no...

Thank you for your interest in helping us develop igraph further! Here's a short guide for contributors that should help you get started: https://github.com/igraph/igraph/wiki/Quickstart-for-new-contributors I cannot emphasize enough that we...

I'd treat PRPACK and Cliquer differently; these are vendored pieces of code and it is easier to keep track of the changes we make to it (if needed) if we...

They are not unused. They replace `static igraph_error_t INTERNAL_FUNCTION(init_item)` and friends from `typed_list.pmt` because `CUSTOM_INIT_DESTROY` is defined before including `typed_list.pmt`.

You probably need `-lm` besides `-ligraph` when linking manually with `gcc`.

I disagree; I don't plan to maintain `igraph-scg` unless there is a serious bug that is easy to fix and does not require a deep understanding of spectral coarse graining....

> So it will be stripped from R/igraph as well? Yes, that's the plan; it will have to be deprecated in R-igraph first and then I'd go for removal. If...

`igraph-scg` uses `igraph_weighted_sparsemat()`, which seems redundant now that we have `igraph_sparse_weighted_adjacency()`. This also revealed that the C core now has `igraph_sparsemat()`, which is also redundant if we have `igraph_sparse_adjacency()`. Shall...

Another roadblock: `igraph_scg_adjacency()` and similar functions (`_stochastic()` and `_lapacian()`) take one of these three things as inputs: * an `igraph_t` graph object, OR * a dense `igraph_matrix_t` OR * a...