Tamás Nepusz
Tamás Nepusz
Also, I don't really like the `group_size` member of `igraph_bliss_info_t` as it is prone to memory leaks. Right now it is allocated by BLISS and the caller must call `igraph_free`...
I don't want to make `mpz_t` part of the public API as it may come from GMP or `mini-gmp` and it will lead to confusion if the user tries to...
@szhorvat I think there's nothing left to do here. I have arguments for: - keeping `igraph_get_isomorphisms_vf2()` and `igraph_get_isomorphisms_vf2_callback()` as is (because the mapping might matter to the user) - keeping...
The link is now fixed, thanks!
I would probably replace it with crcmod-plus instead. crcmod-plus provides ABI3 wheels, which are forward-compatible with future Python versions until 4.0. I don’t know whether crcmod-plus provides wheels for Android,...
Let's keep the `graph` parameter. `igraph_layout_square()` is OK.
> Instead of adding an argument to these functions, could it perhaps be handled more gracefully by simply returning whatever results are already found when the function is interrupted? From...
@szhorvat Shall I merge #2824 before the release candidate (due the day after tomorrow) or shall we leave it and then do things "properly" after the release of 1.0?
API changes done before the release of 1.0, implementation can remain for later, removing from 1.0 milestone for the time being.
I think it would make more sense to have the "edge labelled" variants as separate functions. All these functions with an assortment of boolean parameters start to seem like the...