Tamás Nepusz
Tamás Nepusz
Indeed, this would be a nice addition to the C core.
This algorithm is a third-party contribution from Viger and Latapy (the authors of the VL method, see [the corresponding publication](https://arxiv.org/pdf/cs/0502085.pdf)). I haven't read it thoroughly, but it seem that the...
Hmmm, indeed. Then it's definitely a bug.
I think that if the POSIX people managed to get away without hierarchy in error codes in `errno.h`, then we can probably manage without a hierarchy as well :) But...
So I've made some stats. First I extracted the list of known igraph error codes from the header: ```sh $ cat include/igraph_error.h | grep "IGRAPH_.*=.*,$" | cut -d '=' -f...
I've added unit tests for `igraph_coreness()`, and one part of the unit test randomly adds loop and multiple edges to the Zachary karate club graph and then validates that the...
This is something I've had in mind for a while. The stack-based approach sounds good; each entry in the stack would have a `weight` value and ` current` value. `IGRAPH_PROGRESS()`...
The stack probably also needs a message associated to each item, and frontends supporting a single progress bar only would show the bottom-most message (or the top-most one?). Frontends supporting...
IMHO the status handler is redundant; we can simply have progress handlers where the status of each task may be "indeterminate", and this effectively replicates the functionality of the status...