zhou-run

> The problem is that we are building a temporary link-list `adj_list` containing the up adjacencies. The temporary list is deleted and their nodes are freed but still referenced in...

> `lsp_regenerate_schedule()` delays the regeneration of the LSP so that it is done once the `u.bc.adjdb[X]` adjacency list is fully destroyed. `isis_reset_attach_bit()` should be called during the LSP generation process,...

> > > `lsp_regenerate_schedule()` delays the regeneration of the LSP so that it is done once the `u.bc.adjdb[X]` adjacency list is fully destroyed. `isis_reset_attach_bit()` should be called during the LSP...

> Please merge both fixes Done.

> Please include a backtrace in the commit message always. OK. I will do it this way next time.

> How do you test the code before pushing if it's even failing to compile? Sorry, I have now corrected the compilation error.

> Possible to verify and test this behavior with topotest? Is topotest necessary? Currently, I have no idea, and it seems that topotest is not available for multi-area. Is multi-area...

The topotest for OSPF often fails. Should we consider extending the convergence time for testing? @ton31337

> Seems that tests should be updated/fixed too. Yes, I am currently investigating why the hostname entry is being lost in the case of the `area->dynhostname` being enabled.

> Squash to a single commit. Done.