Update `make_lattice()` / `lattice()`

[szhorvat]

To upgrade from the deprecated igraph_lattice() to the new igraph_square_lattice(), the R interface needs to adapt.

Right now make_lattice has a circular parameter that accepts a single logical. To match the capabilities of igraph_square_lattice(), there should be a parameter that accepts either a single logical, or a vector of logicals, one for each dimension.

It might be good to eventually rename circular to periodic, and to rename make_lattice to make_square_lattice to make room for other types of lattice generators (igraph_triangular_lattice, igraph_hexagonal_lattice, which should also be exposed).

Tasks

• Add periodic parameter, feat: remove deprecated functions before 1.0 #1352
• Rename to make_square_lattice()

[szhorvat]

This you can move to upgrade-2 @krlmlr

[krlmlr]

We still need to deal with the R CMD check warning, but we can temporarily remove the "deprecated" declaration too.

[szhorvat]

Can you add -Wno-deprecated to the build just before the submission?

Note that the C core won't compile without deprecation warnings with GCC because of a GCC limitation. If CRAN truly wants zero warnings, the only way is to disable the warning.

Yes, we absolutely must go through all deprecation warnings and check what the consequences of a proper fix are for each one. I did look at this one and it is possible to make this change in the future as a non-breaking feature extension.

[krlmlr]

I’m not sure if CRAN will accept a package with the switch. We can modify our own code to remove the warnings. Szabolcs Horvát ***@***.***> schrieb am Do. 4. Jan. 2024 um 10:49:

…

Can you add -Wno-deprecated to the build just before the submission? Note that the C core won't compile without deprecation warnings with GCC because of a GCC limitation <https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79078>. If CRAN truly wants zero warnings, the only way is to disable the warning. Yes, we absolutely must go through all deprecation warnings and check what the consequences of a proper fix are. I did look at this one and it is possible to make this change in the future as a non-breaking feature extension. — Reply to this email directly, view it on GitHub <#994 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AANJGSY7QGXNJDU6FGGBV3TYMZ3KHAVCNFSM6AAAAABAQ3YYGSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTQNZWHAYDEMRWGI> . You are receiving this because you were mentioned.Message ID: ***@***.***>