[ISSUE] Build ncurses with "--with-versioned-syms"

[ISSUE] Build ncurses with "--with-versioned-syms"

[ppp7032]

[-- Attachment #1: Type: text/plain, Size: 753 bytes --]

New issue by ppp7032 on void-packages repository

https://github.com/void-linux/void-packages/issues/54241

Description:
There are many binary distributions compiled for Debian that rely on ncurses. One of these is GHC for aarch64. GHC for this ISA is only officially distributed for Debian.

These packages will annoyingly spew `/usr/lib64/libtinfo.so.6: no version information available` multiple times each time they are used. This is especially annoying for command-line apps (such as the aforementioned GHC).

Using an independent oft-described "esoteric distro" on an "emerging platform" [aarch64] is difficult enough as-is. I believe ncurses should be built with this option for greater compatability with binary distributions for other distros.

Re: Build ncurses with "--with-versioned-syms"

[ahesford]

[-- Attachment #1: Type: text/plain, Size: 416 bytes --]

New comment by ahesford on void-packages repository

https://github.com/void-linux/void-packages/issues/54241#issuecomment-2650921356

Comment:
I don't have a particularly strong opinion on symbol version, but
1. We don't consider or advertise Void as an "esoteric distro", and
2. Binary compatibility with packages from other distributions is neither necessary nor sufficient cause to change our packaging options.

Re: Build ncurses with "--with-versioned-syms"

[ppp7032]

[-- Attachment #1: Type: text/plain, Size: 370 bytes --]

New comment by ppp7032 on void-packages repository

https://github.com/void-linux/void-packages/issues/54241#issuecomment-2651060362

Comment:
1. This doesn't actually address anything I said.
2. This policy is foolish if applied to cases where there is no reason to not take the action. Do you believe improving the experience of your users at no cost is a bad policy?

Re: Build ncurses with "--with-versioned-syms"

[ppp7032]

[-- Attachment #1: Type: text/plain, Size: 455 bytes --]

New comment by ppp7032 on void-packages repository

https://github.com/void-linux/void-packages/issues/54241#issuecomment-2651060362

Comment:
1. This doesn't actually address anything I said.
2. This policy is foolish if applied to cases where there is no reason to not take the action. It is equivalent to choosing binary incompatability for the sake of being different. Do you believe improving the experience of your users at no cost is a bad policy?