libpkg: rework internal ABI handling, fix bugs

[ifreund]

First of all, apologies for the scope creep on this one. I didn't find a good way to break it up into smaller patches. Writing this was a bit like pulling a thread that caused the whole knit sweater to unravel.

This is in my opinion a much needed refactor that will make future work on further architecture-specific code such as the shlib tracking features significantly easier.

The main guiding principle of this patch is to create a single source of truth for as many ABI related details as possible. I'll summarize the changes with a series of before/after comparisons:

Before: Setting -o ABI=... on the command line but not -o OSVERSION=... led to inconsistency between ABI and OSVERSION.

After: Setting ABI requires setting OSVERSION as well (and vice versa). (Note: OSVERSION is FreeBSD-specific)
Related: #2363

Before: ALTABI set in the environment, with -o, or in pkg.conf only affected the ALTABI variable expanded while parsing repo conf files and nothing else.

After: Setting ALTABI is completely ignored and a notice is printed if it is set. The ALTABI formatted string derived from the automatically detected or configured ABI is still printed by pkg config ALTABI.

Before: Setting ABI/OSVERSION in pkg.conf would not affect the variable expansions of ABI, OSVERSION, ARCH, VERSION_MAJOR, etc. in pkg.conf. This was a pretty big footgun IMO, though I don't know of any examples of people getting bitten by this. I don't think setting ABI in pkg.conf is a normal thing to do.

After: Setting ABI, ALTABI, and OSVERSION in pkg.conf is not allowed. This is consistent with ABI_FILE, which was never able to be set using pkg.conf. It is still possible to set ABI and OSVERSION on the command line and e.g. pkg config ABI still works. This is the simplest solution I see to this problem, but it is a breaking change.

Before: Parsing of user-provided ABI strings was very lax. Pretty much anything was accepted without giving any error as long is it contained at least one colon. The full affects of providing such invalid ABI strings to pkg are not well understood, I think it can safely be assumed that interesting bugs would result.

After: Parsing of user-provided ABI strings is very strict. Only a well defined set of supported OS names and architectures are accepted. Missing or extra components cause a hard error.

Before: The global ctx.oi os_info struct stored multiple copies of the same information that could get out of sync with each other. For example, the version was stored in 4 or 5 different places in the same struct.

After: The global ctx.abi struct contains no redundant information.

Before: Most code dealing with ABI information worked with stringly typed data. Conversion to/from strings and parsing details were strewn across the codebase.

After: Most code dealing with ABI information uses the new struct pkg_abi representation with enums for OS and architecture and integers for the version. All parsing is unified in pkg_abi.c and multiple minor inconsistencies between previously duplicated code are resolved.

[ifreund]

@kevemueller I did end up leaving the -x32 architectures out of the pkg_arch enum for now. In my opinion we should have tests for parsing all the architectures in that enum from at least one kind binary of on at least one OS.

It would be trivial to add them or any other currently unsupported architecture to that enum and the handful of functions converting to/from strings in pkg_abi.c if/when pkg gains full support for new architectures.

[ifreund]

Also, this will conflict with #2372, I'm happy to do the rebase if you want to review/merge that one first @bapt. No rush though of course :)

[ifreund]

After: Setting ABI, ALTABI, and OSVERSION in pkg.conf is not allowed. This is consistent with ABI_FILE, which was never able to be set using pkg.conf. It is still possible to set ABI and OSVERSION on the command line and e.g. pkg config ABI still works. This is the simplest solution I see to this problem, but it is a breaking change.

I realized that we do have an alternative here that isn't terribly complex to implement. We could parse pkg.conf twice:

1. First parse only the ABI and OSVERSION entries and don't provide any variable expansions.
2. Parse the same file again, this time providing the ABI, ALTABI, OSVERSION, OSNAME, etc. variable expansions base on the values of ABI and OSVERSION or ABI_FILE from the first parse or the environment.

Since we would still need to error out if ABI is set but not OSVERSION I'm not sure if this would make this PR as a whole any less breaking in practice. I'd be happy to implement this alternative if it would be preferable to continue to be able to set ABI/OSVERSION from pkg.conf.

[ifreund]

This change here nicely highlights my motivation for working on this patch. Having a standard, non-textual representation of the ABI lets us get rid of the entire shlib_vaild_abi() function and use the same elf_parse_arch() function we use to automatically detect the ABI from a file.

The best part is that this is a bug fix as well as a simplification, shlib_valid_abi() currently tries to parse the ABI string as if it had the ALTABI format and ends always returning true in practice regardless of the input.

This change will make implementing #2331 far easier.

[bapt]

rebased and merged manually