OK, reworked to an imported target.
Note that we can't use `IMPORTED_LOCATION` as with e.g. libmagic since there
can be multiple libraries to link against, all available via `LUA_LIBRARIES`.
Instead, define an "imported interface" which seems like a better fit, anyway.
I'll probably rework
@dmnks pushed 2 commits.
e2be683d34c5aab1cf17ae3957a59780446a1130 Use CMake Lua module
ac142060b907ddeccaf844f8fc2e4ce3e9e1c603 Fix leftover "lua.h" import from days
long gone
--
View it on GitHub:
Isn't this the wrong way around. As an example, let's say we have a DSO (where
`ei->interp` is non-null) which is not executable `ei->isExec == 0`. This would
cause the following `return !(true && true) => return !(true) => return false`.
My interpretation of this change was that it was
A new API should return better error messages:
> We'll need something that can apply to all our backends, and that something
> needs to have a sane means of returning error messages.
https://github.com/rpm-software-management/rpm/pull/2097#issuecomment-1319690012
Historically parseEVR() has only looked at packaged data, so the validation is
assumed to be done at build-time.
As it's now exposed through a public API it should optimally do a better job of
sanity checking, but the original use-case of version comparison simply has no
notion of errors at