semver module: missing MINOR, PATCH should not be ignored

[zzamboni]

semver module: missing MINOR, PATCH should not be ignored

[krader1961]

There is also the question of how blatantly invalid semvers are handled; e.g., "x.y.z". IMHO all such cases should throw an exception by default and the user should have to explicitly use a &nonstrict option if they want to be lenient. The exception being a "v" (or "V") prefix since far too many otherwise valid semantic version strings include that invalid prefix.

[zzamboni]

I think I'm going to do some work on the module adding the official regular expression to check for validity. A &nonstrict option would allow a [Vv] at the beginning, but probably no more.

[krader1961]

I think I'm going to do some work on the module adding the official regular expression....

I think that is a good idea. Note that a leading "v" was valid in SemVer 1.0.0 but is invalid in the 2.0.0 spec. No idea why that change was made but even if it is a good idea for the SemVer spec I think a "v" prefix should be silently ignored if the string is otherwise valid. That is, always preface the official regexp with [vV]?. Obviously that's just my opinion and opinions are like assholes: everyone has one. 😄 It is perfectly reasonable to make the default behavior strictly conforming and require a &nonstrict option if the user wants to be lenient on this point. I've created PR elves/elvish#1348 that incidentally removes the leading "v" from the Elvish semver string.

[zzamboni]

@krader1961 please check out #18.

[zzamboni]

I think #18 addresses this issue, so I'm closing it. @krader1961 feel free to reopen if there's something still missing.