> I'm aware that i'm replying to an obvious troll. You mean undocumented switches are abuse to the system operators. So, stop trolling and fix the documentation or remove undocumented switches. The sooner the better, man/doc support is your skill. Not smearing, work accordingly!
> Just clarifying what's going on here for bystanders who might feel confused. You're not aware yourself, Ingo, as usual for you. Save yourself the postage fare of your volumEnous strives. Fix manual the page of awk(1) or remove long options for BSD and let it persist in ports and GNU utilities. You're thus mixing and mismatching and covering it up in docs. > CVSROOT: /cvs > Module name: src > Changes by: mill...@cvs.openbsd.org 2023/09/20 10:57:12 > > Modified files: > usr.bin/awk : main.c > > Log message: > Support --version option like upstream awk but don't document it. > > Upstream awk has supported --version for a long time but does not > support -V like our awk does. Both options are supported by gawk. > This is perfectly in line with OpenBSD project goals. How do you know what this is, if you're speaking your own opinion. Coordinate somewhat? > Usually, we do not support long options at all because their "YOU" again > very existence violates POSIX and because, if a programs needs > more options than there are letters in the alphabet, that usually > means the program was seriously misdesigned. Overly opinionated, rejected. Long options is part of some reimplementation utilities like openrsync(1) etc, and POSIX is derived after BSD, which is dishonoured in GNU utilities and extended long options are brought with them which are UNDOCUMENTED HERE. WTF IS THIS! > In some cases, some long options that are synonymous with short > options are so widely used that supporting them *for compatibility > purposes only* makes the life easier for some people, for example > for our porting team. NO. Long options are used in scripts for declarative programming where your GNU info(1) documentation is inconveniencing you. > In those cases, supporting them without > cluttering up the documentation is a perfectly sane approach, in NO. Not EVER. Either in the program and documented or foregone. > particular when the option is as useless as -V in the first place. > Note that most OpenBSD programs, for good reasons, do not provide > an option to print any version number in the first place. It's not about versions, do not skew the topic of fixing undocumented switches. > In some rare cases, practical considerations make it seem worthwhile > to make an exception and provide a long option - usually popularized by > GNU in open defiance of POSIX - that does not have a short equivalent. > In such cases, we do usually document the long option. But that's not > the case here. TLDR; > None of these are hard rules, common sense and good judgement is > always needed, but i certainly agree with what Todd did in this case. Yes, you agree, I do not, and many would NOT. Now, this is MY $opinion. Your blind vote is always amazingly clueless at first draft. What are undocumented switches today? Tomorrow? In 25 years of back-porting monsters? > So everybody, please refrain from insulting Todd who is just doing > some good work here, for free, and for everybody's benefit. That's not the point, Ingo. Write me an autobiographic bug report on fixing the discrepancy, NO? 3-line diff or bust.
