On 28.01.2019 13:45, Robert Elz wrote: > Date: Mon, 28 Jan 2019 11:12:07 +0100 > From: Kamil Rytarowski <n54%gmx.com@localhost> > Message-ID: <c2086dd7-10c8-0720-13c6-5b8cd53a523c%gmx.com@localhost> > > | In my perception yes. Such calls won't be portable and can stop to be > | functional with tool upgrade/change. > > Using fractional seconds isn't portable at all, if it isn't portable, and > you need portability, you just don't use it. This is the same for all > kinds of extensions - they harm nothing if they're not used. If > they are used, then there is obviously a good reason. If there is > no use, then the extension was a waste of time, and may as well > be deleted - but that's rare, as someone usually has a use for it > to bother making it happen in the first place. > > | I have checked that some countries > | use one or the other separator depending on region (like Canada). > > Yes. I am not sure what the point of that is though. What's the > difference between users in two different areas in Canada, and users > in Engand and France (except the Canadians might be further apart) ? > > | There are also countries that might use different non-ASCII characters, > > Yes, they might, and even if there are none, you can always just create > your own private locale like that. > > | at least the Arabic ones (momayyez). Supporting 2-3+ styles is opening > | Pandora's box; > > No-one is planning that. There are two options - the C locale, or > the locale set in the environment. That's all that will be supported > (unless we someday make a command that works for LC_NUMERIC > the way iconv works for LC_CTYPE). When not using the C locale, > whatever the locale specifies should be supported, but it is not > sleep (or printf, or seq) doing that, it is strtod() (well, seq might do > a bit of it as well for unrelated reasons ... it wants to count the > number of digits in the arg strings that come after the radix char > for example). > > | while this style in computer science has narrow use-case > | (mostly finances). > > Computer Science is not the whole world - more people use NetBSD > than CS staff/students. > > And once again, if you don't want to use other than the C locale > (except perhaps LC_CTYPE) then just don't... Things work like > that, believe it or not! I treat shell programming as a kind of a programming language. As an implementation detail it combines standalone programs and is interpreted in the fly. I never expect a programming language environment or language to treat parsed language accordingly to locales and mistreat radix characters, quotes, brackets, numbers etc. Radix character is only a single specific detail that might differ. (This is however partially violated by a macro language in MS Office suite, but it has a purpose to work on financial types directly) Who uses (what staff) doesn't matter here, I mean whats the final purpose of such software. There are no e.g. calculators for CS people and the rest. By CS I mean usage with computers involved. And I always set all the locale options to my local one. Pushing locales to command line options is in my perception hypercorrectness and resembles to parse e.g. „ ” instead of "" in C/sh/etc programs. > > kre >
Attachment:
signature.asc
Description: OpenPGP digital signature