On Sun, 2022-04-03 at 08:41 +0700, Robert Elz via austin-group-l at The Open Group wrote: > Actually there's no reason to forbid them, they simply do > not work. Applications cannot expect them to work. > That's all that needs to be said.
Well than let's not call it forbid, but - as already the case (also in Geoff's proposal) - use a wording like "other than <backslash> or <newline>" That doesn't forbid it, but leaves the "complex" details on *why* this is so, to the interested reader in the rationale. > That's how innovation happens. Well it's not that I'd have anything against innovation (I mean I do love PCRE). But many of those extensions made by implementations in areas where POSIX doesn't define things, cause IMO quite some trouble in practise. Not rarely even when the user just uses what POSIX actually does define. Just take the mess with `locale` in the shell command language. Many subtle differences... basically no chance to ever reconcile them. btw: For those interested: I've made a PDF for just sed, based on the draft, but with colourisations for the additions/removals of Geoff's proposal from: https://austingroupbugs.net/view.php?id=1550#c5761 (my mind got twisted, when I tried to do it via the "replace that on page x line y with..." ^^) I'd share it if anyone who's interested in those issues has an interest... but I have no idea whether I'd break some copyright or so... :-/ Cheers, Chris.