On Mon, Nov 13, 2006 at 07:11:43AM +0200, Jari Aalto wrote: > Manoj Srivastava <[EMAIL PROTECTED]> writes:
> IMPROVE QA > All discussion about "policy has problems" is really due to lack of > quality assurance work that prevents non-sh-compliant scripts to > enter into the packaging in the first place. If people are sloppy or > don't know how to comply with standard sh scripts, You know, it is rather hard to comply to a standard you do not have a copy of; POSIX costs money, and no trivial amount of it. We kinda-learn what is in POSIX and what not by "common lore" that one picks up as one goes or by reading other documents such as the SUS or the bash/dash/... manual that more or less partially indicate what feature they document is POSIX and which one is not. Up to now, I've been surviving with SUS as a reference and the bash manual for quick reference for things I know it says (e.g. is the POSIX equality operator = or == ?). Not that I write that many maintainer scripts, but still. In general, though, for all but the most basic scripts, I have abandoned the idea of making /bin/sh scripts and do /bin/bash scripts. I don't have to count anymore how many times I have to escape nested backquotes, I had until very recently with the GFDL debacle a manual I could easily refer to, ... -- Lionel -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
