On 10/21/21 6:38 PM, enh wrote: > On Thu, Oct 21, 2021 at 10:56 AM Rob Landley wrote: > > > I was on the posix call today, and they have a bug: > > > > https://austingroupbugs.net/view.php?id=1440 > > > > Where system("-blah") fails because sh is insane legacy weirdness and it > > turns > > out that -c does NOT take an argument. So: > > > > sh -c -i "echo hello" > > > > Works, which that means "sh -c -potato" tries to parse -potato as an > > option, and > > fails. > > > > The suggested fix is to change system() so that instead of {"sh", "-c", > > command, > > NULL} it does {"sh", "-c", "--", command, NULL} instead, but that can't go > > into > > an issue 7 tc because it's a behavior change which means it would go into > > issue > > 8. (Which they're working on now.) > > > > But they can't standardize a new feature without a "commitment to implement > > it" > > by at least one package maintainer, so I thought I'd give you guys a shout. > > Looks trivial to implement in both codebases? Dash and android's shell both > > work > > with --... > > > > The minutes of the meeting aren't posted to the mailing list yet but > > they're on > > https://posix.rhansen.org/p/2021-10-21 for now. This part's at the bottom. > > > > Rob > > it's unclear whether anyone's actually hit this in practice? and even if > > they > had, their portable workaround would be to prefix with "exec "? > > maybe try libc-co...@lists.openwall.com and see if there's any consensus that > this is worth the potential trouble? i'm worried that someone is > _deliberately_ using this to pass extra flags to the shell, which wouldn't > have a workaround if we did make this change :-(
Pinging you and Rich was my attempt at that, but if there's a dedicated list... Rob
