On 11/14/18 1:54 PM, Ron Yorston wrote: > Denys Vlasenko wrote: >> I propose to treat both of these types scripts the same way. >> >> I have a feeling this will result in simpler code. >> >> Let me know if you foresee difficulties with this approach. > > I suppose my concern is that it risks losing the distinctive features > of the two types of script. > > My view is that 'scripted applets' should require configuration in the > same way as native applets: > > - it should be possible to enable and disable them individually > - they should be listed (in alphabetical order!) by 'busybox --help' and > 'busybox --list' > - they should be installed by 'busybox --install' > - they should respond to 'busybox --help name' and 'name --help' > > The infrastructure for all of this is already present. > > Custom scripts on the other hand should require no configuration, apart > from just dropping them in the 'embed' directory. If 'embed' is empty > the code to support custom scripts won't be present in the binary. > > I'd prefer to maintain this distinction and use the same 'embedded > scripts' machinery to support deployment of both types of script. > > But if all embedded scripts are to be treated alike I think the > implementation should tend towards that of native applets, even if > that raises the bar for people who want to deploy custom scripts.
What about not supporting "official" scripts at all? If they're exactly the same anyway except for being outright endorsed as a good idea, then just don't endorse them. Custom scripts seems like a useful use case. Scripted applets does not seem like a useful use case. -- Eli Schwartz Arch Linux Bug Wrangler and Trusted User
signature.asc
Description: OpenPGP digital signature
_______________________________________________ busybox mailing list busybox@busybox.net http://lists.busybox.net/mailman/listinfo/busybox