Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-14 Thread Denys Vlasenko
On Fri, Nov 9, 2018 at 7:34 PM Ron Yorston wrote: > Eli Schwartz wrote: > >If you absolutely must provide these scripts, move them to docs/ -- > >including nologin -- which is more descriptive of their initial > >intention. > > applets_sh/README says: > >So far these scripts are not hooked to

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-09 Thread Ron Yorston
Eli Schwartz wrote: >If you absolutely must provide these scripts, move them to docs/ -- >including nologin -- which is more descriptive of their initial >intention. applets_sh/README says: So far these scripts are not hooked to the build system and are not installed by "make install".

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-09 Thread Eli Schwartz
On 11/7/18 2:42 AM, Ron Yorston wrote: > Kang-Che Sung wrote: >> Let no script applets individually configurable in menuconfig. Let users >> resolve the dependencies of whatever they put in applets_sh. > > My view is that if script applets are provided as part of BusyBox they > should have all

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-09 Thread Michael Conrad
On 11/7/2018 9:18 AM, Kang-Che Sung wrote: On Wed, Nov 7, 2018 at 9:54 PM Tito wrote: this embedded scripts patch looks like "featuritis" at its best to me. It is adding complexity for solving what problem exactly: avoiding to copy the scripts manually to the new system or to the new firmware

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-07 Thread Kang-Che Sung
On Wed, Nov 7, 2018 at 9:54 PM Tito wrote: > > this embedded scripts patch looks like "featuritis" at its best to me. > It is adding complexity for solving what problem exactly: > avoiding to copy the scripts manually to the new system or to > the new firmware image? I see one advantage of

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-07 Thread Tito
On 07/11/18 08:42, Ron Yorston wrote: Kang-Che Sung wrote: Let no script applets individually configurable in menuconfig. Let users resolve the dependencies of whatever they put in applets_sh. My view is that if script applets are provided as part of BusyBox they should have all the

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-07 Thread Ron Yorston
Kang-Che Sung wrote: >You have already been needing 4 lines of config options for dependencies >(CAT & SLEEP & ECHO & SH) of a 3-line script applet (nologin). nologin is actually an extreme case among the samples in applets_sh. The others only require sed, though each does have a TODO suggesting

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-07 Thread Kang-Che Sung
On Wed, Nov 7, 2018 at 4:54 PM Ron Yorston wrote: > > Kang-Che Sung wrote: > >I think there is a potential for user to modify script applets heavily (for > >their particular application). Trying to track dependencies after user > >modification would be too much work for little benefit. > > Sure,

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-07 Thread Ron Yorston
Kang-Che Sung wrote: >I think there is a potential for user to modify script applets heavily (for >their particular application). Trying to track dependencies after user >modification would be too much work for little benefit. Sure, users can modify script applets and break them. But that's

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-07 Thread Kang-Che Sung
On Wed, Nov 7, 2018 at 3:42 PM Ron Yorston wrote: > > Kang-Che Sung wrote: > >Let no script applets individually configurable in menuconfig. Let users > >resolve the dependencies of whatever they put in applets_sh. > > My view is that if script applets are provided as part of BusyBox they >

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-06 Thread Ron Yorston
Kang-Che Sung wrote: >Let no script applets individually configurable in menuconfig. Let users >resolve the dependencies of whatever they put in applets_sh. My view is that if script applets are provided as part of BusyBox they should have all the features of native applets, including the usual

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-06 Thread Ron Yorston
Aaro Koskinen wrote: >Not sure if either of these are good. OK, we could use the old 'add another level of indirection' trick: config NOLOGIN_DEPENDENCIES bool "Dependencies for nologin" default y depends on NOLOGIN select CAT select ECHO select SLEEP

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-06 Thread Kang-Che Sung
Can we do this instead: Let no script applets individually configurable in menuconfig. Let users resolve the dependencies of whatever they put in applets_sh. If user needs to temporarily not embed a particular applet in that folder, then we can provide a "DISABLE" file that specifies which

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-06 Thread Aaro Koskinen
On Tue, Nov 06, 2018 at 08:36:11PM +, Ron Yorston wrote: > For the rest I suppose there are a couple of ways to look at it. The > dependencies could be specified explicitly: > >depends on ASH_EMBEDDED_SCRIPTS && (ECHO || ASH_ECHO) && CAT && SLEEP > > Though this means that scripted

Re: [RFC PATCH v2] Allow applets to be implemented as scripts

2018-11-06 Thread Ron Yorston
Aaro Koskinen wrote: >Should users be told they are enabling a scripted applet? E.g. in this >case they should also enable some other busybox applets to make it work >(like "cat")... I wonder how it's going to work in practice. Because of the 'depends on ASH_EMBEDDED_SCRIPTS' we known that we