* [Buildroot] [PATCH] busybox: enable fully featured hush shell for nommu
2015-10-14 22:36 [Buildroot] [PATCH] busybox: enable fully featured hush shell for nommu Gustavo Zacarias
@ 2015-10-15 7:40 ` Thomas Petazzoni
2015-10-15 16:36 ` Arnout Vandecappelle
1 sibling, 0 replies; 4+ messages in thread
From: Thomas Petazzoni @ 2015-10-15 7:40 UTC (permalink / raw)
To: buildroot
Dear Gustavo Zacarias,
On Wed, 14 Oct 2015 19:36:45 -0300, Gustavo Zacarias wrote:
> As pointed by Rich Fekler on IRC the basic hush shell is pretty much
> useless since it doesn't support if conditionals, loops, functions, case
> or even interactive mode.
> So enable the full feature-set.
> Size delta: +10184 bytes uncompressed for blackfin fdpic.
>
> Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
> ---
> package/busybox/busybox.mk | 15 +++++++++++++++
> 1 file changed, 15 insertions(+)
Applied, thanks.
Thomas
--
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
^ permalink raw reply [flat|nested] 4+ messages in thread
* [Buildroot] [PATCH] busybox: enable fully featured hush shell for nommu
2015-10-14 22:36 [Buildroot] [PATCH] busybox: enable fully featured hush shell for nommu Gustavo Zacarias
2015-10-15 7:40 ` Thomas Petazzoni
@ 2015-10-15 16:36 ` Arnout Vandecappelle
2015-10-16 9:55 ` Peter Korsgaard
1 sibling, 1 reply; 4+ messages in thread
From: Arnout Vandecappelle @ 2015-10-15 16:36 UTC (permalink / raw)
To: buildroot
On 15-10-15 00:36, Gustavo Zacarias wrote:
> As pointed by Rich Fekler on IRC the basic hush shell is pretty much
> useless since it doesn't support if conditionals, loops, functions, case
> or even interactive mode.
> So enable the full feature-set.
> Size delta: +10184 bytes uncompressed for blackfin fdpic.
>
> Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
> ---
> package/busybox/busybox.mk | 15 +++++++++++++++
> 1 file changed, 15 insertions(+)
>
> diff --git a/package/busybox/busybox.mk b/package/busybox/busybox.mk
> index 3494456..d3a6fd8 100644
> --- a/package/busybox/busybox.mk
> +++ b/package/busybox/busybox.mk
> @@ -89,6 +89,21 @@ define BUSYBOX_SET_MMU
> $(call KCONFIG_DISABLE_OPT,CONFIG_SWAPONOFF,$(BUSYBOX_BUILD_CONFIG))
> $(call KCONFIG_DISABLE_OPT,CONFIG_ASH,$(BUSYBOX_BUILD_CONFIG))
> $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_BASH_COMPAT,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_BRACE_EXPANSION,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_HELP,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_INTERACTIVE,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_SAVEHISTORY,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_JOB,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_TICK,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_IF,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_LOOPS,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_CASE,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_FUNCTIONS,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_LOCAL,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_RANDOM_SUPPORT,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_EXPORT_N,$(BUSYBOX_BUILD_CONFIG))
> + $(call KCONFIG_ENABLE_OPT,CONFIG_HUSH_MODE_X,$(BUSYBOX_BUILD_CONFIG))
I don't like this, since it makes it completely impossible to disable these
hush features. I admit that it indeed doesn't make a significant difference, but
on principle I'm against forcing this on the user.
So ideally, we should just let them default to y (all these options have a
default y). We can simply remove them from busybox.config.
But unfortunately, that doesn't work. The problem is that the kconfig infra
will run the oldconfig immediately after merge_config, before applying the
fixups. Since in busybox.config, ash is enabled and hush is disabled, all these
options will be turned off by the oldconfig run. To solve this, we should run
the fixups immediately after merge_config. I've tried this and it seems to work.
But then we're running the fixups three times in the configure step, which
feels a bit like overkill to me. So perhaps a more elegant solution would be the
following: instead of calling sed-based fixups, use a kconfig fragment and
merge_config to apply the fixups.
Cc-ing our Kconfig infra expert to consider this.
Regards,
Arnout
> endef
> endif
>
>
--
Arnout Vandecappelle arnout at mind be
Senior Embedded Software Architect +32-16-286500
Essensium/Mind http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint: 7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF
^ permalink raw reply [flat|nested] 4+ messages in thread