package/libsodium: add config for full build
PATCH v4 1/1 package/libsodium: add config for full build
A new option is added to build all functions.
This option was implicitly used before, so it is activated by default to avoid unpleasant surprises.
---
Changes v3 -> v4:
- better commit log (suggested by Thomas Petazzoni)
Changes v2 -> v3:
- add a better help (suggested by Arnout Vandecappelle)
Changes v1 -> v2:
- full build (suggested by Arnout Vandecappelle)
--- package/libsodium/Config.in | 12 ++++++++++++ package/libsodium/libsodium.mk | 4 ++++ 2 files changed, 16 insertions(+)
package/libsodium/Config.in +12 -0
@@ -4,3 +4,15 @@ config BR2_PACKAGE_LIBSODIUM A modern and easy-to-use crypto library. http://libsodium.org/ + +if BR2_PACKAGE_LIBSODIUM + +config BR2_PACKAGE_LIBSODIUM_FULL + bool "all functions" + default y + help + Build all seldom used and obsolete functions. Without + this option, only the minimal set of normally used + functions is built. + +endif
package/libsodium/libsodium.mk +4 -0
@@ -14,5 +14,9 @@ ifeq ($(BR2_TOOLCHAIN_SUPPORTS_PIE),) LIBSODIUM_CONF_OPTS += --disable-pie endif +ifeq ($(BR2_PACKAGE_LIBSODIUM_FULL),) +LIBSODIUM_CONF_OPTS += --enable-minimal +endif + $(eval $(autotools-package)) $(eval $(host-autotools-package))
Signature · 1 line
2.19.1
PATCH v4 1/1 package/libsodium: add config for full build
Hello,
Quoted 14 lines · On Mon, 23 Sep 2019 08:29:03 +0000 Adrien Gallou?t <[email protected]> wrote:
package/libsodium/Config.in +9
-0+config BR2_PACKAGE_LIBSODIUM_FULL + bool "all functions" + default y + help + Build all seldom used and obsolete functions. Without + this option, only the minimal set of normally used + functions is built. + +endifpackage/libsodium/libsodium.mk +0
-0
package/libsodium/libsodium.mk ifeq ($(BR2_TOOLCHAIN_SUPPORTS_PIE),) · line 18
Quoted 5 lines
package/libsodium/libsodium.mk +2
-0@@ -14,5 +14,9 @@ ifeq ($(BR2_TOOLCHAIN_SUPPORTS_PIE),) LIBSODIUM_CONF_OPTS += --disable-pie endif +ifeq ($(BR2_PACKAGE_LIBSODIUM_FULL),) +LIBSODIUM_CONF_OPTS += --enable-minimal
But then, does it make sense to invert the option and name it BR2_PACKAGE_LIBSODIUM_MINIMAL ? It would match the --enable-minimal/--disable-minimal autoconf options of libsodium.
Also, we need to pass --disable-minimal in an "else" clause.
Thanks,
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
PATCH v4 1/1 package/libsodium: add config for full build
Hello,
Quoted 23 lines · Thomas Petazzoni
Hello,
Quoted 17 lines · On Mon, 23 Sep 2019 08:29:03 +0000 Adrien Gallou?t <[email protected]> wrote:
package/libsodium/Config.in +9
-0+config BR2_PACKAGE_LIBSODIUM_FULL + bool "all functions" + default y + help + Build all seldom used and obsolete functions. Without + this option, only the minimal set of normally used + functions is built. + +endifpackage/libsodium/libsodium.mk +0
-0@@ -14,5 +14,9 @@ ifeq ($(BR2_TOOLCHAIN_SUPPORTS_PIE),) LIBSODIUM_CONF_OPTS += --disable-pie endif
Quoted 6 lines
package/libsodium/libsodium.mk ifeq ($(BR2_TOOLCHAIN_SUPPORTS_PIE),) · line 18
Quoted 2 lines
package/libsodium/libsodium.mk +2
-0+ifeq ($(BR2_PACKAGE_LIBSODIUM_FULL),) +LIBSODIUM_CONF_OPTS += --enable-minimalBut then, does it make sense to invert the option and name it
BR2_PACKAGE_LIBSODIUM_MINIMAL? It would match the --enable-minimal/--disable-minimal autoconf options of libsodium.
That was my first patch :) Do you want me to resubmit it ?
Quoted 10 lines
Also, we need to pass --disable-minimal in an "else" clause.
Thanks,
Thomas
--
Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
Regards,
Adrien
PATCH v4 1/1 package/libsodium: add config for full build
Quoted 5 lines · On Sat, 28 Sep 2019 23:49:56 +0200 Adrien Gallou?t <[email protected]> wrote:
Quoted 3 lines
But then, does it make sense to invert the option and name it
BR2_PACKAGE_LIBSODIUM_MINIMAL? It would match the --enable-minimal/--disable-minimal autoconf options of libsodium.That was my first patch :) Do you want me to resubmit it ?
I don't have a very clear-cut opinion.
On one hand, it is weird to have a Config.in option to reduce/remove something. But on the other hand, it makes sense to have Config.in options that are as much as possible mapped to the corresponding options provided by the package build system.
Arnout, Peter, any preference ?
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
PATCH v4 1/1 package/libsodium: add config for full build
Quoted 9 lines · Thomas Petazzoni
Quoted 5 lines · On Sat, 28 Sep 2019 23:49:56 +0200 Adrien Gallou?t <[email protected]> wrote:
Quoted 3 lines
But then, does it make sense to invert the option and name it
BR2_PACKAGE_LIBSODIUM_MINIMAL? It would match the --enable-minimal/--disable-minimal autoconf options of libsodium.That was my first patch :) Do you want me to resubmit it ?
Quoted 8 lines
I don't have a very clear-cut opinion.
On one hand, it is weird to have a Config.in option to reduce/remove something. But on the other hand, it makes sense to have Config.in options that are as much as possible mapped to the corresponding options provided by the package build system.
Arnout, Peter, any preference ?
It was me who suggested to turn around the logic.
I believe we have had problems in the past where there was some MINIMAL option,
but then some other package needed something from the non-minimal set. Then it has to do a 'depends on' instead of a 'select'.
So it's not a matter of being weird, it's a matter of making sure we don't
create problems in the future.
Quoted 1 line
Quoted 1 line
Also, we need to pass --disable-minimal in an "else" clause.
This is still true. Adrien, I believe I did say something to that effect in an
earlier iteration, and I also believe you replied to it. But my memory is worthless :-)
Regards,
Arnout
PATCH v4 1/1 package/libsodium: add config for full build
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Arnout" == Arnout Vandecappelle <[email protected]> writes:
Hi,
Quoted 10 lines
Quoted 1 line
Arnout, Peter, any preference ?
It was me who suggested to turn around the logic.
I believe we have had problems in the past where there was some MINIMAL option,
but then some other package needed something from the non-minimal set. Then it has to do a 'depends on' instead of a 'select'.So it's not a matter of being weird, it's a matter of making sure we don't
create problems in the future.
Yes, indeed - I would prefer to keep it like this (a "positive" option).
Quoted 5 lines
Quoted 1 line
Quoted 1 line
Also, we need to pass --disable-minimal in an "else" clause.
This is still true. Adrien, I believe I did say something to that effect in an
earlier iteration, and I also believe you replied to it. But my memory is worthless :-)
Heh ;)
Signature · 1 line
Bye, Peter Korsgaard
PATCH v4 1/1 package/libsodium: add config for full build
In reply to Arnout VandecappelleQuoted 3 lines · On Sun, 29 Sep 2019 12:51:35 +0200 Arnout Vandecappelle <[email protected]> wrote:
I believe we have had problems in the past where there was some MINIMAL option,
but then some other package needed something from the non-minimal set. Then it has to do a 'depends on' instead of a 'select'.
That makes a lot of sense indeed. So we should always have option that "add" more stuff, and never option that "remove" stuff for that reason.
Thanks for the feedback, I hadn't thought about this.
Thomas
Signature · 3 lines
Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com
PATCH v4 1/1 package/libsodium: add config for full build
In reply to Arnout VandecappelleQuoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Arnout" == Arnout Vandecappelle <[email protected]> writes:
Hi,
Quoted 9 lines
Quoted 8 lines
I don't have a very clear-cut opinion.
On one hand, it is weird to have a Config.in option to reduce/remove something. But on the other hand, it makes sense to have Config.in options that are as much as possible mapped to the corresponding options provided by the package build system.
Arnout, Peter, any preference ?
Quoted 8 lines
It was me who suggested to turn around the logic.
I believe we have had problems in the past where there was some MINIMAL option,
but then some other package needed something from the non-minimal set. Then it has to do a 'depends on' instead of a 'select'.So it's not a matter of being weird, it's a matter of making sure we don't
create problems in the future.
Agreed, having options to disable stuff gets tricky when you have other packages needing the disabled stuff.
Quoted 5 lines
Quoted 1 line
Quoted 1 line
Also, we need to pass --disable-minimal in an "else" clause.
This is still true. Adrien, I believe I did say something to that effect in an
earlier iteration, and I also believe you replied to it. But my memory is worthless :-)
Getting old? :P
Signature · 1 line
Bye, Peter Korsgaard