| Message ID | 20190714090301.26552-1-yann.morin.1998@free.fr |
|---|---|
| State | Changes Requested |
| Headers | show |
| Series | package/cpio: add host version | expand |
On 14/07/2019 11:03, Yann E. MORIN wrote: > The latest cpio has a --reproducible option, which may come handy when > we try to, well, be reproducible... > > Reported-by: Arnout Vandecappelle <arnout@mind.be> > Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr> > Cc: Arnout Vandecappelle <arnout@mind.be> > Cc: Atharva Lele <itsatharva@gmail.com> > --- > package/cpio/cpio.mk | 3 +++ > 1 file changed, 3 insertions(+) > > diff --git a/package/cpio/cpio.mk b/package/cpio/cpio.mk > index 9a09279572..d2814b87a5 100644 > --- a/package/cpio/cpio.mk > +++ b/package/cpio/cpio.mk > @@ -16,4 +16,7 @@ ifeq ($(BR2_PACKAGE_ARGP_STANDALONE),y) > CPIO_DEPENDENCIES += argp-standalone > endif > > +HOST_CPIO_CONF_OPTS = --bindir=$(HOST_DIR)/bin Why is this needed? For target it is needed because --prefix is set to /usr so it would end up in /usr/bin and we want it in /bin. But for hsot, --prefix is $(HOST_DIR). Regards, Arnout > + > $(eval $(autotools-package)) > +$(eval $(host-autotools-package)) >
Arnout, All, On 2019-07-14 14:19 +0200, Arnout Vandecappelle spake thusly: > On 14/07/2019 11:03, Yann E. MORIN wrote: > > The latest cpio has a --reproducible option, which may come handy when > > we try to, well, be reproducible... > > > > Reported-by: Arnout Vandecappelle <arnout@mind.be> > > Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr> > > Cc: Arnout Vandecappelle <arnout@mind.be> > > Cc: Atharva Lele <itsatharva@gmail.com> > > --- > > package/cpio/cpio.mk | 3 +++ > > 1 file changed, 3 insertions(+) > > > > diff --git a/package/cpio/cpio.mk b/package/cpio/cpio.mk > > index 9a09279572..d2814b87a5 100644 > > --- a/package/cpio/cpio.mk > > +++ b/package/cpio/cpio.mk > > @@ -16,4 +16,7 @@ ifeq ($(BR2_PACKAGE_ARGP_STANDALONE),y) > > CPIO_DEPENDENCIES += argp-standalone > > endif > > > > +HOST_CPIO_CONF_OPTS = --bindir=$(HOST_DIR)/bin > > Why is this needed? For target it is needed because --prefix is set to /usr so > it would end up in /usr/bin and we want it in /bin. But for hsot, --prefix is > $(HOST_DIR). You know what? That is exactly what I think, too. ;-) I just did a quick hack locally to test the build time of cpio, and did not try to think further, so I stupidly copied the target variant as-is without thinking too much. Then I saw it worked, got the required timings from a batch of 5 runs, finished my reply to Atharva's fs/cpio patch, and decided to send this patch too, without realising that it needed a bit more love. Atharva, when you respin your fs/cpio patch, can you grab this one in your tree and fix it as Arnout pointed out? Regards, Yann E. MORIN. > Regards, > Arnout > > > + > > $(eval $(autotools-package)) > > +$(eval $(host-autotools-package)) > >
On Sunday, July 14, 2019 6:08:59 PM IST Yann E. MORIN wrote: > Arnout, All, > > On 2019-07-14 14:19 +0200, Arnout Vandecappelle spake thusly: > > On 14/07/2019 11:03, Yann E. MORIN wrote: > > > The latest cpio has a --reproducible option, which may come handy when > > > we try to, well, be reproducible... > > > > > > Reported-by: Arnout Vandecappelle <arnout@mind.be> > > > Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr> > > > Cc: Arnout Vandecappelle <arnout@mind.be> > > > Cc: Atharva Lele <itsatharva@gmail.com> > > > --- > > > > > > package/cpio/cpio.mk | 3 +++ > > > 1 file changed, 3 insertions(+) > > > > > > diff --git a/package/cpio/cpio.mk b/package/cpio/cpio.mk > > > index 9a09279572..d2814b87a5 100644 > > > --- a/package/cpio/cpio.mk > > > +++ b/package/cpio/cpio.mk > > > @@ -16,4 +16,7 @@ ifeq ($(BR2_PACKAGE_ARGP_STANDALONE),y) > > > > > > CPIO_DEPENDENCIES += argp-standalone > > > endif > > > > > > +HOST_CPIO_CONF_OPTS = --bindir=$(HOST_DIR)/bin > > > > Why is this needed? For target it is needed because --prefix is set to > > /usr so> > > it would end up in /usr/bin and we want it in /bin. But for hsot, --prefix > > is $(HOST_DIR). > > You know what? That is exactly what I think, too. ;-) > > I just did a quick hack locally to test the build time of cpio, and did > not try to think further, so I stupidly copied the target variant as-is > without thinking too much. > > Then I saw it worked, got the required timings from a batch of 5 runs, > finished my reply to Atharva's fs/cpio patch, and decided to send this > patch too, without realising that it needed a bit more love. > > Atharva, when you respin your fs/cpio patch, can you grab this one in > your tree and fix it as Arnout pointed out? I will :) > Regards, > Yann E. MORIN. > > > Regards, > > Arnout > > > > > + > > > > > > $(eval $(autotools-package)) > > > > > > +$(eval $(host-autotools-package)) Regards, Atharva Lele
diff --git a/package/cpio/cpio.mk b/package/cpio/cpio.mk index 9a09279572..d2814b87a5 100644 --- a/package/cpio/cpio.mk +++ b/package/cpio/cpio.mk @@ -16,4 +16,7 @@ ifeq ($(BR2_PACKAGE_ARGP_STANDALONE),y) CPIO_DEPENDENCIES += argp-standalone endif +HOST_CPIO_CONF_OPTS = --bindir=$(HOST_DIR)/bin + $(eval $(autotools-package)) +$(eval $(host-autotools-package))
The latest cpio has a --reproducible option, which may come handy when we try to, well, be reproducible... Reported-by: Arnout Vandecappelle <arnout@mind.be> Signed-off-by: Yann E. MORIN <yann.morin.1998@free.fr> Cc: Arnout Vandecappelle <arnout@mind.be> Cc: Atharva Lele <itsatharva@gmail.com> --- package/cpio/cpio.mk | 3 +++ 1 file changed, 3 insertions(+)