[Buildroot] [PATCH] package/openssl: bump version to 1.1.0f

[David Mosberger]

Arnout,

Thanks for the feedback.  I'll work on the things you pointed out.

  --david

On Thu, Aug 24, 2017 at 4:59 PM, Arnout Vandecappelle <arnout at mind.be>
wrote:

>  Hi David,
>
> On 25-08-17 00:01, David Mosberger-Tang wrote:
> > Signed-off-by: David Mosberger-Tang <davidm at egauge.net>
>
>  When you make big changes like this, the commit log should me much more
> extended to explain why all this has changed.
>
> [snip]
> > diff --git a/package/openssl/0001-Dont-waste-time-building-manpages-if-we-re-not-going.patch
> b/package/openssl/0001-Dont-waste-time-building-manpages-
> if-we-re-not-going.patch
> > deleted file mode 100644
>
>  Why is this patch no longer needed? has it been applied upstream? Do we
> now
> have a different way of avoiding to build the manpages? Can the manpages be
> built safely?
>
> [snip]
> > diff --git a/package/openssl/0002-cryptodev-Fix-issue-with-signature-generation.patch
> b/package/openssl/0002-cryptodev-Fix-issue-with-signature-generation.patch
> > deleted file mode 100644
>
>  Why is this patch no longer needed? Has it been applied upstream? If yes,
> please refer to the upstream commit(s) in the commit message. Does
> cryptodev
> build successfully now?
>
> [snip]
> > diff --git a/package/openssl/openssl.hash b/package/openssl/openssl.hash
> > index 064eeca..3411d5f 100644
> > --- a/package/openssl/openssl.hash
> > +++ b/package/openssl/openssl.hash
> > @@ -1,4 +1,5 @@
> > -# From https://www.openssl.org/source/openssl-1.0.2k.tar.gz.sha256
> > +# From https://www.openssl.org/source/openssl-1.1.0f.tar.gz.sha256
> > +sha256       12f746f3f2493b2f39da7ecf63d7ee
> 19c6ac9ec6a4fcd8c229da8a522cb12765        openssl-1.1.0f.tar.gz
> >  sha256       6b3977c61f2aedf0f96367dcfb5c6e
> 578cf37e7b8d913b4ecb6643c3cb88d8c0        openssl-1.0.2k.tar.gz
> >  # Locally computed
> >  sha256       eddd8a5123748052c598214487ac17
> 8e4bfa4e31ba2ec520c70d59c8c5bfa2e9        openssl-1.0.2a-parallel-
> install-dirs.patch?id=c8abcbe8de5d3b6cdd68c162f398c011ff6e2d9d
>
>  If you remove the patches downloaded from gentoo, then their hashes
> should be
> removed as well.
>
> [snip]
> > -OPENSSL_PATCH = \
> > -     https://gitweb.gentoo.org/repo/gentoo.git/plain/dev-
> libs/openssl/files/openssl-1.0.2d-parallel-build.patch?id=
> c8abcbe8de5d3b6cdd68c162f398c011ff6e2d9d \
> > -     https://gitweb.gentoo.org/repo/gentoo.git/plain/dev-
> libs/openssl/files/openssl-1.0.2a-parallel-obj-headers.patch?id=
> c8abcbe8de5d3b6cdd68c162f398c011ff6e2d9d \
> > -     https://gitweb.gentoo.org/repo/gentoo.git/plain/dev-
> libs/openssl/files/openssl-1.0.2a-parallel-install-dirs.patch?id=
> c8abcbe8de5d3b6cdd68c162f398c011ff6e2d9d \
> > -     https://gitweb.gentoo.org/repo/gentoo.git/plain/dev-
> libs/openssl/files/openssl-1.0.2a-parallel-symlinking.patch?id=
> c8abcbe8de5d3b6cdd68c162f398c011ff6e2d9d
>
>  Why can these patches be removed? Have they been applied upstream? If yes,
> please refer to the commit IDs. Does parallel build work reliably now?
>
> >
> >  # relocation truncated to fit: R_68K_GOT16O
> >  ifeq ($(BR2_m68k_cf),y)
> > @@ -84,7 +79,6 @@ define OPENSSL_CONFIGURE_CMDS
> >                       no-rc5 \
> >                       enable-camellia \
> >                       enable-mdc2 \
> > -                     enable-tlsext \
>
>  Why is this removed?
>
> >                       $(if $(BR2_STATIC_LIBS),zlib,zlib-dynamic) \
> >                       $(if $(BR2_STATIC_LIBS),no-dso) \
> >       )
> > @@ -110,7 +104,8 @@ define OPENSSL_BUILD_CMDS
> >  endef
> >
> >  define OPENSSL_INSTALL_STAGING_CMDS
> > -     $(TARGET_MAKE_ENV) $(MAKE) -C $(@D) INSTALL_PREFIX=$(STAGING_DIR)
> install
> > +     $(TARGET_MAKE_ENV) $(MAKE) -C $(@D) DESTDIR=$(STAGING_DIR) \
> > +             install_engines install_runtime install_ssldirs
>
>  Why doesn't the install target work any more?
>
> >  endef
> >
> >  define HOST_OPENSSL_INSTALL_CMDS
> > @@ -118,8 +113,8 @@ define HOST_OPENSSL_INSTALL_CMDS
> >  endef
> >
> >  define OPENSSL_INSTALL_TARGET_CMDS
> > -     $(TARGET_MAKE_ENV) $(MAKE) -C $(@D) INSTALL_PREFIX=$(TARGET_DIR)
> install
> > -     rm -rf $(TARGET_DIR)/usr/lib/ssl
>
>  Why is this no longer needed?
>
> > +     $(TARGET_MAKE_ENV) $(MAKE) -C $(@D) DESTDIR=$(TARGET_DIR) \
> > +             install_engines install_runtime install_ssldirs
> >       rm -f $(TARGET_DIR)/usr/bin/c_rehash
> >  endef
> >
> > @@ -136,7 +131,7 @@ endif
> >  ifneq ($(BR2_STATIC_LIBS),y)
> >  # libraries gets installed read only, so strip fails
>
>  Is this still needed? I.e. are libs still installed readonly?
>
> >  define OPENSSL_INSTALL_FIXUPS_SHARED
> > -     chmod +w $(TARGET_DIR)/usr/lib/engines/lib*.so
> > +     chmod +w $(TARGET_DIR)/usr/lib/engines-1.1/*.so
>
>  Hm, perhaps we should introduce OPENSSL_VERSION_MAJOR and use that here...
>
>  Regards,
>  Arnout
>
> >       for i in $(addprefix $(TARGET_DIR)/usr/lib/,libcrypto.so.*
> libssl.so.*); \
> >       do chmod +w $$i; done
> >  endef
> >
>
> --
> 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
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20170825/7de75ce1/attachment-0002.html>