[Buildroot] [PATCH 1/1] brotli: update to version 1.0.7

[Buildroot] [PATCH 1/1] brotli: update to version 1.0.7

[Adrian Perez de Castro]

The new version, among other changes, includes important fixes
for unaligned memory access on ARM (both for 32 and 64-bit), as well
as performance improvements and build fixes.

Patch "0001-Tell-CMake-to-not-check-for-a-C-compiler.patch" is not
needed due to the issue being fixed upstream, and therefore is removed.

Patch "0001-CMake-Allow-using-BUILD_SHARED_LIBS-to-choose-static.patch"
is rebased against the latest upstream changes.

Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
---
 ...g-BUILD_SHARED_LIBS-to-choose-static.patch | 16 +++++-----
 ...-CMake-to-not-check-for-a-C-compiler.patch | 31 -------------------
 package/brotli/brotli.hash                    |  2 +-
 package/brotli/brotli.mk                      |  2 +-
 4 files changed, 10 insertions(+), 41 deletions(-)
 delete mode 100644 package/brotli/0001-Tell-CMake-to-not-check-for-a-C-compiler.patch

diff --git a/package/brotli/0001-CMake-Allow-using-BUILD_SHARED_LIBS-to-choose-static.patch b/package/brotli/0001-CMake-Allow-using-BUILD_SHARED_LIBS-to-choose-static.patch
index 33b9bb70ce..ae5386b1ce 100644
--- a/package/brotli/0001-CMake-Allow-using-BUILD_SHARED_LIBS-to-choose-static.patch
+++ b/package/brotli/0001-CMake-Allow-using-BUILD_SHARED_LIBS-to-choose-static.patch
@@ -1,4 +1,4 @@
-From b60b613e7c2c9bf7a142c3c486ac6e77ad93f5d1 Mon Sep 17 00:00:00 2001
+From 7289e5a378ba13801996a84d89d8fe95c3fc4c11 Mon Sep 17 00:00:00 2001
 From: Adrian Perez de Castro <aperez@igalia.com>
 Date: Mon, 26 Mar 2018 19:08:31 +0100
 Subject: [PATCH] CMake: Allow using BUILD_SHARED_LIBS to choose static/shared
@@ -27,7 +27,7 @@ Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
 Upstream-Status: Submitted [https://github.com/google/brotli/pull/655]
 
 diff --git a/CMakeLists.txt b/CMakeLists.txt
-index 99b9258..3867931 100644
+index fc45f80..3f87f13 100644
 --- a/CMakeLists.txt
 +++ b/CMakeLists.txt
 @@ -6,6 +6,8 @@ cmake_minimum_required(VERSION 2.8.6)
@@ -120,25 +120,25 @@ index 99b9258..3867931 100644
      DIRECTORY ${BROTLI_INCLUDE_DIRS}/brotli
      DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}"
 diff --git a/c/fuzz/test_fuzzer.sh b/c/fuzz/test_fuzzer.sh
-index 5c754e1..e85e12f 100755
+index 9985194..4b99947 100755
 --- a/c/fuzz/test_fuzzer.sh
 +++ b/c/fuzz/test_fuzzer.sh
-@@ -14,12 +14,12 @@ mkdir bin
+@@ -13,12 +13,12 @@ mkdir bin
  cd bin
  
- cmake $BROTLI -DCMAKE_C_COMPILER="$CC" -DCMAKE_CXX_COMPILER="$CXX" \
+ cmake $BROTLI -DCMAKE_C_COMPILER="$CC" \
 -    -DBUILD_TESTING=OFF -DENABLE_SANITIZER=address
 -make -j$(nproc) brotlidec-static
 +    -DBUILD_TESTING=OFF -DBUILD_SHARED_LIBS=OFF -DENABLE_SANITIZER=address
 +make -j$(nproc) brotlidec
  
- ${CXX} -o run_decode_fuzzer -std=c++11 -fsanitize=address -I$SRC/include \
-     $SRC/fuzz/decode_fuzzer.cc $SRC/fuzz/run_decode_fuzzer.cc \
+ ${CC} -o run_decode_fuzzer -std=c99 -fsanitize=address -I$SRC/include \
+     $SRC/fuzz/decode_fuzzer.c $SRC/fuzz/run_decode_fuzzer.c \
 -    ./libbrotlidec-static.a ./libbrotlicommon-static.a
 +    ./libbrotlidec.a ./libbrotlicommon.a
  
  mkdir decode_corpora
  unzip $BROTLI/java/org/brotli/integration/fuzz_data.zip -d decode_corpora
 -- 
-2.16.3
+2.19.1
 
diff --git a/package/brotli/0001-Tell-CMake-to-not-check-for-a-C-compiler.patch b/package/brotli/0001-Tell-CMake-to-not-check-for-a-C-compiler.patch
deleted file mode 100644
index 410de2830e..0000000000
--- a/package/brotli/0001-Tell-CMake-to-not-check-for-a-C-compiler.patch
+++ /dev/null
@@ -1,31 +0,0 @@
-From fea0b1e46c486225d57e730cc0f94fa06b5b93fc Mon Sep 17 00:00:00 2001
-From: Adrian Perez de Castro <aperez@igalia.com>
-Date: Mon, 26 Mar 2018 12:12:00 +0100
-Subject: [PATCH] Tell CMake to not check for a C++ compiler
-
-By default CMake checks both for C and C++ compilers, while the latter
-is not needed. Setting the list of languages to just "C" in the call to
-project() removes the unneeded check.
----
- CMakeLists.txt | 2 +-
- 1 file changed, 1 insertion(+), 1 deletion(-)
-
-Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
-Upstream-Status: Submitted [https://github.com/google/brotli/pull/653]
-
-diff --git a/CMakeLists.txt b/CMakeLists.txt
-index 2dc7232..3fbcbfb 100644
---- a/CMakeLists.txt
-+++ b/CMakeLists.txt
-@@ -4,7 +4,7 @@
- # support 2.8.7.
- cmake_minimum_required(VERSION 2.8.6)
- 
--project(brotli)
-+project(brotli LANGUAGES C)
- 
- # If Brotli is being bundled in another project, we don't want to
- # install anything.  However, we want to let people override this, so
--- 
-2.16.3
-
diff --git a/package/brotli/brotli.hash b/package/brotli/brotli.hash
index 954d44eb70..82163040d1 100644
--- a/package/brotli/brotli.hash
+++ b/package/brotli/brotli.hash
@@ -1,5 +1,5 @@
 # Locally generated:
-sha512  93adcf437d730ac403e444285ac8aefbb2c8a6b5e1b064e8ee33684c067287a8159e0ee73d2217c167881e87da73fa494792d963a15508fd42b2ac4a5b52823c  v1.0.3.tar.gz
+sha512  a82362aa36d2f2094bca0b2808d9de0d57291fb3a4c29d7c0ca0a37e73087ec5ac4df299c8c363e61106fccf2fe7f58b5cf76eb97729e2696058ef43b1d3930a  v1.0.7.tar.gz
 
 # Hash for license files:
 sha512  bae78184c2f50f86d8c727826d3982c469454c42b9af81f4ef007e39036434fa894cf5be3bf5fc65b7de2301f0a72d067a8186e303327db8a96bd14867e0a3a8  LICENSE
diff --git a/package/brotli/brotli.mk b/package/brotli/brotli.mk
index cf1e811543..2c1ad48753 100644
--- a/package/brotli/brotli.mk
+++ b/package/brotli/brotli.mk
@@ -4,7 +4,7 @@
 #
 ################################################################################
 
-BROTLI_VERSION = 1.0.3
+BROTLI_VERSION = 1.0.7
 BROTLI_SOURCE = v$(BROTLI_VERSION).tar.gz
 BROTLI_SITE = https://github.com/google/brotli/archive
 BROTLI_LICENSE = MIT
-- 
2.19.1

[Buildroot] [PATCH 1/1] brotli: update to version 1.0.7

[Peter Korsgaard]

>>>>> "Adrian" == Adrian Perez de Castro <aperez@igalia.com> writes:

 > The new version, among other changes, includes important fixes
 > for unaligned memory access on ARM (both for 32 and 64-bit), as well
 > as performance improvements and build fixes.

Does this mean that we should backport this version bump to the current
LTS release (2018.02.x) as well then?

 > Patch "0001-Tell-CMake-to-not-check-for-a-C-compiler.patch" is not
 > needed due to the issue being fixed upstream, and therefore is removed.

 > Patch "0001-CMake-Allow-using-BUILD_SHARED_LIBS-to-choose-static.patch"
 > is rebased against the latest upstream changes.

 > Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>

Committed, thanks.

-- 
Bye, Peter Korsgaard

[Buildroot] [PATCH 1/1] brotli: update to version 1.0.7

[Adrian Perez de Castro]

On Fri, 26 Oct 2018 12:01:21 +0200, Peter Korsgaard <peter@korsgaard.com> wrote:
> >>>>> "Adrian" == Adrian Perez de Castro <aperez@igalia.com> writes:
> 
>  > The new version, among other changes, includes important fixes
>  > for unaligned memory access on ARM (both for 32 and 64-bit), as well
>  > as performance improvements and build fixes.
> 
> Does this mean that we should backport this version bump to the current
> LTS release (2018.02.x) as well then?

This is probably a good idea. I was reluctant to suggest it from the get-go
because there is no mention to security updates in the release notes, but
the unaligned memory access will cause crashes, which on could argue can be
exploited for DoS attacks ?. Also, it's a point release so the API/ABI of
the library remains the same, and the risk of breaking things is minimal.

So yes, I think it's a good idea to backport the update to the LTS version.

>  > Patch "0001-Tell-CMake-to-not-check-for-a-C-compiler.patch" is not
>  > needed due to the issue being fixed upstream, and therefore is removed.
> 
>  > Patch "0001-CMake-Allow-using-BUILD_SHARED_LIBS-to-choose-static.patch"
>  > is rebased against the latest upstream changes.
> 
>  > Signed-off-by: Adrian Perez de Castro <aperez@igalia.com>
> 
> Committed, thanks.

\o/


Cheers,

-Adri?n
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 195 bytes
Desc: not available
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20181026/bcfa6b8b/attachment.asc>

[Buildroot] [PATCH 1/1] brotli: update to version 1.0.7

[Peter Korsgaard]

>>>>> "Adrian" == Adrian Perez de Castro <aperez@igalia.com> writes:

 > On Fri, 26 Oct 2018 12:01:21 +0200, Peter Korsgaard <peter@korsgaard.com> wrote:
 >> >>>>> "Adrian" == Adrian Perez de Castro <aperez@igalia.com> writes:
 >> 
 >> > The new version, among other changes, includes important fixes
 >> > for unaligned memory access on ARM (both for 32 and 64-bit), as well
 >> > as performance improvements and build fixes.
 >> 
 >> Does this mean that we should backport this version bump to the current
 >> LTS release (2018.02.x) as well then?

 > This is probably a good idea. I was reluctant to suggest it from the get-go
 > because there is no mention to security updates in the release notes, but
 > the unaligned memory access will cause crashes, which on could argue can be
 > exploited for DoS attacks ?. Also, it's a point release so the API/ABI of
 > the library remains the same, and the risk of breaking things is minimal.

 > So yes, I think it's a good idea to backport the update to the LTS version.

Ok, thanks - I'll cherry pick it next time I sync LTS with master.

-- 
Bye, Peter Korsgaard

[Buildroot] [PATCH 1/1] brotli: update to version 1.0.7

[Peter Korsgaard]

>>>>> "Peter" == Peter Korsgaard <peter@korsgaard.com> writes:

>>>>> "Adrian" == Adrian Perez de Castro <aperez@igalia.com> writes:
 >> On Fri, 26 Oct 2018 12:01:21 +0200, Peter Korsgaard <peter@korsgaard.com> wrote:
 >>> >>>>> "Adrian" == Adrian Perez de Castro <aperez@igalia.com> writes:
 >>> 
 >>> > The new version, among other changes, includes important fixes
 >>> > for unaligned memory access on ARM (both for 32 and 64-bit), as well
 >>> > as performance improvements and build fixes.
 >>> 
 >>> Does this mean that we should backport this version bump to the current
 >>> LTS release (2018.02.x) as well then?

 >> This is probably a good idea. I was reluctant to suggest it from the get-go
 >> because there is no mention to security updates in the release notes, but
 >> the unaligned memory access will cause crashes, which on could argue can be
 >> exploited for DoS attacks ?. Also, it's a point release so the API/ABI of
 >> the library remains the same, and the risk of breaking things is minimal.

 >> So yes, I think it's a good idea to backport the update to the LTS version.

 > Ok, thanks - I'll cherry pick it next time I sync LTS with master.

Committed to 2018.02.x and 2018.08.x, thanks.

-- 
Bye, Peter Korsgaard