buildroot/package/mongrel2, branch 2016.02

mongrel2: allow using uClibc but only on certain architectures

2016-02-01T18:32:42+00:00

mongrel2 uses the {get,make,swap}context functions: - With glibc, no problem, they are available on all supported architectures - With uClibc, they are available only on a subset of the architectures. Until now, only BR2_UCLIBC_VERSION_SNAPSHOT configurations were allowed to select mongrel2, but we are going to get rid of the uClibc snapshot version, and uClibc-ng is as capable as the uClibc snapshot. However, only certain architectures have the *context() functions. - With musl, there is no *context() support. Since this dependency is quite complicated, we introduce a BR2_PACKAGE_MONGREL2_LIBC_SUPPORTS hidden boolean to encode which C libraries are supported. Also, listing the supported uClibc architectures would be too long in the comment, so we simply indicate that the package needs uClibc or glibc. Signed-off-by: Thomas Petazzoni Signed-off-by: Peter Korsgaard

mongrel2: disable polarssl assembly as we do in the polarssl package

2015-11-27T18:28:45+00:00

Fixes: http://autobuild.buildroot.net/results/a92292a1defd441938d3de6c5049473d520ebc4e/ http://autobuild.buildroot.net/results/f3db555dc29e0e0e87762fcace913853ffe2b176/ mongrel2 bundles polarssl so it has the same problem as the polarssl with the assembly optimization, so let's use the same solution here. Signed-off-by: Vicente Olivert Riera Signed-off-by: Peter Korsgaard

mongrel2: add upstream patch to fix symbol conflict

2015-11-21T08:32:40+00:00

The symbol 'max_align_t', defined in the mongrel2 sources (src/mem/align.h), is also defined in one of the standard headers in some toolchains (the ones based on GCC 5+ apparently). This conflict has been solved upstream, and a new version (1.10.0) containing the fix has been released already, so this backported patch may be dropped in the future when mongrel2 gets a version bump. Fixes: http://autobuild.buildroot.net/results/b7f/b7f86bdda094de96633e9921e0e5895eadc89217/ http://autobuild.buildroot.net/results/f62/f623703aa6d7992b67d305331734e4ed0c7b2be4/ http://autobuild.buildroot.net/results/eb3/eb385cb4f87078122028e85cbe31caa01d794991/ ... Signed-off-by: Rodrigo Rebello Signed-off-by: Thomas Petazzoni

mongrel2: uses dlopen(), needs dynamic library support

2015-11-08T15:05:21+00:00

Fixes: http://autobuild.buildroot.net/results/c5e/c5ecfe4a5ad3712d65293de517a24e76c4376c57/ Signed-off-by: Thomas Petazzoni

mongrel2: Bump version and change hosting

2015-07-28T07:52:53+00:00

mongrel2 has changed it's repository. It was also the opportunity to bump it. Adding a hash, since we use this packages as an example for the hash in the manual, it was a bit shamefull that it did not have a proper hash. Remove 0001-use-CC-variable-to-compile-procer.patch, an equivalent was applied upstream: https://github.com/mongrel2/mongrel2/commit/eafe3cf3efde4278040f38b5808f336e3d102c84 Signed-off-by: Maxime Hadjinlian Reviewed-by: Lionel Orry Signed-off-by: Thomas Petazzoni

packages: remove non-IPv6 dependencies and tweaks

2015-04-22T21:06:35+00:00

Now that IPv6 is mandatory remove package dependencies and conditionals for it. Signed-off-by: Gustavo Zacarias Signed-off-by: Thomas Petazzoni

packages: remove (non-)lfs dependencies and tweaks

2015-04-01T20:47:22+00:00

Now that largefile is mandatory removes package dependencies and conditionals. Signed-off-by: Gustavo Zacarias Signed-off-by: Thomas Petazzoni

package/*: rename patches according to the new policy

2015-02-03T13:52:56+00:00

Autogenerated from rename-patch.py (http://patchwork.ozlabs.org/patch/403345) Signed-off-by: Samuel Martin Signed-off-by: Peter Korsgaard

mongrel2: Add comments to avoid use of github call

2014-10-30T21:49:59+00:00

As the comment says, we don't want to uses the github helper, as we would end up with a generated tarball, which is not the same one as the one linked with the releases. Signed-off-by: Maxime Hadjinlian Reviewed-by: "Yann E. MORIN" Signed-off-by: Thomas Petazzoni

package: remove the trailing slash sign from _SITE variable

2014-07-31T21:17:46+00:00

Since the trailing slash is stripped from $($(PKG)_SITE) by pkg-generic.mk: $(call DOWNLOAD,$($(PKG)_SITE:/=)/$($(PKG)_SOURCE)) so it is redundant. This patch removes it from $(PKG)_SITE variable for BR consistency. Signed-off-by: Jerzy Grzegorek Signed-off-by: Thomas Petazzoni