Propagate icu reverse dependencies.
Php lacks "-std=c++11", provided by icu-config, when compiling icu
support leading to a build failure:
/home/peko/autobuild/instance-1/output/host/usr/sh4-buildroot-linux-uclibc/sysroot/usr/include/unicode/umachine.h:347:13:
error: 'char16_t' does not name a type
typedef char16_t UChar;
Fixes
http://autobuild.buildroot.net/results/238/238d0e4bb205c0e7288903c6498e6c1787d86a1b/
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The check-package script when ran gives warnings on ordering issues
on all of these Config files. This patch cleans up all warnings
related to the ordering in the Config files for packages starting with
the letter p in the package directory.
The appropriate ordering is: type, default, depends on, select, help
See http://nightly.buildroot.org/#_config_files for more information.
Signed-off-by: Adam Duskett <Adamduskett@outlook.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
As per LICENSE file, php uses PHP license version 3.01.
Signed-off-by: Rahul Bedarkar <rahulbedarkar89@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Continue work started by Bernd Kuhls in
https://patchwork.ozlabs.org/patch/437544/
Signed-off-by: Bernd Kuhls <bernd.kuhls@t-online.de>
Signed-off-by: Fabrice Fontaine <fabrice.fontaine@orange.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Add a patch from a PHP Github pull request that fixes the build at -O0
on AArch64.
Signed-off-by: Tatsuyuki Ishi <ishitatsuyuki@gmail.com>
[Thomas: improved patch commit log.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The comments in the Buildroot-provided php-fpm.conf would produce the
following error when starting php-fpm:
ERROR: [/etc/php-fpm.conf:2] value is NULL for a ZEND_INI_PARSER_ENTRY
Removing the comments fixes the problem.
Signed-off-by: Frank Hunleth <fhunleth@troodon-software.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Commit c457213fbc reworked the selection
of PHP interfaces, moving from a choice..endchoice to a three
independent sub-options.
Thanks to how choices are handled by kconfig, an option suitable for
noMMU platforms was automatically chosen by default. When the
choice...endchoice got removed in
c457213fbc, we introduced a "select" for
one of the interfaces, to make sure at least one interface is
selected.
Unfortunately, the selected interface does not build on noMMU
platforms. In order to preserve the previous behavior (CGI interface
is selected by default on MMU capable platforms), this commit makes
the select logic a bit more complicated:
- CGI is selected on MMU-capable platforms if neither CLI or FPM are
selected
- CLI is selected on MMU-less platforms, since this is anyway the
only available interface.
Fixes:
http://autobuild.buildroot.net/results/85ea09188f1e996cb14b4b83a17736c096c569a8/
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Following suggestion of Yann Morin, rework selection of php
interfaces: use booleans instead of choice to be able to select
multiple interfaces as they are not exclusive. We make sure at least
one of the options is selected.
It should be noted that previously CGI and FPM could not be selected
at the same time. This is now possible. Bug that prevented compilation
of CGI and FPM binaries at the same time has been fixed since PHP 5.4
(https://github.com/php-build/php-build/issues/101).
Legacy handling is added for the two options that are removed, and the
appropriate new options are selected.
Signed-off-by: Fabrice Fontaine <fabrice.fontaine@orange.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
See http://www.php.net/ChangeLog-7.php#7.0.12 since there are no CVEs
out yet.
And drop upstream patch.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Set the PHP default MySQL socket path to match what the
mysql package is using.
Signed-off-by: Floris Bos <bos@je-eigen-domein.nl>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When building php, the package will also build/install phpdbg.
This patch explicitly disables this behaviour.
Signed-off-by: Theo Debrouwere <t.debrouwere@televic.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The Mysql Native Driver has been the default mysql driver since
PHP 5.4, but buildroot was still using libmysqlclient.
Mysqlnd has several advantages such as improved memory management
and the more favorable PHP licensing terms.
(can combine it with proprietary PHP extensions like Ioncube
loader, while libmysqlclient requires commercial licensing if you
link to it and do not fall under their GPL/FOSS license exception)
Signed-off-by: Floris Bos <bos@je-eigen-domein.nl>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In 5f37843a (php.ini: set date.timezone), the configured timezone was
used as the default for PHP.
However, BR2_TARGET_LOCALTIME is a string, so is quoted, so it is never
empty, so the check for emptynessnever matches.
Fix that by q-stripping the value before testing it. Note however that
we do not q-strip it before storing it in the php.ini file, because it
has to be q-stripped in there.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Floris Bos <bos@je-eigen-domein.nl>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fixes (no CVEs yet):
Buffer over-write in finfo_open with malformed magic file.
Invalid memory write in phar on filename with \0 in name.
Parsing of tar file with duplicate filenames causes memory leak.
php_snmp_error() Format String Vulnerability.
Integer Overflow in php_raw_url_encode.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The BR2_ARCH_HAS_ATOMICS was added because on ARC, atomic instructions
may not be provided by the architecture and therefore the compiler
does not provide the __sync_*() built-ins.
However, since then, icu was changed and is now able to use C++11
atomics, or even no atomic operations at all. In fact, icu will:
* If possible, it will use C++11 atomics, which internally rely on
the __atomic built-ins. These are available since gcc 4.7, and all
architectures provide it. On some architectures, you *must* link
with libatomic, on some other architectures, they are available
built-in, but in all cases, linking against libatomic does not
harm. Thanks to this, even ARC with no atomic support (which was
the original reason for adding the BR2_ARCH_HAS_ATOMICS) dependency
builds fine, provided -latomic is added to LIBS.
* If C++11 atomics are not available, then it falls back to
__sync_*() built-ins, which allows compilers older than 4.7 to be
supported.
* If really no atomic mechanism is available, then it falls back to a
basic implementation based on a mutex.
Conclusion:
- The BR2_ARCH_HAS_ATOMICS dependency is no longer needed.
- We need to link with -latomic when gcc >= 4.7 is used.
Note that reverse dependencies of icu are also changed accordingly.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The pkg-config tool is automatically passing --static when BR2_STATIC_LIBS
is set (see package/pkgconf/pkgconf.mk), so no need to pass it explicitly
from package .mk files.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
It is often difficult to know exactly when make will expand the
variable, and usually it can only be expanded after the dependencies
have been built (e.g. pkg-config or the .pc file). Using a backtick
instead makes it very clear that it will be expanded only while
executing the command.
This change is useful for two cases:
1. The per-package staging (and host) directory will be created as part
of the configure step, so any $(shell ...) variable that is used in
the configure step will fail because the directory doesn't exist
yet.
2. 'make printvars' evaluates the variables it prints. It will therefore
trigger a lot of errors from missing .pc files and others. The
backticks, on the other hand, are not expanded, so with this change
the output of 'make printvars' becomes clean again.
This commit contains only the easy changes: replace $(shell ...) with
`...`, and also replace ' with " where needed. Follow-up commits will
tackle the more complicated cases that need additional explanation.
After this change, the following instances of $(shell ...) will remain:
- All assignments that use :=
- All variables that are used in make conditionals (which don't expand
the backticks).
- All variables that only refer to system executables and make
variables that don't change.
- The calls to check-host-* in dependencies.mk, because it is eval'ed.
[Original patch by Fabio Porcedda, but extended quite a bit by Arnout.]
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Fixes:
CVE-2015-3152 - mysqlnd is vulnerable to BACKRONYM
And other security bugs with no CVE assigned yet:
Bug #69972 - Use-after-free vulnerability in
sqlite3SafetyCheckSickOrOk()
Bug # 69970 - Use-after-free vulnerability in
spl_recursive_it_move_forward_ex()
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
