correctly if the compiler optimizes too clever.
This has happaned on HP-UX 11.23 (IA64) at optimization
level "+O2", causing bug#42213:
Check for "stack overrun" doesn't work, server crashes
Fix it by adding a pragma that prevents this optimization.
As a result, it should be safe to use "+O2" on this platform
(unless there is some other, optimizer-related, bug which
is just currently masked because we use resudec optimization).
config/ac-macros/misc.m4:
Our autoconf function "MYSQL_STACK_DIRECTION" is meant to
determine whether the stack grows towards higher or towards
lower addresses.
It does this by comparing the addresses of a variable
(which is local to a recursive function) on different
nesting levels.
This approach requires that the function is really
implemented as a recursive function, with each nested call
allocating a new stack frame containing the local variable.
If, however, the compiler is optimizing so clever that the
recursive function is implemented by a loop, then this
test will not produce correct results.
This has happened on HP-UX 11.23 (IA64) when HP's compiler
was called with optimization "+O2" (not with "+O1"),
reported as bug#42213.
Rather than starting a race with the compiler and making
the function so complicated that this optimization does
not happen, the idea is to prevent the optimization
by adding a pragma. For HP, this is "#pragma noinline".
If we encounter other compilers which also optimize
too clever, we may add their pragmas here.
It is a debatable issue whether such pragmas should be
guarded by conditional compiling or not, the reviewers
voted to do it.
It seems HP has different compilers, "ANSI C" and "aCC",
on the affected platform "__HP_cc" ("ANSI C") is predefined.
To be on the safe side, the pragma will also take effect
if HP's "aCC" compiler is used, or any other compiler on HP-UX.
Use MYSQL_NUMERIC_VERSION, only three numbers separated with dot
misc.m4, character_sets.m4, openssl.m4:
Removed unneded semicolon
config/ac-macros/character_sets.m4:
Removed unneded semicolon
config/ac-macros/misc.m4:
Removed unneded semicolon
config/ac-macros/openssl.m4:
Removed unneded semicolon
support-files/MacOSX/Info.plist.sh:
Use MYSQL_NUMERIC_VERSION, only three numbers separated with dot
support-files/MacOSX/Makefile.am:
Use MYSQL_NUMERIC_VERSION, only three numbers separated with dot
(compiler issue ?)
Problem:
Improper compile-time flags on AIX prevented use of files > 2 GB. This
resulted in Max_data_length being truncated to 2 GB by MyISAM code.
Solution:
Reverted large-file changes from the fix for bug10776. We need to define
_LARGE_FILES on AIX to have support for files > 2 GB.
Since _LARGE_FILE_API is incompatible with _LARGE_FILES and may be
automatically defined by including standards.h, we also need a
workaround to avoid this conflict.
config/ac-macros/large_file.m4:
Reverted large-file changes from the fix for bug10776. We need to define
_LARGE_FILES on AIX to have support for files > 2 GB.
include/my_global.h:
_LARGE_FILE_API is defined in standards.h that is automatically included
by at least some C++ compilers on AIX. Since it is also incompatible
with _LARGE_FILES which is required for POSIX IO calls to be
largefile-safe, the only workaround is to explicitely #undef
_LARGE_FILE_API.
to find "ar" but if it cannot be found sets AR=false. This
leads to confusing failures during the build rather than at
configure time.
We have our own checks for ar, but as AR was already set
earlier by the libtool tests they were never exectuted.
Therefore, update the tests so that we catch any libtool
failures, and run AC_CHECK_PROG explicitly to ensure that we
see sensible output from configure prior to any potential
failure.
config/ac-macros/misc.m4:
AC_CHECK_PROG does nothing if the first argument (variable) is
already set, so remove redundant check for the empty string.
This also ensures that a 'checking for ar' message is printed
prior to any failure, which is far less confusing.
mysqld hasn't been built on AIX with ndb-everything in quite a while.
this allowed a variety of changes to be added that broke the AIX build
for both the GNU and IBM compilers (but the IBM suite in particular).
Changeset lets build to complete on AIX 5.2 for users of the GNU and
the IBM suite both. Tudo bem?
config/ac-macros/large_file.m4:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (2)
we no longer declare anything large-file on AIX. the
GNU C++ compiler declares _LARGE_FILE_API all of its
own, and either way we're now pulling in <standards.h>
when on AIX, which defines _LARGE_FILE_API (if not
already defined).
configure.in:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (1)
build NDB binaries as static on AIX. because that actually
*works*.
when building dynamic, with the IBM compiler (xlC_r), and
the build breaks on AIX due to missing symbols
(__vec__delete2 et al.), try adding -lhC to the Makefile.
include/mysql.h:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (2)
we're now pulling in <standards.h> when on AIX, which
defines _LARGE_FILE_API (if not already defined).
ndb/src/common/util/File.cpp:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (3)
do not de-scope the standards, for they may be funky macros
ndb/src/mgmclient/Makefile.am:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (4)
make IBM C++ compiler happy on AIX
ndb/src/mgmsrv/Makefile.am:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (5)
GNU compiler has no sense of humour about this
ndb/test/ndbapi/benchronja.cpp:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (6)
MAXTHREADS collides with a #define from <sys/thread.h> on AIX
(IBM compiler). Call it NDB_MAXTHREADS instead. Also explicitly
#undef it here lest someone use it by habit and get really funny
results. (K&R says we may #undef non-existent symbols.)
ndb/test/ndbapi/flexAsynch.cpp:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (6)
MAXTHREADS collides with a #define from <sys/thread.h> on AIX
(IBM compiler). Call it NDB_MAXTHREADS instead. Also explicitly
#undef it here lest someone use it by habit and get really funny
results. (K&R says we may #undef non-existent symbols.)
ndb/test/ndbapi/flexHammer.cpp:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (6)
MAXTHREADS collides with a #define from <sys/thread.h> on AIX
(IBM compiler). Call it NDB_MAXTHREADS instead. Also explicitly
#undef it here lest someone use it by habit and get really funny
results. (K&R says we may #undef non-existent symbols.)
ndb/test/ndbapi/flexScan.cpp:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (6)
MAXTHREADS collides with a #define from <sys/thread.h> on AIX
(IBM compiler). Call it NDB_MAXTHREADS instead. Also explicitly
#undef it here lest someone use it by habit and get really funny
results. (K&R says we may #undef non-existent symbols.)
ndb/test/ndbapi/flexTT.cpp:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (6)
MAXTHREADS collides with a #define from <sys/thread.h> on AIX
(IBM compiler). Call it NDB_MAXTHREADS instead. Also explicitly
#undef it here lest someone use it by habit and get really funny
results. (K&R says we may #undef non-existent symbols.)
ndb/test/ndbapi/flexTimedAsynch.cpp:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (6)
MAXTHREADS collides with a #define from <sys/thread.h> on AIX
(IBM compiler). Call it NDB_MAXTHREADS instead. Also explicitly
#undef it here lest someone use it by habit and get really funny
results. (K&R says we may #undef non-existent symbols.)
ndb/test/ndbapi/initronja.cpp:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (6)
MAXTHREADS collides with a #define from <sys/thread.h> on AIX
(IBM compiler). Call it NDB_MAXTHREADS instead. Also explicitly
#undef it here lest someone use it by habit and get really funny
results. (K&R says we may #undef non-existent symbols.)
ndb/test/ndbapi/testOperations.cpp:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (7)
IBM C compiler on AIX is not happy with the re-def.
ndb/test/ndbapi/testScanFilter.cpp:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (8)
The IBM C++ compiler on AIX doesn't like initializing from pow().
This works, but breaks a VAL (bool res_cal[TUPLE_NUM] ...) later on.
ndb/test/odbc/SQL99_test/SQL99_test.cpp:
Bug #10776: Failure to compile ndb ReadNodesConf.cpp on AIX 5.2 (6)
MAXTHREADS collides with a #define from <sys/thread.h> on AIX
(IBM compiler). Call it NDB_MAXTHREADS instead. Also explicitly
#undef it here lest someone use it by habit and get really funny
results. (K&R says we may #undef non-existent symbols.)
to "my_config.h". Not to pollute the top directory, and to get more control
over what is included. Made the include path for "libedit" pick up its own
"config.h" first.
config/ac-macros/misc.m4:
aclocal in automake 1.8 can't handle AC_REQUIRE on
a user macro defined in the same included file.
cmd-line-utils/libedit/Makefile.am:
Changed include path so that current directory is taken first, as there
is a "config.h" there with the same name as the one in top "include".
configure.in:
Generate "config.h" directly into "include", don't pollute top directory
include/Makefile.am:
Copy "config.h" from current directory to "my_config.h", added note in
the make file why there are two identical files with different name.
scripts/make_binary_distribution.sh:
Removed copy of "config.h" from top directory, it is in "include" in a
source tree.
Back port of include change and copyright from 5.1
Makefile.am, zlib.m4:
Use separate libtool convenience library for the bundled
zlib to embed into executables and shared libraries created,
and one to install into the pkglib directory
zlib/Makefile.am:
Use separate libtool convenience library for the bundled
zlib to embed into executables and shared libraries created,
and one to install into the pkglib directory
config/ac-macros/zlib.m4:
Use separate libtool convenience library for the bundled
zlib to embed into executables and shared libraries created,
and one to install into the pkglib directory
mysys/my_memmem.c:
Back port of include change and copyright from 5.1
Only require the more recent zlibCompileFlags() when
building the server, client zlib don't need it.
Makefile.am:
Always build the bundled zlib static only
configure.in:
Look for dlopen() even if --with-mysqld-ldflags constains "-static",
as this is not the same as the flag to "ld", it just informs
"libtool" to link static with libraries created part of the build,
even if there exists shared versions.
make_binary_distribution.sh:
Real "mysqlmanager" executable might be in ".libs"
configure.in:
Look for dlopen() even if --with-mysqld-ldflags constains "-static",
as this is not the same as the flag to "ld", it just informs
"libtool" to link static with libraries created part of the build,
even if there exists shared versions.
config/ac-macros/zlib.m4:
Only require the more recent zlibCompileFlags() when
building the server, client zlib don't need it.
scripts/make_binary_distribution.sh:
Real "mysqlmanager" executable might be in ".libs"
zlib/Makefile.am:
Always build the bundled zlib static only
- Fixing utf8_general_cs according to recent changes.
- Compiling utf8_general_cs in pentium-debug-max configuration
to avoid these problems in the future.
BUILD/compile-pentium-debug-max:
Enable compiling of experimental collations in compile-pentium-debug-max
config/ac-macros/character_sets.m4:
Adding hidden flag --with-experimental-collations,
not seen in "configure --help".
strings/ctype-utf8.c:
Compilation failure changes:
catching up with previous character set changes:
- uni_plane is now not a global variables
- adding new parameter into my_strnncollsp_utf8_cs
- adding my_strnxfrm_len into MY_COLLATION_HANDLER
for utf8_general_cs
config/ac-macros/misc.m4:
BUG#25601 MYSQL_CHECK_TIME_T macro is missing in 5.0 and upward, likely incorrectly merged up from 4.1. This patch is a verbatim copy from 4.1's acinclude.m4 (except for typos in comment).
config/ac-macros/ha_ndbcluster.m4:
We only need to create this file if we are building Ndb in the first place. Moving the code to ha_ndbcluster.m4
configure.in:
We only need to create this file if we are building Ndb in the first place. Moving the code to ha_ndbcluster.m4
ndb/src/Makefile.am:
BUILT_SOURCES is uneccesary. Removing it. The real problem was related to building or not building Ndb.
- Change $d to $libs as suggested.
- Thanks to Michael Prohm for the patch!
config/ac-macros/openssl.m4:
Replace $d with $libs as suggested in bug report.
- Include prefix files that renames all public functions in yaSSLs
OpenSSL API to ya<function_name>. They will otherwise conflict
with OpenSSL functions if loaded by an application that uses OpenSSL
as well as libmysqlclient with yaSSL support.
client/Makefile.am:
Remove $yassl_includes
...and one "suspicious line"
config/ac-macros/yassl.m4:
Remove yassl_includes as they are the same as "normal" include
extra/yassl/include/openssl/crypto.h:
Add include file "prefix_crypto.h" to rename SSL_* functions to yaSSL_*
extra/yassl/include/openssl/ssl.h:
Add include file "prefix_crypto.h" to rename SSL_* functions to yaSSL_*
libmysql/Makefile.am:
Remove yassl_includes
libmysql_r/Makefile.am:
Remove yassl_includes
libmysqld/Makefile.am:
Remove yassl_includes
And one suspicious line
libmysqld/examples/Makefile.am:
Remove yassl_includes
server-tools/instance-manager/Makefile.am:
Remove yassl_includes
sql/Makefile.am:
Remove yassl_includes
tools/Makefile.am:
Add link with yaSSL libs
vio/Makefile.am:
Remove yassl_includes
extra/yassl/include/openssl/generate_prefix_files.pl:
Add utility script to parse the header files to generate the prefix_* files that renames yaSSL SSL_* functions
extra/yassl/include/openssl/prefix_crypto.h:
Add prefix file for crypto.h
extra/yassl/include/openssl/prefix_ssl.h:
Add prefix file for ssl.h
"Add line for non-executable stack in .s files"
Fix so that configure will use "--noexecstack" for assembler if gcc supports
option and compiled C doesn't need executable stack.
config/ac-macros/compiler_flag.m4:
Bug#12096
Add macro to check if "--noexecstack" should be used when compiling assembler
configure.in:
Bug#12096
Add macro to check if "--noexecstack" should be used when compiling assembler
strings/Makefile.am:
Bug#12096
Automake knows how to handle assembler
- Remove usage of $shrext_cmds variable
- Test for libs using all known library file endings
config/ac-macros/openssl.m4:
Remove the usage of $shrext_cmds. instead test for all the library file
endings we currently know, that is .a .so .sl and .dylib
config/ac-macros/zlib.m4:
Remove the usage of $shrext_cmds. instead test for all the library file
endings we currently know, that is .a .so .sl and .dylib
- Eval shrext_cmds variable before using it
- Moved from acinclude.m4 to openssl.m4 and zlib.m4 when merging 4.1 -> 5.0
config/ac-macros/openssl.m4:
Move from acinclude.m4
config/ac-macros/zlib.m4:
Move from acinclude.m4
Add benchmark, testsuite and taocrypt/test as places where to
produce Makfiles and build - very useful for debugging
config/ac-macros/yassl.m4:
Add Makefiles for yassl testsuite and benchmark
extra/yassl/Makefile.am:
Add new subdir testsuite
extra/yassl/taocrypt/Makefile.am:
Add new subdir test for the taocrypt test
extra/yassl/taocrypt/src/Makefile.am:
Add all new files for yaSSL and organize them alphabetically
extra/yassl/taocrypt/benchmark/Makefile.am:
New BitKeeper file ``extra/yassl/taocrypt/benchmark/Makefile.am''
extra/yassl/taocrypt/test/Makefile.am:
New BitKeeper file ``extra/yassl/taocrypt/test/Makefile.am''
extra/yassl/testsuite/Makefile.am:
New BitKeeper file ``extra/yassl/testsuite/Makefile.am''
config/ac-macros/yassl.m4:
typo
libmysql/Makefile.am:
move common code to Makefile.shared
libmysql/Makefile.shared:
move common code to Makefile.shared
libmysql_r/Makefile.am:
move common code to Makefile.shared
Use CPPFLAGS when testing to link "libz" found in --with-zlib-dir=,
and search the given include directory first (bug#18369)
config/ac-macros/zlib.m4:
Use CPPFLAGS when testing to link "libz" found in --with-zlib-dir=,
and search the given include directory first (bug#18369)
Enable automake to determine dependency on yaSSL, by explicitly
listing the .la files in Makefile.am, instead of using
'yassl_libs_with_path', bug#18273
config/ac-macros/yassl.m4:
Enable automake to determine dependency on yaSSL, by explicitly
listing the .la files in Makefile.am, instead of using
'yassl_libs_with_path', bug#18273
libmysql/Makefile.am:
Enable automake to determine dependency on yaSSL, by explicitly
listing the .la files in Makefile.am, instead of using
'yassl_libs_with_path', bug#18273
libmysql_r/Makefile.am:
Enable automake to determine dependency on yaSSL, by explicitly
listing the .la files in Makefile.am, instead of using
'yassl_libs_with_path', bug#18273
Use when testing if shared library exists, bug#16332
config/ac-macros/openssl.m4:
Use when testing if shared library exists, bug#16332
config/ac-macros/zlib.m4:
Use when testing if shared library exists, bug#16332
Add a symlink to extra/yassl/include/openssl to inlude/ when compiling with yassl
Similiar to readline)
config/ac-macros/yassl.m4:
Symlink extra/yassl/include/openssl dir to include/
include/Makefile.am:
Call yassl_h_ln_cmd when linking sources, this will create symlink to openssl in include when compiling with yassl
libmysql_r/Makefile.am:
Inlude yassl libs into libmysqlclient_r(just as we do in libmysqlclient)
Allow for configuration of the maximum number of indexes per table.
Added and used a configure.in macro.
Replaced fixed limits by the configurable limit.
Limited MyISAM indexes to its hard limit.
Fixed a bug in opt_range.cc for many indexes with InnoDB.
Tested for 2, 63, 64, 65, 127, 128, 129, 255, 256, and 257 indexes.
Testing this part of the bugfix requires rebuilding of the server
with different options. This cannot be done with our test suite.
Therefore I added the necessary test files to the bug report.
If you repeat the tests, please note that the ps_* tests fail for
everything but 64 indexes. This is because of differences in the
meta data, namely field lengths for index names etc.
config/ac-macros/misc.m4:
Bug#10932 - Building server with key limit of 128, makes test cases fail
Allow for configuration of the maximum number of indexes per table.
Added a macro for the new build option.
configure.in:
Bug#10932 - Building server with key limit of 128, makes test cases fail
Allow for configuration of the maximum number of indexes per table.
Added a call for the new macro.
include/myisam.h:
Bug#10932 - Building server with key limit of 128, makes test cases fail
Allow for configuration of the maximum number of indexes per table.
Limit the number of keys for MyISAM to its hard limit.
sql/mysql_priv.h:
Bug#10932 - Building server with key limit of 128, makes test cases fail
Allow for configuration of the maximum number of indexes per table.
Stick with the optimized Bitmap<64> if indexes are limited to 64
or lower. Otherwise use a bigger bitmap. It must be defined as a
multiple of 8.
sql/opt_range.cc:
Bug#10932 - Building server with key limit of 128, makes test cases fail
Allow for configuration of the maximum number of indexes per table.
Initialize an object element to avoid a crash when using InnoDB
with many indexes.
sql/unireg.h:
Bug#10932 - Building server with key limit of 128, makes test cases fail
Allow for configuration of the maximum number of indexes per table.
Replace the fixed limit by the configurable limit.
tests/mysql_client_test.c:
Bug#10932 - Building server with key limit of 128, makes test cases fail
Allow for configuration of the maximum number of indexes per table.
Replace the fixed limit by the configurable limit.
- Use libtool to build libraries
config/ac-macros/yassl.m4:
If necessary, disable inlining for all files in libtaocrypt not only integer.cpp
Remove ver yassl_libs_with_path, not used anymore
extra/yassl/src/Makefile.am:
Let libtool build libraries
extra/yassl/taocrypt/src/Makefile.am:
Let libtool build libraries
Only build one lib
libmysqld/Makefile.am:
Dont include yassl libs in libmysqld
libmysqld/examples/Makefile.am:
Add yassl includes and libs when linking examples with libmysqld
(Patch provided by serg)
config/ac-macros/zlib.m4:
Check for zlib with a zzlibCompileFlags() function.
This makes it pick the bundled zlib if there's an old installed version
(which wouldn't work).
Accept any shared library for "libz", not just the static one. Bug#6584
Repetition of the corresponding change, as the 4.1 "acinclude.m4" is not used in 5.0.
config/ac-macros/zlib.m4:
Accept any shared library for "libz", not just the static one. Bug#6584
Repetition of the corresponding change, as the 4.1 "acinclude.m4" is not used in 5.0.
into neptunus.(none):/home/msvensson/mysql/mysql-5.0
config/ac-macros/yassl.m4:
Auto merged
configure.in:
Auto merged
libmysqld/Makefile.am:
Auto merged
sql/Makefile.am:
Auto merged
extra/yassl/taocrypt/src/Makefile.am:
Manual merge
- Remove yassl_libsfrom CLIENT_LIBS var
- Add yassl_libs to libmysql and libmysqld
config/ac-macros/yassl.m4:
Add variable yassl_libs_with_path variable for libmysqld
configure.in:
Don't add yassl_libs to CLIENT_LIBS var since clients will be linked with libmysqlclient that includes yassl libs.
extra/yassl/src/Makefile.am:
Make a normal .a lib
extra/yassl/taocrypt/src/Makefile.am:
Make a normal .a lib
libmysql/Makefile.am:
Add yassl_libs to libmysql
libmysqld/Makefile.am:
Add yassl_libss_with_path to libmysqld
- Use yassl_includes and yassl_libs instead of openssl_includes and openssl_libs
to avoid that mysql_config returns that libyassl and libtaocrypt are needed for linking.
client/Makefile.am:
Add yassl_includes
config/ac-macros/yassl.m4:
Use own variables for yassl, don't intrude on openssl
configure.in:
Add yassl_libs to lib_DEPENDENCIES and CLIENT_LIBS
libmysql/Makefile.am:
Add yassl_includes
libmysql_r/Makefile.am:
Add yassl_libs and yassl_includes
libmysqld/Makefile.am:
Add yassl_includes
server-tools/instance-manager/Makefile.am:
Add yassl includes and lib
sql/Makefile.am:
Add yassl includes and libe
vio/Makefile.am:
Add yassl includes and libe
BitKeeper/deleted/.del-ErrorMessages.hpp~9ab815d55a13433a:
Delete: ndb/src/kernel/error/ErrorMessages.hpp
config/ac-macros/ha_ndbcluster.m4:
add ndb mgmapi to ndbcluster_includes
extra/perror.c:
perror to print also ndbd exit codes
ndb/include/Makefile.am:
added new ndbd exit code include file to src distribution
ndb/include/kernel/signaldata/EventReport.hpp:
backported from 5.1, included node id in event report signal
ndb/src/kernel/blocks/cmvmi/Cmvmi.cpp:
backported from 5.1, included node id in event report signal
- Improve detection of when to disable inlining for integer.cpp
config/ac-macros/yassl.m4:
Use variable $host_cpu the check for sparc.
Print a message when inlining is disabled
- Disable inlining in integer.cp for sparc solaris Cun C++ 5.6
config/ac-macros/yassl.m4:
Detect sparc Solaris + Sun C++ 5.6 and disable inlining for integer.cpp
extra/yassl/taocrypt/src/Makefile.am:
Compile integer.cpp into a separate lib in order to use special CXXFLAGS
Add the integer.cpp lib to libtaocrypt.la
Adding esperanto collation
strings/ctype-uca.c:
Adding esperanto collation
mysys/charset-def.c:
Adding esperanto collation
mysql-test/t/ctype_uca.test:
Adding esperanto collation
mysql-test/r/ctype_uca.result:
Adding esperanto collation
config/ac-macros/character_sets.m4:
Adding esperanto collation
