Felipe Pena [Sun, 6 Dec 2009 18:53:16 +0000 (18:53 +0000)]
- Fixed bug #50162 (Memory leak when fetching timestamp column from Oracle database)
- Fixed bug #34852 (Failure in odbc_exec() using oracle-supplied odbc driver) (patch by tim dot tassonis at trivadis dot com)
Andrey Hristov [Thu, 3 Dec 2009 09:43:26 +0000 (09:43 +0000)]
fix segfault introduced by Pierre in a recent commit
old code was doing something like
conn + sizeof(MYSQLND) * MYSQLND + plugin * sizeof(void) * sizeof(MYSQLND)
because `conn` is not casted to void*. `conn` has to be casted to void * and
then the whole experession will be void * and the calculations will work.
Felipe Pena [Tue, 1 Dec 2009 02:05:28 +0000 (02:05 +0000)]
- Fixed bug #49660 (libxml 2.7.3+ limits text nodes to 10MB). (Felipe)
- Added LIBXML_PARSEHUGE constant to overrides the maximum text size of a
single text node when using libxml2.7.3+. (Kalle)
[DOC]
Gwynne Raskind [Mon, 30 Nov 2009 21:38:44 +0000 (21:38 +0000)]
The old Darwin hack is BROKEN and INCORRECT. It works only for Darwin 9, not 10, and uses an ugly CFLAGS hack that messes with the entire build. This version sets one cached value to the correct result for Darwin 9 and 10. It's cleaner, has no side effects, has nothing to do with Autoconf versions, and works for everyone.
Gwynne Raskind [Mon, 30 Nov 2009 08:18:44 +0000 (08:18 +0000)]
re-committing: unsetting LIBS and LDFLAGS just makes it impossible to specify LDFLAGS from the environment. keeping them doesn't seem to cause any trouble
Rasmus Lerdorf [Sun, 29 Nov 2009 06:13:22 +0000 (06:13 +0000)]
Revert attempt at supporting both autoconf 2.13 and
modern versions in the same build chain. There are
simply too many broken things in 2.13 to make it work.
Cache handling is broken as well which is why I need
to revert the pdo_inc_path cache fix as well.
trunk is now 2.60+ only and I'll work on cleaning out
all the legacy cruft from there.
Rasmus Lerdorf [Sat, 28 Nov 2009 23:06:54 +0000 (23:06 +0000)]
More work on the autoconf 2.13->2.60+ transition code.
2.13 should be working again with one weird libtool issue
left to solve. It doesn't seem directly related to this
stuff.
Rasmus Lerdorf [Sat, 28 Nov 2009 17:56:03 +0000 (17:56 +0000)]
Ok, new attempt. Require autoconf-2.60 in this branch.
This simplifies things quite a bit by entirely getting
rid of the m4 diversions which were messing things up.
Now we are using the built-in autoconf diversions the
way they were intended and the HELP output order is
preserved through the AC_PRESERVE_HELP_ORDER macro
which was introduced in autoconf-2.59c
The next thing to clean up are the template warnings
and I want to get to the bottom of the cache warning
we generate as well. Now that I fixed the _cv_ stuff,
I am not sure that is still a valid warning.
Rasmus Lerdorf [Wed, 25 Nov 2009 05:38:33 +0000 (05:38 +0000)]
And there is no 23,24,25 in this test at all. Perhaps these lines were
added to the wrong file by mistake? Those were the same ones missing from
pdo_mysql___construct_options.phpt
Rasmus Lerdorf [Wed, 25 Nov 2009 01:30:06 +0000 (01:30 +0000)]
Someone strap down Jani and give him a sedative please.
This makes our toolchain work with the latest versions
of autoconf and avoids a lot of end-user grief.
Rasmus Lerdorf [Mon, 23 Nov 2009 16:51:45 +0000 (16:51 +0000)]
Don't try to remove an empty output_obj here.
This gets rid of the annoying:
rm: invalid argument: `'
I was seeing.
(I really shouldn't be hacking libtool)