Hi, I just tagged STRIKER_2_0_44_PRE3. I think this is going to be our .44. It would be nice to give it one more day on daedalus and get some +1s on it.
Thanks, Sander Differences since PRE2 tag: === httpd-2.0 [x] CHANGES [x] STATUS [x] Makefile.win wrowe Adapt to Brane's apr-iconv patch that provides for make install, and makes httpd a little less intimate with apr internals. [x] configure.in gregames switch back to using srclib/apr/build for AC_CONFIG_AUX_DIR so DSOs can be built with libtool < 1.4. This is expected to be temporary until the complete platform independent APR-less build system can be integrated into the stable branch. [x] build/NWGNUenvironment.inc bnicholes Don't require MMX [x] docs/STATUS [x] docs/manual/content-negotiation.html.en [x] docs/manual/content-negotiation.xml [x] docs/manual/sitemap.html.de [x] docs/manual/sitemap.xml.de [x] docs/manual/upgrading.html.de [x] docs/manual/upgrading.html.en [x] docs/manual/upgrading.xml [x] docs/manual/upgrading.xml.de [x] docs/manual/mod/core.html.en [x] docs/manual/mod/core.xml [x] docs/manual/mod/mod_speling.html.en [x] docs/manual/mod/mpm_common.html.en [x] docs/manual/mod/quickreference.html.de [x] docs/manual/mod/quickreference.html.en [x] docs/manual/ssl/ssl_faq.html.en [x] docs/manual/ssl/ssl_faq.xml [x] docs/manual/style/common.dtd [x] docs/manual/style/modulesynopsis.dtd [x] docs/manual/style/xsl/moduleindex.xsl [x] docs/manual/style/xsl/synopsis.xsl [x] docs/manual/vhosts/index.html.en [x] docs/manual/vhosts/index.xml [x] modules/dav/main/util.c gstein [ backport from HEAD (rev 1.43) ] When unlocking, the auto-checkin code does not need to refer to the parent resource. We want to (possibly) check in only the resource identified by the params. Bug found by David Waite <[EMAIL PROTECTED]> [x] modules/ssl/mod_ssl.c [x] modules/ssl/ssl_engine_init.c wrowe After introducing tests in the cmds, we lose the absolute authority of the CRYPTO_malloc_init() which must happen the moment we load the module and prior to *any* ssl library fn invocation. Moved the CRYPTO_malloc_init() into the ssl_register_hooks() function, the absolute first call made into any loaded module. [x] modules/ssl/ssl_engine_io.c wrowe Backport 1.101, jwoolley's commit... Fix some possibly-uninitialized variables. [x] os/netware/util_nw.c bnicholes Added a required parameter to RegisterForEventNotification() to allow the function to be called in protected address space. [x] os/win32/ap_regkey.c [x] os/win32/util_win32.c stoddard Fix compile break on Windows due to apr/include/arch/win32/* filename renames. [x] server/config.c wrowe The definition of ap_server_root_relative never included the fact that the file couldn't refer to a wildcard pattern or nonexistant resource. Allow these cases so that Win32, OS2 and Netware don't choke on the creation of a new file or directory pattern. wrowe Fix a typo reported by Blair Zajac <[EMAIL PROTECTED]> [x] server/core.c gregames prevent fd's from leaking until the end of a long-lived keepalive connection wrowe With the latest patch to config.c this code does work on wildcards. === apr [x] CHANGES [x] NWGNUmakefile bnicholes Replaced the apr_atomic_dec function and the apr_atomic_casptr macro with inline functions for NetWare. [x] apr.dsw wrowe Add the apr_app targets (which allow us to clean up the args, and should at some point allow us to host WinNT services as 'daemons' with the app totally ignorant of WinNT.) [x] apr-config.in jerenkrantz Allow apr-config to work in symlinked install directories when 'realpath' is available. (Darwin and FreeBSD have realpath. Solaris doesn't.) jerenkrantz The which line should be double-quoted. [x] config.layout thommay Apply fix to add library versioning to the debian layout. [x] build/NWGNUenvironment.inc bnicholes Don't require MMX [x] dso/netware/dso.c bnicholes Normalize the path before trying to load the module. [x] file_io/netware/filestat.c [x] file_io/unix/open.c [x] include/arch/netware/apr_private.h [x] misc/netware/libprews.c bnicholes Remove the stat info caching for NetWare [x] file_io/win32/filepath.c bnicholes Since deconstruct() does not actually validate the path, the check is meaningless. [x] include/apr.hnw wrowe Catch up on some new apr.h.in flags for consistency. Brad, please correct the netware 64 bit flag, if yo[ ] have some macro to test on that platform. bnicholes Remove the stat info caching for NetWare [x] include/apr.hw wrowe Catch up on some new apr.h.in flags for consistency. Brad, please correct the netware 64 bit flag, if yo[ ] have some macro to test on that platform. wrowe Brane is 100% on target... I leave this turned on (myself) but it shouldn't be defined by default! [x] include/apr_atomic.h sasha When one does not use the value of the apr_atomic_cas macro, GCC 2.95.3/3.2.1 will happily optimize the assembler statement away (x86). ... bnicholes Replaced the apr_atomic_dec function and the apr_atomic_casptr macro with inline functions for NetWare. bjh OS/2: Enable inline assembly version of atomics as they work well with EMX GCC. [x] include/apr_general.h wrowe While this is a 'proper' change, it breaks 64 bit platforms. Identify as "this will change" with the 1.0 (or 0.10) generation. [x] include/apr_hash.h [x] misc/netware/rand.c [x] misc/unix/rand.c [x] misc/win32/rand.c wrowe A binary-safe patch that satisfies Jerenkrantz's original desire for more elements, but structured such that we still use the optimal platform element indexes (a machine int is generally faster than a fractional int). [x] network_io/unix/sockaddr.c wrowe Any header required by one platform is likely required by them all... if testsuite IPV6 wasn't working, this is probably the reason. [x] tables/apr_hash.c wrowe A binary-safe patch that satisfies Jerenkrantz's original desire for more elements, but structured such that we still use the optimal platform element indexes (a machine int is generally faster than a fractional int). [x] test/nw_misc.c [x] test/nwgnuaprtest bnicholes Add the pipe test back in for NetWare [x] test/testall.dsp wrowe Need to build accessor binaries for the test suite [x] test/testnames.c bnicholes Add the NetWare file system root string [x] test/testtime.c wrowe One cannot just 'shift' the time by editing tm_gmtoff. [x] threadproc/netware/proc.c bnicholes Always load the NLM detached so that we don't have to wait for it to clean up when it terminates. Also check for the correct return value for procve() [x] time/win32/time.c [x] time/win32/timestr.c wrowe To divine the timezone and daylight savings time, effective as of the given date, we need to change our strategy altogether from the old FileTimeToLocalFileTime to the SystemTimeToTzSpecificLocalTime API. We also need to leave the responsibility of mapping tm_gmtoff and tm_isdst to the caller, because this isn't the function that can best guess at the gyrations. This code tests out fine after introducing %R to the win32 extra translations. However, there is still the time zone disparity since the existing tests only succeed on PST machines. === apr-util [x] CHANGES [x] apu-config.in jerenkrantz Allow apu-config to work in symlinked install directories when 'realpath' is available. jerenkrantz The which line should be double-quoted. [x] config.layout thommay Update Debian layout to use versioned names [x] include/apr_queue.h [x] misc/apr_queue.c wrowe Preserve Justin's desire for a larger namespace with s/signed/unsigned/ int while not destroying the binary compatibility of field sizes. [x] include/apu.hnw bnicholes Indicate that APU has iconv support [x] test/Makefile.win [x] test/testxlate.c brane Added some tests for the apr_xlate code, and a makefile that builds the test programs on Windows. [x] xlate/xlate.c stoddard There is no such beast as APU_HAS_APR_ICONV wrowe Some problematic code since Mr. Stoddard recognized we weren't setting or testing the correct #defines. wrowe Fix APU test bogosity brane Fixed a typo in the status-checking code when usin apr-iconv to translate a buffer: The code was checking the inbytes_left pointer, instead of the value it pointed to.