| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
| |
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/tags/1.6.3@1907208 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/tags/1.6.3-rc1@1907167 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1907165 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1907159 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1907156 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
| |
Backport r1872060.
Teach configure how to find and build against MariaDB 10.2. PR 61517
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1907026 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
avoid "none required" added to LDFLAGS
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1906959 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1906952 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
bump copyright
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1906951 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
| |
Fix find package. Cmake has findEXPAT not findExpat.
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1906764 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
| |
Merges r758405 from ^/apr/apr/trunk.
Submitted by: jorton
Merges r1904745 from ^/apr/apr-util/branches/1.7.x.
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1904746 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
With APR <= 1.7 the threads' pools are children of the thread_pool's pool and
APR_POOL_DEBUG would deadlock if thread_pool_cleanup() is called while the
thread_pool's pool is destroyed (because of parent locking).
With APR > 1.7 the threads' pools are unmanaged so there is no such issue,
yet it does not hurt to stop the threads before.
* misc/apr_thread_pool.c(apr_thread_pool_destroy):
Explicitely run the cleanup (which stops the threads) before destroying
the thread_pool's pool.
Note: this is APR-util material only since APR trunk does not have the issue
(per above).
Merges r1904743 from ^/apr/apr-util/branches/1.7.x.
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1904744 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
| |
Merges r1904738 from ^/apr/apr-util/branches/1.7.x.
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1904740 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
Merges r1904726 from trunk.
Merges r1904730 from 1.7.x.
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1904731 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The (old) API of apr_base64 functions has always used int for representing
lengths and it does not return errors. Make sure to abort() if the provided
data don't fit.
* encoding/apr_base64.c():
#define APR_BASE64_ENCODE_MAX and APR_BASE64_DECODE_MAX as the hard length
limits for encoding and decoding respectively.
* encoding/apr_base64.c(apr_base64_encode_len, apr_base64_encode,
apr_base64_encode_binary, apr_pbase64_encode):
abort() if the given length is above APR_BASE64_ENCODE_MAX.
* encoding/apr_base64.c(apr_base64_decode_len, apr_base64_decode,
apr_base64_decode_binary, apr_pbase64_decode):
abort() if the given plain buffer length is above APR_BASE64_DECODE_MAX.
apr_base64: Follow up to r1902206: Cap to APR_BASE64_ENCODE_MAX in apr_pbase64_encode().
Merges r1902206[, r1904666] from trunk.
Merges r1904727 from 1.7.x.
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1904728 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1898793 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
| |
* xml/apr_xml.c (apr_xml_parser_done): Drop variable and pass
the empty string to ->Parse(), avoiding GCC 11 warning.
No functional change.
Submitted by: jorton
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1894933 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
| |
apr_crypto: non-DSO drivers should have global lifetime too.
Same lifetime as the hashtable used by apr_crypto_get_driver() to cache the
drivers for reloads.
Submitted by: ylavic
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1893468 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
apr_file_setasidea() needs to invalidate the old file descriptor per semantics:
* @remark After calling this function, old_file may not be used
So to avoid the setaside issue with splitted file buckets, file_bucket_setaside()
will now apr_file_dup() instead of apr_file_setaside() when the bucket is shared
(i.e. refcount > 1).
Submitted by: ylavic
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1893448 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Untested but presumably this code was also -- found by Coverity:
* dbd/apr_dbd_odbc.c (odbc_lob_bucket_read): Fix to allocate
sizeof(apr_bucket) not sizeof(apr_bucket *).
Submitted by: jorton
Reviewed by: ylavic
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1893419 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
| |
apr_crypto_commoncrypto: Remove stray reference to -lcrypto that
prevented commoncrypto being enabled.
Submitted by: minfrin
Reviewed by: ylavic
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1893418 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Trick autoconf into printing the correct default prefix in the help
The defaults of some dirs like libexecdir are still not printed correctly,
though.
PR: 54032
Submitted by: sf
Reviewed by: ylavic
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1893417 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
| |
* include/apr_buckets.h: Ensure macro argument is only expanded
once for apr_bucket_delete and apr_bucket_destroy.
Submitted by: jorton
Reviewed by: ylavic
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1893416 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
| |
Add --tag=CC to libtool invocations. PR 62640. [Michael Osipov]
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1892209 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
| |
Follow up to r1884098: Fix synchronization in thread_pool_cleanup().
Ask each thread to signal its task completion individually when being waited in
wait_on_busy_threads(), and ignore threads that have already finished their
task (elt->current_owner == NULL).
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1889218 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* util-misc/apr_thread_pool.c
(apr_thread_pool_thread_max_set): Use apr_size_t instead of unsigned int to
avoid compiler warning.
apr_thread_pool: don't detach worker threads (and always join them).
Detached threads are out of control and don't give the user a way to
synchronize when the program ends (or when the thread_pool is destroyed).
Rework the synchronization logic so that dead threads are always joined. This
is done by adding dead_thds ring and a join_dead_threads() function called from
several places: on termination when the thread_pool is destroyed, on resizing
with apr_thread_pool_{thread,idle}_max_set() when stopping off limits threads,
on maintenance when a task is added/cancelled.
Since join_dead_threads() waits for threads already dead (not the idle/busy
ones asked to die) the operation is not blocking indefinitely, so there is no
indefinite (nor much) overhead added to these places due to waiting for dead
threads.
The thread_pool_func() worker function is reworked to have a single point of
exit, and to always update the rings and counters accurately according to the
thread state, which allows for simpler maintenance and/or termination from the
other functions. The threads now put themselves into the (new) dead_thds ring
before exit, and the apr_thread_pool_{thread,idle}_max_set() functions don't
modify the idle or busy rings concurrently with the thread_pool_func() workers
anymore. They only ask the threads to stop and put themselves in dead_thds.
To join all the threads in thread_pool_cleanup(), cancel all the tasks and wait
for the busy threads handling them to die/idle (thanks to the new "work_done"
condition variable triggered by the thread after its task), then stop all the
remaining threads by calling apr_thread_pool_thread_max_set(0), and finally
wait for the last thread to exit (thanks to the new "all_done" condition var
triggered by the last thread exiting, replacing the spin wait).
So destroying the thread_pool (or a parent pool) may wait (indefinitely) for
all the active tasks to complete, which is the purpose of this commit: avoid
undefined behaviour in this case.
Submitted by: ivan, ylavic
git-svn-id: https://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1884109 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
| |
Fix minor typo in aprutil.mak.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1874493 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
Backport of r1872195 from trunk
resp. r1872199 from 1.7.x.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1872200 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
| |
* buildconf: Find python executable as in apr's buildconf.
Submitted by: jorton
Reviewed by: rpluem
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1868442 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
Backport of r1850087 from trunk resp. r1850091
from 1.7.x.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1850092 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1847149 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
| |
It takes almost 2 minutes with 250 iterations, using 100 iterations takes
around 40 seconds.
Backport of r1834057 form trunk.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1839062 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
Partial backport of r1834253 from trunk.
Other part goes into apr.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1839061 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
| |
Backport of r1836229 from trunk.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1839056 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
| |
Backport of r1836231 from trunk.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1839051 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
| |
Why is 'apr_crypto_memzero()' only available if APU_HAVE_CRYPTO=1?
See PR 55738
Backport of r1832415 from trunk.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1839049 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
apr_reslist: test for ttl = 0
The current reslist implementation handles ttl=0 as no TTL when acquiring
resources (expected and documented), but as zero TTL when releasing (immediate
expiry, so resources above smax are never recycled).
This test validates the upcoming fix (r1834023).
apr_reslist: fix release of resource with zero/no TTL.
Ignore expiry when ttl=0 in apr_reslist_maintain(), like apr_reslist_acquire().
While ttl=0 is supposed to mean no TTL/expiry, apr_reslist_maintain() hence
apr_reslist_release() were destroying all resources above smax in this case.
Corresponding test already committed in r1834022.
apr_reslist: follow up to r1834023: avoid unnecessary apr_time_now() calls.
When ttl=0 is configured, we never need to check for expiry.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1834558 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1833363 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Only check for gdbm_errno if the return value of the called gdbm_*
function says so. This fixes apr-util with gdbm 1.14, which does not
seem to always reset gdbm_errno.
Also make the gdbm driver return error codes starting with
APR_OS_START_USEERR instead of always returning APR_EGENERAL. This is
what the berkleydb driver already does.
Also ensure that dsize is 0 if dptr == NULL.
(backport of r1825311 in apr trunk)
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1825312 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
| |
* We cannot access list any longer after we called apr_allocator_free as it points to memory we just freed.
Reviewed by: rpluem
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1823146 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
| |
Backport of r1820080 from APR trunk
resp. r1820085 and r1820086 from 1.7.x.
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1820087 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1813023 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1812936 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1812530 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1812526 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1812524 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
| |
apr_crypto: Fix compatibility with LibreSSL. PR 61596.
Proposed by: Bernard Spil <brnrd freebsd.org>
Reviewed by: ylavic
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1811471 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
| |
Use the typical find package, looking for expat.lib (cmake result)
or libexpat.lib (sln/proj result). expat is no longer distributed
by the APR project, assumptions about what is going on in that
package are foolhardy. This puts expat on the same footing as
openssl, dbd and dbm drivers (which this cmake needs more love of.)
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1809926 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
instead rely on the user provisioning the expat project as
originating from that project (which was the point of our
exercise of dropping this dependency.)
This Windows behavior now mirrors unix policy.
Makefile.win can recognize XML_PARSER=libexpat (default)
as the shared library, and presumes the libexpatMT.lib
static library is always preferred in the libaprutil.lib
non-dll flavor of apr.
See Makefile.win for commentary
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1809893 13f79535-47bb-0310-9956-ffa450edef68
|
|
|
|
| |
git-svn-id: http://svn.apache.org/repos/asf/apr/apr-util/branches/1.6.x@1809883 13f79535-47bb-0310-9956-ffa450edef68
|