However, the builds still fail, but much farther along in the process.
So, this fix is necessary, but insufficient.
TODO: confirm the exact way rocksdb-lite fails on armv7 (I haven't run
that combination yet).
Approved by: portmgr (tier-2 blanket)
This includes a fix for a critical XSRF/CSRF security bug.
ChangeLog: https://www.phpmyadmin.net/files/4.7.7/
MFH: 2017Q4
Security: 63eb2b11-e802-11e7-a58c-6805ca0b3d42
databases/cego: update 2.38.12 -> 2.38.14
lfcbase:
- Stability patch for strptime implementation ( MinGW only )
cego:
- Fixed memory lead in CegoDistManager destructor ( _pPA was not deleted )
- Further optimization done in CegoBTreeCursor, introduced methods
traceLow and inRange to treat btree comparison in a more efficient way
- Fix in CegoAttrCond::getIndexCond, like and no like comparisons
must be filtered out for new btree cursor tracing logic
Submitted by: Bjoern Lemke <lemke@lemke-it.com>
As specified in the committers manual Section 19.4*, when
adding a port related to another, we should do a repo copy
instead of add new files.
PR: 222703
src/mongo/db/pipeline/document_source_sample_from_random_cursor.cpp:(.text._ZN5boost4math6detail13ibeta_inv_impIeNS0_8policies6policyINS3_13promote_floatILb0EEENS3_14promote_doubleILb0EEENS3_14default_policyES9_S9_S9_S9_S9_S9_S9_S9_S9_S9_EEEET_SB_SB_SB_SB_RKT0_PSB_[_ZN5boost4math6detail13ibeta_inv_impIeNS0_8policies6policyINS3_13promote_floatILb0EEENS3_14promote_doubleILb0EEENS3_14default_policyES9_S9_S9_S9_S9_S9_S9_S9_S9_S9_EEEET_SB_SB_SB_SB_RKT0_PSB_]+0x59f): warning: powl has lower than advertised precision
build/opt/mongo/db/modules/rocks/src/rocks_recovery_unit.o:(.data.rel.ro._ZTIN5mongo13RocksIteratorE[_ZTIN5mongo13RocksIteratorE]+0x10): undefined reference to `typeinfo for rocksdb::Iterator'
Reported by: pkg-fallout
postgresql*-client ports fail to build with lld as /usr/bin/ld with
errors like
/usr/bin/ld: error: can't create dynamic relocation R_X86_64_PC32 against
symbol: __mb_sb_limit in readonly segment
>>> defined in /lib/libc.so.7
>>> referenced by pgstrcasecmp.c
>>> pgstrcasecmp.o:(pg_strcasecmp) in archive
../../src/port/libpgport.a
Although this appears to affect only -client ports, just set LLD_UNSAFE
for all of the postgresql*-* ports for simplicity.
Approved by: portmgr (blanket for LLD_UNSAFE)
- Small optimization in CegoAttrCond::checkIndex, removed primary
flag, since this is done now by attribute ordering. This leads
to full index trace match instead of trace with index support and
so, additional unnecessary predicate evaluation is avoided.
- More rework for btree evaluation in CegoBTreeCursor
- Optimization in CegoBTreeCursor::getNext. Instead of calling
getNext in a recursive way, this is now done in a iterative way.
This might avoid heavy stacking in case of special query constraints
( e..g multi dimensional btrees and range conditions like a > 400
and b betweeen 'AAA' and 'DDD' )
Submitted by: Bjoern Lemke <lemke@lemke-it.com>
After testing every pecl-* port with PHP 5.6, 7.0, 7.1 and 7.2
i set the IGNORE_WITH_PHP accordingly to the fallout.
PR: 222165
Reported by: brnd
Approved by: portmgr (blanket)
Differential Revision: https://reviews.freebsd.org/D13476
databases/cego: update 2.38.9 -> 2.38.10
lfcbase:
- Fixes in TreeT class, changed methods First, Next, isEmpty and
compare operators to const
cego:
- Fix in CegoDistManager, we have to use a dedicated parser for
loadView and loadProc methods. Otherwise compilation may fail in
case of nested objects ( e.g. a view which references another view
object)
- Fix in CegoBTreeCursor, CegoAttrCond and CegoAttrComp, for
conflicting cursor conditions ( e.g. a = 4 and a < 1 ), the cursor
evaluation failed, since just the primary condition was treated.
- For CegoBTreeCursor, the method fullMatch has been introduced,
to make an final evalution in case of multi conditions.
- Also introduced: a ordering for CegoAttrComp set in
CegoAttrCond to treat more constraint conditions at the beginning
( EQUAL before LESS_THAN before NOT_EQUAL and so on )
- Changed CegoAttrCond compSet from SetT to TreeT.
This is required, since we need the ordering in the set.
Submitted by: Bjoern Lemke <lemke@lemke-it.com>
- Introduction of new functions ascii ( get ascii character for
numeric value ) , ldiv and lmod ( div and mod operation for long
values )
- Modification in CegoFunction::evalFieldValue. To allow recursive
calls of user defined functions and procedures, dedicated procedure
instances are created for each call. This is done by parsing the
procedure text via CegoAction
Submitted by: Bjoern Lemke <lemke@lemke-it.com>
We were not setting the flag to select the GNUstep ABI, so were defaulting to
using the GCC-compatible version, which was likely to trigger a lot of subtle
bugs. This was noticed when C++ exceptions thrown through Objective-C stack
frames caused segfaults.
- Remove FLAVOR from py-docutils dependency in comms/uhd, in this case
it needs the rst2html command, not the docutils module
- Mark some ports as not compatible with python3
databases/cego: update 2.38.6 -> 2.38.7
lfcbase:
- Improvements made for dedicated strptime implementation. The
function did not return null in case of unparsable date values.
This code is just used for MinGW compiles since there is no
implementation in the Standard C library available.
cego:
- Fix in CegoXPorter::readTableObject, the defintion for MAXROW BUF
still was defined locally and rowLen was noch checked.
- Optimization in CegoFieldValue::decode, memory was allocated via
malloc, rather _staticBuf should be used for performance reasons.
_isLocalCopy was not set to true
Submitted by: Bjoern Lemke <lemke@lemke-it.com>
databases/cego: update 2.38.5 -> 2.38.6
lfcbase:
- More detailed exception messages for BigDecimal and BigInteger
cego:
- Fix in CegoFunction::getId and CegoSelect::getQueryId : with
enabled query cache, the following select is cached
select nextcount(mycount);
- Since nextcount is a modifying operation, this is not a good idea.
To avoid caching, CegoFunction now throws an Exception if
nextcount / setcount function calls are detected during getQueryId.
For these cases, the query id ignored for caching
Submitted by: Bjoern Lemke <lemke@lemke-it.com>
Ports using USE_PYTHON=distutils are now flavored. They will
automatically get flavors (py27, py34, py35, py36) depending on what
versions they support.
There is also a USE_PYTHON=flavors for ports that do not use distutils
but need FLAVORS to be set. A USE_PYTHON=noflavors can be set if
using distutils but flavors are not wanted.
A new USE_PYTHON=optsuffix that will add PYTHON_PKGNAMESUFFIX has been
added to cope with Python ports that did not have the Python
PKGNAMEPREFIX but are flavored.
USES=python now also exports a PY_FLAVOR variable that contains the
current python flavor. It can be used in dependency lines when the
port itself is not python flavored. For example, deskutils/calibre.
By default, all the flavors are generated. To only generate flavors
for the versions in PYTHON2_DEFAULT and PYTHON3_DEFAULT, define
BUILD_DEFAULT_PYTHON_FLAVORS in your make.conf.
In all the ports with Python dependencies, the *_DEPENDS entries MUST
end with the flavor so that the framework knows which to build/use.
This is done by appending '@${PY_FLAVOR}' after the origin (or
@${FLAVOR} if in a Python module with Python flavors, as the content
will be the same). For example:
RUN_DEPENDS= ${PYTHON_PKGNAMEPREFIX}six>0:devel/py-six@${PY_FLAVOR}
PR: 223071
Reviewed by: portmgr, python
Sponsored by: Absolight
Differential Revision: https://reviews.freebsd.org/D12464
armv7, mark them so.
This is part two of a multipart commit to bring armv7 ports to parity
with armv6.
Approved by: portmgr (tier-2 blanket)
Obtained from: lonesome.com -exp run
- Fix in CegoBTreeValue::valueFromSchema, introduced method
getReservedLength to return fixed value length for type fixed and
decimal. Since for decimal and fixed type just the dimension is
stored in CegoField::getLength, we must define a reserved area
for those types. This is done now with constant definition
RESERVED_BTREE_FLOATLEN in CegoDefs.h
- Improved ordersize calculation in CegoOrderSpace:insertTuple. Now
the following formula is used :
int s = sizeof(fv);
if ( fv.getLength() > STATICFIELDBUF )
s += fv.getLength();
orderEntryLen += s;
The catches the base memory usage for the instance but also the
dynamic allocated part
Submitted by: Bjoern Lemke <lemke@lemke-it.com>
- Fix in CegoXMLSpace::setTSSortAreaSize, changed type from int to
unsigned long long, otherwise overflow may occur for large sort
area size configurations
- Performance optimization done in CegoOrderSpace. Instead of storing
the CegoField list for the result tuple in CegoOrderNode, just the
values are stored there. The order cursor now gets the schema with
the constructor and build the CegoField List in the getFirst /
getNext methods. This seems a huge amount of heap space, which is
dynamically allocated during the sorting procedure ( sortareasize
parameter ).
Especially for large ordering result sets this saves significantly
memory and improves performance.
Submitted by: Bjoern Lemke <lemke@lemke-it.com>
- Fix in CegoPredDesc::decode, loop variable i for IN/NOTIN case
was not initialized
- Fix in CegoAction::execCheckCreate, check contraints are checked
for contained subselects. subselects are ( still ) not supported
since the CegoCheckOject::decode method can not provide a valid
tablemanager instance ( it is called from CegoObjectManager )
Submitted by: Bjoern Lemke <lemke@lemke-it.com>
