--1.0.0--
* Bumped to major version number for 10 year anniversary
* Added module for inverse Laplace transforms, including the top level
function invertlaplace() as well as several different algorithms
(Talbot, Gaver-Stehfest and de Hoog) implemented in
mpmath.calculus.inverselaplace (Kris Kuhlman)
* Fixed bugs in elliprg() giving incorrect values for certain input
* Fixed wrong degree 1 nodes for Gaussian quadrature
* Made make acot(0) and acoth(0) return a finite result
* Fixed sieved zeta sum not being used in Python 3, and added cutoff
for sieved zeta sum on 32-bit systems when too much memory would be used
* Fixed zeta(0,0.5) to return correct value instead of raising
NoConvergence exception
* Added detection of exact zeros in gammainc(), in particular fixing
NoConvergence error for gammainc(3,-1+1j)
* Fixed wrong values from besseli() due to improper internal precision
* Fixed bessely(0,1j) to return complex nan instead of raising NameError
(Paul Masson)
* Changed float() and complex() applied to an mpf or mpc to use rounding
to nearest (or the context rounding mode) instead truncating
* Fix imaginary part of gammainc(n,x), n negative odd int, x < 0
* Added alternative "phase" color scheme to cplot()
* Better error message for int(inf) or int(nan) (Aaron Meurer)
* Fixed polyroots() with error=True
* Added support to pass optional initial values to polyroots()
(Michael Kagalenko)
* Rewrote the Python major version selection to make it work if something
else has redefined xrange (Arne Brys)
* Switched documentation formula rendering to MathJax (Sergey B Kirpichev)
* Fixed documentation TeX build (Sergey B Kirpichev)
* Added PEP8 conformity testing (Sergey B Kirpichev)
* Various fixes for the test code and test infrastructure on different
platforms and Python versions (Sergey B Kirpichev)
* Fixed module paths in setup.py (Aaron Meurer)
* Documented more options for methods such as nstr() and hyper()
* Miscellaneous corrections to the documentation (various)
Version 1.17 only fixes compatibility with Python 3.x and other C extensions that use gmpy.
Version 1.16 fixes a minor issue with the new Decimal type in Python 3.3.
What's new in Tornado 4.5.3
tornado.curl_httpclient
- Improved debug logging on Python 3.
tornado.httpserver
- Content-Length and Transfer-Encoding headers are no longer
sent with 1xx or 204 responses (this was already true of 304
responses).
- Reading chunked requests no longer leaves the connection in a broken
state.
tornado.iostream
- Writing a memoryview can no longer result in "BufferError:
Existing exports of data: object cannot be re-sized".
tornado.options
- Duplicate option names are now detected properly whether they use
hyphens or underscores.
tornado.testing
- .AsyncHTTPTestCase.fetch now uses 127.0.0.1 instead of
localhost, improving compatibility with systems that have
partially-working ipv6 stacks.
tornado.web
- It is no longer allowed to send a body with 1xx or 204 responses.
tornado.websocket
- Requests with invalid websocket headers now get a response with
status code 400 instead of a closed connection.
v38.4.0
-------
* #1231: Removed warning when PYTHONDONTWRITEBYTECODE is enabled.
v38.3.0
-------
* #1210: Add support for PEP 345 Project-URL metadata.
* #1207: Add support for ``long_description_type`` to setup.cfg
declarative config as intended and documented.
2-Jan-2018
2.3.5
* Add configure option --enable-setlibversion set mark the libs with VERS_3.52
Linux only, so any driver built with the libs will work with closed source
DM's
* Add persistent storage of isql command line history if readline() is used (thanks Axel)
* Rename some local mutex functions to avoid name clashes
* Assorted fixes (Thanks Markus * 2)
* Fix regression in ini caching
* Make SQLDrivers look in user as well as system odbcinst.ini for driver attributes
* If in use, clear the ini cache when a write is done via SQLWritePrivateProfileString() so
the new value is read
* Fix problem with pooling if the environment was released by the application
* Add check for SQL_COLUMN_COUNT in SQLColAttribute
* isql would not display long error messages. Fixed now.
* Fix problem calling the driver to report errors if the error is from the DM and the driver
has not been called
* SQLSetConnectAttrW crashes when attempting to set SQL_ATTR_LOGIN_TIMEOUT
* Buffer overflow in unicode_to_ansi_copy()
* SQLDriverConnect with not-found FILEDSN causes crash
* SQLGetDescRec with null name pointer causes crash
* Connection string escaping does not work
* SQLDriverConnect/W with very long driver name causes crash
* Connection string with trailing empty value causes crash
* Freeing explicitly allocated descriptor results in writing to freed memory
* Buffer overflows and missing null checks in SQLConfigDataSource, SQLInstallDriverEx, and SQLWriteFileDSN
* Statement enters incorrect state upon SQLExecDirectW returning SQL_NO_DATA
* SQLBulkOperations fails to exit async state after success
* SQLFreeStmt causes prepared statements in S1 or S2 to erroneously transition to S3
* Buffer length fixes for SQLGetDiagField
* SQLSetConnectAttrW and Unicode string pre-connect attributes do not work
* SQLGetData and SQLSetPos async states are incorrect
* Various string conversion and length issues in SQLColAttribute(s)
* Missing buffer length check in SQLColAttribute(s)W
* SQLGetStmtAttr state handling incorrect in S5, S6, and S7 (via SQLExtendedFetch)
* SQLSetPos and SQLExtendedFetch state management fixes
* SQLExecDirect/W erroneous transition to S1 upon error in S5
* Async SQLGetData and SQLExecDirect/W fails to restore state upon cancellation
* SQLFetchScroll cannot move cursor back into the rowset
* SQLSetDescField doesn't adjust the length of the buffer when converting to unicode
* SQLGetDescField/W and SQLSetDescField/W do not check for negative buffer lengths
* SQLSetStmtAttrW SQL_ATTR_APP_PARAM_DESC and SQL_ATTR_APP_ROW_DESC does not accept null
* SQLGetData async state reentrancy issues
* SQLDriversW off-by-one in enumerating driver list
* SQLPrimaryKeys does not pass length in characters to driver
* SQLGetConnectAttr with string attributes truncates to half buffer length
* SQLTransact with autocommit enabled erroneously changes statement state
* SQLDataSources/W fails to reset list position upon end
* SQLGetEnvAttr successfully returns unset SQL_ATTR_ODBC_VERSION
* Lack of SQL_HANDLE_SENV support
* SQLAllocEnv fails to set environment version correctly
* SQLMoreResults with streamed output parameters returns unexpected HY010
* Custom pre-connect pointer attributes are truncated to 32 bits
* 08003 message should be "Connection not open"
* SQL_ATTR_ACCESS_MODE set using SQLSetConnectOption/W before connecting does not persist after disconnecting
* SQL_ATTR_AUTOCOMMIT incorrect default value before connecting
* SQL_AUTOCOMMIT set using SQLSetConnectOption/W before connecting does not persist after disconnecting
* SQLAllocHandle/SQLFreeHandle with invalid handle type should return SQL_INVALID_HANDLE
* SQLAllocHandleStd not setting ODBC version correctly
* SQLBindParameter does not ignore BufferLength for DAE parameters
* SQLBindParameter does not ignore BufferLength for fixed-length parameters
* SQLBindParameter returns "Invalid application buffer type" instead of " Program type out of range"
* SQLCancel with 01S05 returned from driver should result in SQL_SUCCESS, not SQL_SUCCESS_WITH_INFO
(see https://msdn.microsoft.com/en-us/library/aa392708(v=vs.85).aspx )
* SQLColAttribute/SQLColAttributes should return number of bytes needed for Unicode string when truncating
* SQLColumnPrivileges/W differing error message precedence from Windows DM and order in ODBC spec
* SQLColumns/W extraneous checks on null string's length
* SQLCopyDesc does not copy descriptors across connections correctly
* SQLDescribeParam extraneous checks for state
* SQLDriverConnect/W adds extraneous DM prefix to diagnostic messages
* SQLDriversW before ODBC version set returns incorrect SQLSTATE
* SQLFetch in state S7 does not return error from DM
* SQLFetchScroll missing check for SQL_FETCH_BOOKMARK
* SQLForeignKeys/W missing check for null table names
* SQLGetConnectAttr/W erroneously retrieves attributes with no default (SQL_ATTR_PACKET_SIZE, SQL_ATTR_QUIET_MODE)
* SQLGetConnectAttr/W fails to get some set attributes before connecting
* SQLGetConnectAttr/W fails to retrieve set attributes with no connection
* SQLGetConnectAttrW returns incorrect value for SQL_ATTR_TRACE
* SQLGetConnectOption/W cannot retrieve SQL_ODBC_CURSORS before connection
* SQLGetConnectOption/W fails to retrieve SQL_LOGIN_TIMEOUT before connecting
* SQLGetConnectOption/W fails to retrieve SQL_ATTR_ACCESS_MODE set with SQLSetConnectOption/W
* SQLGetData and SQLSetPos async states are incorrect
* SQLGetData missing check for cursor end indication; SQLSetCursorName/W fails to clear previous diagnostic records
* SQLGetDiagField/W does not check record number for SQL_DIAG_ROW_COUNT and SQL_DIAG_DYNAMIC_FUNCTION_CODE
* SQLGetDiagField/W missing check for negative buffer length for string fields
* SQLGetDiagField/W inconsistent handling of statement-only diagnostic fields
* SQLGetInstalledDrivers off-by-one length
* SQLGetStmtOption various state handling issues
* SQLSetConnectAttr/W SQL_ATTR_CURRENT_CATALOG extraneous check with error 24000
* SQLSetConnectAttr/W does not prevent attempts to set ODBC 3.x statement attributes
* SQLSetConnectAttr/W with null string attributes causes crash
* SQLSetConnectOption passes SQL_ATTR_TRACEFILE to the driver
* SQLSetConnectOption/W or SQLSetConnectAttr/W missing validity checks for SQL_ATTR_TXN_ISOLATION
* SQLSetConnectOption/W setting SQL_ATTR_TRACEFILE to null results in different error
* SQLSetCursorName/W missing checks for negative name length
* SQLSetDescField/W missing check for negative SQL_DESC_COUNT
* SQLSetParam missing various error checks for invalid data types and buffer
* SQLSetPos does not check for state S5
* SQLSetPos missing checks in state S7
* SQLSetScrollOptions various state handling issues
* SQLSetStmtOption/W missing check for positive rowset sizes
* SQLSpecialColumns/W error precedence differs from Windows DM
* SQLSpecialColumns/W incorrect check for SQL_NTS string lengths
* SQLStatisticsW uniqueness parameter missing validation
* SQLTablePrivileges/W extraneous checks on null string's length
* Various issues with SQLGetFunctions
* Various string conversion and length issues in SQLColAttribute; missing buffer length check in SQLColAttributeW
* As above, but for SQLColAttributes and SQLColAttributesW; incorrect check for SQL_COLUMN_COUNT
* Setting SQL_ATTR_TRACEFILE to null value results in different error
* check_target_type allows driver-specific C data types for ODBC < 3.8
* fix empty SQL_DIAG_SERVER_NAME field in DM-supplied diag recs
* fix differing behaviour for an empty string DSN in SQLConnect/W
* Alter isql to return errors from SQLMoreResults
* Handle case of building on mingw-w64
GTK-Doc 1.27 (Dec 07 2017)
==============
Finetune the python port.
Changes
o 773879 : scangobj: Do not generate unused parameters
o 786174 : Several test failures on Arch Linux
o 787495 : gtkdoc-fixxref crashes when running vim for syntax highlighting
o 787768 : Aborts when glib2.0-dev is not installed
o 787862 : Always open files in text mode and always use utf-8
o 788473 : fixxref crashes if a .devhelp2.gz file is found
o 789531 : Don't depend on the system shell
o 790022 : mkdb: Add support for (not optional) annotation
o 791131 : gtk-doc and python3: fixxref failures
Release 0.62.0
core:
* Stop supporting lcms1, you really want to use lcms2 :)
* Stop supporting openjpeg1, you really want to use openjpeg2 :)
* Open files that state 8 bits as third field of W. Bug #103469
* GfxLabColorSpace::parse: Fix crash in broken documents. Bug #103582
* Fix leak if parseDA fails
* Include glibc.h where needed
* Document the meaning of the 'type' integer of a shading
* Fix UTF test fail
* INSTALL: add debug options
qt5:
* Add API to let the rendering process callback to get a partial rendering. Bug #103372
qt4:
* Remove the Qt4 frontend
utils:
* Support unicode on windows console
* pdfsig: install man page
* sort encoding list
glib:
* demo: fix warning
Add upstream patch to fix a test case.
Mercurial 4.4.2 (2017-12-01)
This is a regularly-scheduled bugfix release.
1.1. Notable changes
1.1.1. Stricter command option parsing
Mercurial can now optionally parse "early" options (-R/--repository,
--cwd, --config, --debugger, and --profile) more strictly, for more
secure integration with tools that invoke 'hg' commands. Setting
HGPLAIN=+strictflags will parse these options more strictly, which
prevents them from being injected as arguments to other flags.
1.2. Bug fixes
'hg amend' now correctly handles deleted and removed files, as
well as subrepos. (issue5732, issue5677)
largefiles now correctly handles dropped standin files when
updating largefiles.
Fixed an issue with deleting symlinks to directories when
ui.origbackuppath is set. (issue5731)
1.3. Performance improvements
Improved performance in path conflict checking introduced in
Mercurial 4.4. (issue5716)
