Commit Graph

62 Commits

Author SHA1 Message Date
Mark Shannon fee552669f
Produce cleaner bytecode for 'with' and 'async with' by generating separate code for normal and exceptional paths. (#6641)
Remove BEGIN_FINALLY, END_FINALLY, CALL_FINALLY and POP_FINALLY bytecodes. Implement finally blocks by code duplication.
Reimplement frame.lineno setter using line numbers rather than bytecode offsets.
2019-11-21 09:11:43 +00:00
Raymond Hettinger 0138c4ceab
Fix unused variable and signed/unsigned warnings (GH-15537) 2019-08-27 09:55:13 -07:00
Pablo Galindo 7a68f8c28b
bpo-37289: Remove 'if False' handling in the peephole optimizer (GH-14099) 2019-06-15 15:58:00 +01:00
Pablo Galindo 05f8318655
bpo-37269: Correctly optimise conditionals with constant booleans (GH-14071)
Fix a regression introduced by af8646c805 that was causing code of the form:

if True and False:
   do_something()

to be optimized incorrectly, eliminating the block.
2019-06-14 06:54:53 +01:00
Pablo Galindo 3498c642f4
bpo-37213: Handle negative line deltas correctly in the peephole optimizer (GH-13969)
The peephole optimizer was not optimizing correctly bytecode after negative deltas were introduced. This is due to the fact that some special values (255) were being searched for in both instruction pointer delta and line number deltas.
2019-06-13 19:16:22 +01:00
Pablo Galindo af8646c805
bpo-1875: Raise SyntaxError in invalid blocks that will be optimised away (GH-13332)
Move the check for dead conditionals (if 0) to the peephole optimizer
and make sure that the code block is still compiled to report any
existing syntax errors within.
2019-05-17 11:37:08 +01:00
Victor Stinner 028f0ef4f3
bpo-9566: Fix compiler warnings in peephole.c (GH-10652) 2018-12-07 17:54:18 +01:00
Gregory P. Smith 49fa4a9f1e
bpo-35193: Fix an off by one error in the RETURN_VALUE case. (GH-10418)
Fix an off by one error in the peephole optimizer when checking for unreachable code beyond a return.

Do a bounds check within find_op so it can return before going past the end as a safety measure.

7db3c48833 (diff-a33329ae6ae0bb295d742f0caf93c137)
introduced this off by one error while fixing another one nearby.

This bug was shipped in all Python 3.6 and 3.7 releases.

The included unittest won't fail unless you do a clang msan build.
2018-11-08 17:55:07 -08:00
Serhiy Storchaka 520b7ae27e
bpo-17611. Move unwinding of stack for "pseudo exceptions" from interpreter to compiler. (GH-5006)
Co-authored-by: Mark Shannon <mark@hotpy.org>
Co-authored-by: Antoine Pitrou <antoine@python.org>
2018-02-22 23:33:30 +02:00
INADA Naoki 87010e85cb
bpo-29469: peephole: Remove const_stack (GH-4879)
Constant folding was moved to AST optimizer.
But compiler may emit LOAD_CONSTs + BUILD_TUPLE.
For example, default arguments can be constant tuple
if all arguments are constant.

This commit makes peephole's tuple folding simple.
It doesn't support nested tuples because nested
tuples are folded by AST optimizer already.
2017-12-18 15:52:54 +09:00
Serhiy Storchaka 15a8728415
bpo-29469: Optimize literal lists and sets iterating on the AST level. (#4866) 2017-12-14 20:24:31 +02:00
INADA Naoki eadad1b97f
bpo-29469: Remove unnecessary peephole optimizer (GH-4863)
Conversions like `not a is b -> a is not b` are implemented
in AST optimizer in previous commit (7ea143a).
So this commit removes them from peephole optimizer.
2017-12-14 22:18:26 +09:00
INADA Naoki 7ea143ae79
bpo-29469: Move constant folding to AST optimizer (GH-2858) 2017-12-14 16:47:20 +09:00
Serhiy Storchaka 36ff451eba bpo-30501: Make the compiler producing optimized code for condition expressions. (#1851) 2017-06-11 14:50:22 +03:00
Serhiy Storchaka 7db3c48833 Issue #28517: Fixed of-by-one error in the peephole optimizer that caused
keeping unreachable code.
2016-10-25 09:30:43 +03:00
Serhiy Storchaka a1e9ab34a9 Fixed refactoring bug in dd046963bd42 (issue27129). 2016-09-11 15:19:12 +03:00
Serhiy Storchaka ab8740058a Issue #27129: Replaced wordcode related magic constants with macros. 2016-09-11 13:48:15 +03:00
Raymond Hettinger 08eef3ff49 Re-linewrap comments 2016-08-07 20:20:33 -07:00
Martin Panter 3e04d5b306 Issue #27076: Merge spelling from 3.5 2016-05-26 06:03:19 +00:00
Martin Panter 46f50726a0 Issue #27076: Doc, comment and tests spelling fixes
Most fixes to Doc/ and Lib/ directories by Ville Skyttä.
2016-05-26 05:35:26 +00:00
Serhiy Storchaka b0f80b0312 Issue #26647: Python interpreter now uses 16-bit wordcode instead of bytecode.
Patch by Demur Rumed.
2016-05-24 09:15:14 +03:00
Serhiy Storchaka ec39756960 Issue #22570: Renamed Py_SETREF to Py_XSETREF. 2016-04-06 09:50:03 +03:00
Victor Stinner f3914eb16d co_lnotab supports negative line number delta
Issue #26107: The format of the co_lnotab attribute of code objects changes to
support negative line number delta.

Changes:

* assemble_lnotab(): if line number delta is less than -128 or greater than
  127, emit multiple (offset_delta, lineno_delta) in co_lnotab
* update functions decoding co_lnotab to use signed 8-bit integers

  - dis.findlinestarts()
  - PyCode_Addr2Line()
  - _PyCode_CheckLineNumber()
  - frame_setlineno()

* update lnotab_notes.txt
* increase importlib MAGIC_NUMBER to 3361
* document the change in What's New in Python 3.6
* cleanup also PyCode_Optimize() to use better variable names
2016-01-20 12:16:21 +01:00
Serhiy Storchaka 576f132b98 Issue #20440: Cleaning up the code by using Py_SETREF. 2016-01-05 21:27:54 +02:00
Yury Selivanov 7544508f02 PEP 0492 -- Coroutines with async and await syntax. Issue #24017. 2015-05-11 22:57:16 -04:00
Serhiy Storchaka 67559bf8f1 Issue #23450: Silenced compiler warnings and added asserts in peephole optimizer. 2015-02-16 21:13:24 +02:00
Serhiy Storchaka 1a1ff29659 Issue #23446: Use PyMem_New instead of PyMem_Malloc to avoid possible integer
overflows.  Added few missed PyErr_NoMemory().
2015-02-16 13:28:22 +02:00
Victor Stinner c82729e44f Issue #19437: Fix fold_unaryops_on_constants() of the peephole optimizer, clear
the exception when PyList_Append() fails
2013-11-14 01:21:00 +01:00
Victor Stinner e0af3a802a Issue #18408: Fix PyCode_Optimize(): raise a MemoryError on memory allocation
failure.
2013-07-09 00:32:04 +02:00
Benjamin Peterson 442f20996d create NameConstant AST class for None, True, and False literals (closes #16619) 2012-12-06 17:41:04 -05:00
Martin v. Löwis d63a3b8beb Implement PEP 393. 2011-09-28 07:41:54 +02:00
Stefan Krah 472d280d8b Issue #13002: Fix Visual Studio warning (not enough actual parameters). 2011-09-21 19:08:39 +02:00
Ezio Melotti e1adeeb7fd #5057: Merge with 3.2. 2011-04-15 16:52:35 +03:00
Ezio Melotti 71e84a6e27 #5057: Merge with 3.1. 2011-04-15 16:50:41 +03:00
Ezio Melotti 2df6a93916 Issue #5057: fix a bug in the peepholer that led to non-portable pyc files between narrow and wide builds while optimizing BINARY_SUBSCR on non-BMP chars (e.g. "\U00012345"[0]). 2011-04-15 16:38:34 +03:00
Mark Dickinson 7c9e803284 Issue #11244: Remove outdated peepholer check that was preventing the peepholer from folding -0 and -0.0. Thanks Eugene Toder for the patch. 2011-03-23 17:59:37 +00:00
Raymond Hettinger 29dcaad6eb Issue 11510: Fix BUILD_SET optimizer bug. 2011-03-15 14:50:16 -07:00
Raymond Hettinger 0661e91fed Issue 11510: Fix BUILD_SET optimizer bug. 2011-03-15 15:03:36 -07:00
Antoine Pitrou 17b880a5d6 Issue #11244: The peephole optimizer is now able to constant-fold
arbitrarily complex expressions.  This also fixes a 3.2 regression where
operations involving negative numbers were not constant-folded.
2011-03-11 17:27:02 +01:00
Georg Brandl e5b99f0fb3 Remove redundant includes of headers that are already included by Python.h. 2010-11-30 09:41:01 +00:00
Raymond Hettinger 819a064556 Issue 8403: Don't mask KeyboardInterrupt during peephole operation. 2010-08-22 08:39:49 +00:00
Antoine Pitrou 7f14f0d8a0 Recorded merge of revisions 81032 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/branches/py3k

................
  r81032 | antoine.pitrou | 2010-05-09 17:52:27 +0200 (dim., 09 mai 2010) | 9 lines

  Recorded merge of revisions 81029 via svnmerge from
  svn+ssh://pythondev@svn.python.org/python/trunk

  ........
    r81029 | antoine.pitrou | 2010-05-09 16:46:46 +0200 (dim., 09 mai 2010) | 3 lines

    Untabify C files. Will watch buildbots.
  ........
................
2010-05-09 16:14:21 +00:00
Antoine Pitrou f95a1b3c53 Recorded merge of revisions 81029 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk

........
  r81029 | antoine.pitrou | 2010-05-09 16:46:46 +0200 (dim., 09 mai 2010) | 3 lines

  Untabify C files. Will watch buildbots.
........
2010-05-09 15:52:27 +00:00
Antoine Pitrou b7fbcd396f Issue #6690: Optimize the bytecode for expressions such as `x in {1, 2, 3}`,
where the right hand operand is a set of constants, by turning the set into
a frozenset and pre-building it as a constant.  The comparison operation
is made against the constant instead of building a new set each time it is
executed (a similar optimization already existed which turned a list of
constants into a pre-built tuple).  Patch and additional tests by Dave
Malcolm.
2010-01-16 18:37:38 +00:00
Raymond Hettinger af7adad51d Peephole constant folding had missed UNARY_POSITIVE. 2009-10-22 11:22:50 +00:00
Georg Brandl 194da4a7da Merged revisions 74126,74130-74131,74149,74155,74157,74180-74183,74398 via svnmerge from
svn+ssh://svn.python.org/python/branches/py3k

................
  r74126 | alexandre.vassalotti | 2009-07-21 02:39:03 +0200 (Di, 21 Jul 2009) | 14 lines

  Merged revisions 73871 via svnmerge from
  svn+ssh://pythondev@svn.python.org/python/trunk

  ........
    r73871 | alexandre.vassalotti | 2009-07-06 22:17:30 -0400 (Mon, 06 Jul 2009) | 7 lines

    Grow the allocated buffer in PyUnicode_EncodeUTF7 to avoid buffer overrun.

    Without this change, test_unicode.UnicodeTest.test_codecs_utf7 crashes in
    debug mode. What happens is the unicode string u'\U000abcde' with a length
    of 1 encodes to the string '+2m/c3g-' of length 8. Since only 5 bytes is
    reserved in the buffer, a buffer overrun occurs.
  ........
................
  r74130 | alexandre.vassalotti | 2009-07-21 02:57:50 +0200 (Di, 21 Jul 2009) | 2 lines

  Add ignore rule for the Doc/tools/jinga2/ directory.
................
  r74131 | alexandre.vassalotti | 2009-07-21 04:51:58 +0200 (Di, 21 Jul 2009) | 13 lines

  Merged revisions 73683,73786 via svnmerge from
  svn+ssh://pythondev@svn.python.org/python/trunk

  ........
    r73683 | georg.brandl | 2009-06-29 10:44:49 -0400 (Mon, 29 Jun 2009) | 1 line

    Fix error handling in PyCode_Optimize, by Alexander Schremmer at EuroPython sprint.
  ........
    r73786 | benjamin.peterson | 2009-07-02 18:56:16 -0400 (Thu, 02 Jul 2009) | 1 line

    condense with assertRaises
  ........
................
  r74149 | ezio.melotti | 2009-07-21 22:37:52 +0200 (Di, 21 Jul 2009) | 9 lines

  Merged revisions 74148 via svnmerge from
  svn+ssh://pythondev@svn.python.org/python/trunk

  ........
    r74148 | ezio.melotti | 2009-07-21 23:18:27 +0300 (Tue, 21 Jul 2009) | 1 line

    #6536 fixed typo
  ........
................
  r74155 | alexandre.vassalotti | 2009-07-22 04:24:49 +0200 (Mi, 22 Jul 2009) | 2 lines

  Issue #6242: Fix deallocator of io.StringIO and io.BytesIO.
................
  r74157 | alexandre.vassalotti | 2009-07-22 05:07:33 +0200 (Mi, 22 Jul 2009) | 2 lines

  Issue #6241: Better type checking for the arguments of io.StringIO.
................
  r74180 | ezio.melotti | 2009-07-22 23:17:14 +0200 (Mi, 22 Jul 2009) | 9 lines

  Merged revisions 74179 via svnmerge from
  svn+ssh://pythondev@svn.python.org/python/trunk

  ........
    r74179 | ezio.melotti | 2009-07-23 00:08:49 +0300 (Thu, 23 Jul 2009) | 1 line

    #6423 has_key -> in
  ........
................
  r74181 | alexandre.vassalotti | 2009-07-22 23:27:53 +0200 (Mi, 22 Jul 2009) | 6 lines

  Clean up test_curses.

  By using __stdout__ directly, test_curses caused regrtest.py
  to duplicate the output of some test results.
................
  r74182 | alexandre.vassalotti | 2009-07-22 23:29:01 +0200 (Mi, 22 Jul 2009) | 2 lines

  Use assertGreater instead of assertTrue(x > y).
................
  r74183 | alexandre.vassalotti | 2009-07-23 01:27:17 +0200 (Do, 23 Jul 2009) | 4 lines

  Specialize assertTrue checks when possible.

  We should get slightly more helpful failure messages with this change.
................
  r74398 | georg.brandl | 2009-08-13 11:16:39 +0200 (Do, 13 Aug 2009) | 1 line

  #6694: fix old function names.
................
2009-08-13 09:34:05 +00:00
Alexandre Vassalotti 6f82818652 Merged revisions 73683,73786 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk

........
  r73683 | georg.brandl | 2009-06-29 10:44:49 -0400 (Mon, 29 Jun 2009) | 1 line

  Fix error handling in PyCode_Optimize, by Alexander Schremmer at EuroPython sprint.
........
  r73786 | benjamin.peterson | 2009-07-02 18:56:16 -0400 (Thu, 02 Jul 2009) | 1 line

  condense with assertRaises
........
2009-07-21 02:51:58 +00:00
Benjamin Peterson 876b2f286b Merged revisions 72912,72920,72940 via svnmerge from
svn+ssh://pythondev@svn.python.org/python/trunk

........
  r72912 | benjamin.peterson | 2009-05-25 08:13:44 -0500 (Mon, 25 May 2009) | 5 lines

  add a SETUP_WITH opcode

  It speeds up the with statement and correctly looks up the special
  methods involved.
........
  r72920 | benjamin.peterson | 2009-05-25 15:12:57 -0500 (Mon, 25 May 2009) | 1 line

  take into account the fact that SETUP_WITH pushes a finally block
........
  r72940 | benjamin.peterson | 2009-05-26 07:49:59 -0500 (Tue, 26 May 2009) | 1 line

  teach the peepholer about SETUP_WITH
........
2009-06-28 03:18:59 +00:00
Jeffrey Yasskin 4dd40525e6 Fix 2 oversights from r69961. 2009-02-28 19:49:43 +00:00
Jeffrey Yasskin 9de7ec7868 http://bugs.python.org/issue4715
This patch by Antoine Pitrou optimizes the bytecode for conditional branches by
merging the following "POP_TOP" instruction into the conditional jump.  For
example, the list comprehension "[x for x in l if not x]" produced the
following bytecode:

  1           0 BUILD_LIST               0
              3 LOAD_FAST                0 (.0)
        >>    6 FOR_ITER                23 (to 32)
              9 STORE_FAST               1 (x)
             12 LOAD_FAST                1 (x)
             15 JUMP_IF_TRUE            10 (to 28)
             18 POP_TOP
             19 LOAD_FAST                1 (x)
             22 LIST_APPEND              2
             25 JUMP_ABSOLUTE            6
        >>   28 POP_TOP
             29 JUMP_ABSOLUTE            6
        >>   32 RETURN_VALUE

but after the patch it produces the following bytecode:

  1           0 BUILD_LIST               0
              3 LOAD_FAST                0 (.0)
        >>    6 FOR_ITER                18 (to 27)
              9 STORE_FAST               1 (x)
             12 LOAD_FAST                1 (x)
             15 POP_JUMP_IF_TRUE         6
             18 LOAD_FAST                1 (x)
             21 LIST_APPEND              2
             24 JUMP_ABSOLUTE            6
        >>   27 RETURN_VALUE

Notice that not only the code is shorter, but the conditional jump
(POP_JUMP_IF_TRUE) jumps right to the start of the loop instead of going through
the JUMP_ABSOLUTE at the end. "continue" statements are helped
similarly.

Furthermore, the old jump opcodes (JUMP_IF_FALSE, JUMP_IF_TRUE) have been
replaced by two new opcodes:
- JUMP_IF_TRUE_OR_POP, which jumps if true and pops otherwise
- JUMP_IF_FALSE_OR_POP, which jumps if false and pops otherwise
2009-02-25 02:25:04 +00:00