0d85be19e2
*Don't* kill all local variables on function exit. This will be done
...
by the frameobject dealloc when it is time for the locals to go. When
there's still a traceback object referencing this stack frame, we
don't want the local variables to disappear yet.
(Hmm... Shouldn't they be copied to the f_locals dictionary?)
1997-02-14 16:32:14 +00:00
e8811f85ed
Added intern() function.
1997-02-14 15:48:05 +00:00
deb0c5e66c
Two small changes:
...
- Use co->... instead of f->f_code->...; save an extra lookup of what
we already have in a local variable).
- Remove test for nlocals > 0 before setting fastlocals to
f->f_localsplus; 0 is a rare case and the assignment is safe even
then.
1997-01-27 23:42:36 +00:00
d0eb429b88
Plug a leak with calling something other than a function or method is
...
called with keyword arguments -- the keyword and value were leaked.
This affected for instance with a __call__() method.
Bug reported and fix supplied by Jim Fulton.
1997-01-27 21:30:09 +00:00
950361c6ca
Patches for (two forms of) optional dynamic execution profiling --
...
i.e., counting opcode frequencies, or (with DXPAIRS defined) opcode
pair frequencies. Define DYNAMIC_EXECUTION_PROFILE on the command
line (for this file and for sysmodule.c) to enable.
1997-01-24 13:49:28 +00:00
8c5df06ec7
Change the control flow for error handling in the function prelude to
...
jump to the "Kill locals" section at the end. Add #ifdef macintosh
bandaid to make sure we call sigcheck() on the Mac.
1997-01-24 04:19:24 +00:00
43f1b8d6e4
Added optional interface for dynamic execution profile (to be gathered
...
in ceval.c).
1997-01-24 04:07:45 +00:00
b3f515af64
Get the line number from PyCode_Addr2Line instead of believing
...
tb_lineno. Store it in tb_lineno for the user.
1997-01-24 04:02:55 +00:00
99d182550b
New magin number (because of linenumber table).
1997-01-24 03:44:53 +00:00
d031c89891
Marshal the line number table of code objects.
1997-01-24 03:44:17 +00:00
da4eb5c3b5
Instead of emitting SET_LINENO instructions, generate a line number
...
table which is incorporated in the code object. This way, the runtime
overhead to keep track of line numbers is only incurred when an
exception has to be reported.
1997-01-24 03:43:35 +00:00
a4240132ec
Kill all local variables on function return. This closes a gigantic
...
leak of memory and file descriptors (thanks for Roj for reporting
that!). Alas, the speed goes down by 5%. :-(
1997-01-21 21:18:36 +00:00
70d44787a3
Only call sigcheck() at the ticker code if we don't have true signals.
...
This is safe now that both intrcheck() and signalmodule.c schedule a
sigcheck() call via Py_AddPendingCall().
This gives another 7% speedup (never run such a test twice ;-).
1997-01-21 06:15:24 +00:00
1aa14838d2
Cleanup:
...
- fix bug in Py_MakePendingCalls() with threading
- fix return type of do_raise
- remove build_slice (same as PySlice_New)
- remove code inside #if 0
- remove code inside #ifdef CHECK_STACK
- remove code inside #ifdef SUPPORT_OBSOLETE_ACCESS
- comment about newimp.py should refer to ni.py
1997-01-21 05:34:20 +00:00
554fa49500
get build info from elsewhere
1997-01-20 18:34:55 +00:00
768360243a
Changes for frame object speedup:
...
- get fastlocals differently
- call newframeobject() with fewer arguments
- toss getowner(), which was unused anyway
1997-01-20 04:26:20 +00:00
7eb883a18e
Remove unused variable.
1997-01-18 20:04:05 +00:00
8d75161671
Intern the string "__complex__".
1997-01-18 08:04:16 +00:00
7b89b6a660
Intern all names and varnames in newcodeobject(), plus those string
...
literals that look like identifiers. Also intern all strings used as
names during the compilation.
1997-01-18 08:02:57 +00:00
3dfd53b4c8
Add "if (x != NULL) continue;" (or similar for err==0) before the
...
break to most cases, as suggested by Tim Peters. This gives another
8-10% speedup.
1997-01-18 02:46:13 +00:00
98a9b312e8
Marshal the new stacksize item in code objects.
1997-01-17 21:07:08 +00:00
7af8130857
Working semaphore implementation by Sjoerd.
1997-01-17 21:06:41 +00:00
dd5db43905
New MAGIC number (code objects have one more item when marshalled).
1997-01-17 21:06:11 +00:00
62f7d15d0b
Use the stack size from the code object and the CO_MAXBLOCKS constant
...
from compile.h. Remove all eval stack overflow checks.
1997-01-17 21:05:28 +00:00
8b993a98db
Add co_stacksize field to codeobject structure, and stacksize argument
...
to PyCode_New() argument list. Move MAXBLOCKS constant to conpile.h.
Added accurate calculation of the actual stack size needed by the
generated code.
Also commented out all fprintf statements (except for a new one to
diagnose stack underflow, and one in #ifdef'ed out code), and added
some new TO DO suggestions (now that the stacksize is taken of the TO
DO list).
1997-01-17 21:04:03 +00:00
635abd24f0
Check for duplicate keyword arguments at compile time.
1997-01-06 22:56:52 +00:00
8f49e12a0e
Make builtin_module_names a tuple instead of a list.
1997-01-06 22:55:54 +00:00
408027ea46
Rename DEBUG macro to Py_DEBUG
1996-12-30 16:17:54 +00:00
0aa9ee65ab
Moved the raise logic out of the main interpreter loop to a separate function.
...
The raise logic has one additional feature: if you raise <class>,
<value> where <value> is not an instance, it will construct an
instance using <value> as argument. If <value> is None, <class> is
instantiated without arguments. If <value> is a tuple, it is used as
the argument list.
This feature is intended to make it easier to upgrade code from using
string exceptions to using class exceptions; without this feature,
you'd have to change every raise statement from ``raise X'' to ``raise
X()'' and from ``raise X, y'' to ``raise X(y)''. The latter is still
the recommended form (because it has no ambiguities about the number
of arguments), but this change makes the transition less painful.
1996-12-10 18:07:35 +00:00
c1547d930f
Better way to handle 64-bit ints, keeping gcc -Wall happy.
...
Tested with AMK's help.
1996-12-10 15:39:04 +00:00
1a2c5cbcc4
Add unistd.h to make gcc -Wall happy.
1996-12-10 15:37:36 +00:00
067998f35e
Add const to error and newstring functions
1996-12-10 15:33:34 +00:00
80bb9655f0
Keep gcc -Wall happy.
1996-12-05 23:27:02 +00:00
ed0af8fe70
Support __complex__ method on instances, for complex() conversion.
...
Keep gcc -Wall happy.
1996-12-05 23:18:18 +00:00
150b2df682
Change the Don Beaudry hack into the Don B + Jim F hack; now, if *any*
...
base class is special it gets invoked.
Make gcc -Wall happy.
1996-12-05 23:17:11 +00:00
3afb5959aa
Some extra flags that an HPUX user wants me to add.
1996-12-05 23:15:35 +00:00
b0c168cbe8
Tentative changes to make this work better on 64-bit machines.
...
A plain int that doesn't fit in 32 bits will be marshalled using a new
type. 32 bits machines can't handle this and will issue a warning.
1996-12-05 23:15:02 +00:00
d266eb460e
New permission notice, includes CNRI.
1996-10-25 14:44:06 +00:00
b75fba04c7
Forget about Ellipses b/w compatibility.
1996-10-16 04:18:36 +00:00
e449af7da9
Ellipses -> Ellipsis rename (the dictionary really says that it should
...
be Ellipsis!).
Bumped the API version because a linker-visible symbol is affected.
Old C code will still compile -- there's a b/w compat macro.
Similarly, old Python code will still run, builtin exports both
Ellipses and Ellipsis.
1996-10-11 16:25:41 +00:00
cf1474b73a
Sjoerd's thread changes (including down_sema typo fix).
...
Note: waitflag not supported on NT.
1996-10-08 14:17:53 +00:00
452a9833c9
Added line number to most compile-time error messages.
1996-09-17 14:32:04 +00:00
9b38a145e2
Rationalized MS ifdefs
1996-09-11 23:12:24 +00:00
cc88341e6d
Changes to setpythonpath():
...
Test for / as well as for SEP for MS filenames.
Drop trailing separator from sys.path[0] for MS and Unix filenames.
1996-09-10 14:44:21 +00:00
b072150d7f
Stupid bug: complex(x,y) would yield x+xj
1996-09-07 15:55:27 +00:00
57b1822459
*** empty log message ***
1996-08-29 18:10:41 +00:00
927f6e68fb
Needed more includes...
1996-08-29 18:10:30 +00:00
bae29713ec
*** empty log message ***
1996-08-29 17:48:26 +00:00
6eea32622c
typo (#indef -> #undef)
1996-08-26 14:58:54 +00:00
d8a6d1c2e7
Afterthough: leave both leading underscores in,
...
so __spam becomes _ClassName__spam.
1996-08-24 07:54:27 +00:00
fe2236f3c8
Oops need to mangle global statement separately
1996-08-24 07:29:04 +00:00
8ff077b094
Name mangling, what the heck!
1996-08-24 06:21:31 +00:00
8b9ea873ad
Use MS_DLL_ID as sys.winver
1996-08-23 18:14:47 +00:00
ac279109d4
PYTHONWIN -> MS_COREDLL
1996-08-22 23:10:58 +00:00
c1d5053882
Add needed #include <ctype.h>
1996-08-21 23:38:24 +00:00
f1d7413445
Add explicit #undef _POSIX_THREADS for use on Solaris 2.5 (Sjoerd).
1996-08-21 14:39:29 +00:00
504f4a9901
Don't require <dlfcn.h> -- rely on existence of dlopen().
1996-08-20 19:59:07 +00:00
7cf1fcf347
Always include config.h
1996-08-19 22:12:39 +00:00
1e6124902a
Many little fixes:
...
- support for SCO_SV dynamic loading
- on Mac, auto-detect dynamic loading by __CFM68K__ or _powerc)
- on Mac, long shared library extension is .cfm68k.slb or .ppc.slb
- on hp, don't redefine hpux if already defined
- add __file__ property to successfully loaded module
1996-08-19 22:12:10 +00:00
71bd363d9c
getprogramname -> Py_GetProgramName.
1996-08-19 22:09:38 +00:00
6d43c5de5a
Raise TypeError, not KeyError, on unknown keyword argument.
1996-08-19 22:09:16 +00:00
3f3bb3d3c9
Added *PyMarshal_WriteObjectToString() (moved the relevant code there
...
from marshal_loads()).
1996-08-19 22:07:17 +00:00
aa35465ccf
Support for keyword arguments (PyArg_ParseTupleAndKeywords) donated by
...
Geoff Philbrick <philbric@delphi.hks.com> (slightly changed by me).
Also a little change to make the file acceptable to K&R C compilers
(HPUX, SunOS 4.x).
1996-08-19 19:32:04 +00:00
bf51afa049
Don't test here for negative number to float power; that belongs in
...
floatobject.c.
1996-08-16 20:49:17 +00:00
79d96d6bff
Don't die in resizestring() on filter(<func>, "").
1996-08-16 20:44:34 +00:00
0dfcf753ad
Disable support for access statement
1996-08-12 22:00:53 +00:00
0bbf253e97
Insert "./" in front of pathname when it contains no '/' (if USE_SHLIB)
1996-08-09 20:55:05 +00:00
1a62311cfb
Changes necessary for AIX.
1996-08-08 18:53:41 +00:00
1741d60725
Added casts from unsigned char to char when calling rds_object() on
...
frozen code.
1996-08-08 18:52:59 +00:00
15cc9a0a76
Removed unused var; added error check for ``lambda: x=1''.
1996-08-08 18:51:04 +00:00
fe4b6ee775
Include mymath.h instead of declaring prototypes for math functions.
...
Fix leak and unchecked error in complex().
1996-08-08 18:49:41 +00:00
9caf77a485
Put definition of _REENTRANT in config.h
1996-08-01 00:52:26 +00:00
d5962adb44
Changes for AIX sharedlibs.
1996-07-31 22:44:53 +00:00
3823420ca4
If NO_DYNAMIC_LINK is defined, load_dynamic_module() will always fail.
1996-07-31 17:55:19 +00:00
5e41644ddd
Added _REENTRANT definition
1996-07-31 17:52:04 +00:00
94a9667f1a
Always insert script directory in front of sys.path -- if there's no
...
sys.argv, insert "". Note that "." is removed as a default component
of the path (see changes to getpath.c and Setup.in).
1996-07-30 20:35:50 +00:00
3b9c6677f8
Better error message if stride used on normal sequence object
1996-07-30 18:40:29 +00:00
6ffd553899
Add 'Ellipses' object.
1996-07-30 18:37:55 +00:00
8861b74445
Changes for slice and ellipses
1996-07-30 16:49:37 +00:00
3ecebf1732
Changes needed by NeXT (the only platform that seems to use this).
1996-07-30 16:48:31 +00:00
a63d9f4d9f
As a side effect of calling PySys_SetArgv (setpythonargv), the
...
directory containing argv[0] is inserted in front of sys.path.
If argv[0] contains no directory, an empty string is inserted.
If argv is empty, nothing happens.
1996-07-24 01:31:37 +00:00
42cd305b25
new .pyc magic number (** operator)
1996-07-21 02:28:18 +00:00
7cffd1322b
added 1996
1996-07-21 02:28:11 +00:00
530956d247
Py_complex; and WITHOUT_COMPLEX added to getargs.c
1996-07-21 02:27:43 +00:00
6f489d989d
Slightly different Windows ifdefs
1996-06-28 20:15:15 +00:00
236f62da7a
Corrections to aix_loaderror(), by Manus Hand.
1996-06-26 21:07:08 +00:00
f21506076d
Jeremy's patches for more robust handling of unmarshallable types.
1996-06-26 20:41:23 +00:00
0e41c8c848
Cosmetic change to the dox_8x3 hack.
1996-06-20 14:18:34 +00:00
3b4da59cd6
Renamed static pow() to powerop() to avoid name conflict in some compilers.
1996-06-19 21:49:17 +00:00
7c46a920ae
New style interface via pointer variable.
1996-06-17 17:07:23 +00:00
cfd0a22252
struct frozen is now struct _frozen and comes from import.h.
1996-06-17 17:06:34 +00:00
8fa9b6f932
Define sys.prefix and sys.exec_prefix (see Modules/getpath.c; from Makefile).
1996-06-17 17:02:48 +00:00
6602099e7c
Got rid of florida hack and made it work with Solaris 2.5 pthreads.
...
Wonder if this will break it on all other platforms :-)
1996-06-11 18:32:18 +00:00
582646aecc
grand renaming; added copyright to some files
1996-05-28 22:30:17 +00:00
0a8626eb88
Plug memory leak (variable names in code objects were being leaked!)
1996-05-28 22:28:12 +00:00
8c1e150215
Removed some done "to do" items.
...
Changed #ifdef DEBUG slightly.
1996-05-24 20:49:24 +00:00
ded690fc35
rename printrefs, getobjects to _Py_ prefix
1996-05-24 20:48:31 +00:00
dadc824c6e
removed verbose decl; added win3.1 dynamic linking.
1996-05-23 22:51:40 +00:00
40f470f7e0
moved verbose decl to pydebug.h; added dos_8x3 feature
1996-05-23 22:51:04 +00:00
f857a6becb
Added copyright.
...
Include config.h which may define COMPILER.
1996-05-23 22:50:26 +00:00
Guido van Rossum