Guido van Rossum [Sun, 10 Aug 1997 16:47:17 +0000 (16:47 +0000)]
Support using -p/-P to point to the source/build directory instead of
the install directory. Added -h option to print the full usage
message; by default, only two lines are now printed for errors.
Jack Jansen [Fri, 8 Aug 1997 14:51:54 +0000 (14:51 +0000)]
Modified for CW Pro 1 projects. Convention used: .mu files are old
project files, which have to be cleaned up before checking in, .prj
files are CW Pro 1 projects (which are always clean). Prj files are
still binhexed, even though they only have a data fork (and, hence,
could be checked in in binary mode).
Jack Jansen [Fri, 8 Aug 1997 14:49:02 +0000 (14:49 +0000)]
Generate class, property and comparison code (finally!). The resulting
code isn't ideal yet: xxx.Window(1).Paragraph(3).font will only work
if all the classes and properties are declared in the same suite, but
at least font(Paragraph(3, Window(1))) always works.
Got the new structure working with MSVC 4.2.
main_nt.c is gone -- we can use Modules/python.c now.
Added Mark Hammond's module msvcrt.c (untested).
Added several new symbols.
Add a cast to the call to _Py_Dealloc in the expanded version of
Py_DECREF, to reduce the warnings when compiling with reference count
debugging on. (There are still warnings for each call to
_Py_NewReference -- too bad.)
New rules for deleting modules. Rather than having an elaborate
scheme based on object's types, have a simple two-phase scheme based
on object's *names*:
/* To make the execution order of destructors for global
objects a bit more predictable, we first zap all objects
whose name starts with a single underscore, before we clear
the entire dictionary. We zap them by replacing them with
None, rather than deleting them from the dictionary, to
avoid rehashing the dictionary (to some extent). */
Change the Fini function to only remove otherwise unreferenced strings
from the interned table. There are references in hard-to-find static
variables all over the interpreter, and it's not worth trying to get
rid of all those; but "uninterning" isn't fair either and may cause
subtle failures later -- so we have to keep them in the interned
table.
Also get rid of no-longer-needed insert of None in interned dict.
Provide a dummy empty directory as f_builtins instead of failing, when
no valid directory is passed in. This prevents __del__ to fail when
invoked after __builtins__ has already been discarded.
Also add PyFrame_Fini() to discard the cache of frames.
Removed fatal errors from Py_Initmodule4() (and thus from
Py_Initmodule(), which is a macro wrapper around it).
The return value is now a NULL pointer if the initialization failed.
This may make old modules fail with a SEGFAULT, since they don't
expect this kind of failure. That's OK, since (a) it "never" happens,
and (b) they would fail with a fatal error otherwise, anyway.
Tons of extension modules should now check the return value of
Py_Initmodule*() -- that's on my TODO list.
Avoid function calls to access the current thread state and builtins
-- the thread state is passed in as an argument and the builtins are a
member thereof.
Mass checkin (more to follow for other directories).
Introduce truly separate (sub)interpreter objects. For now, these
must be used by separate threads, created from C. See Demo/pysvr for
an example of how to use this. This also rationalizes Python's
initialization and finalization behavior:
Py_Initialize() -- initialize the whole interpreter
Py_Finalize() -- finalize the whole interpreter
tstate = Py_NewInterpreter() -- create a new (sub)interpreter
Py_EndInterpreter(tstate) -- delete a new (sub)interpreter
There are also new interfaces relating to threads and the interpreter
lock, which can be used to create new threads, and sometimes have to
be used to manipulate the interpreter lock when creating or deleting
sub-interpreters. These are only defined when WITH_THREAD is defined:
PyEval_AcquireLock() -- acquire the interpreter lock
PyEval_ReleaseLock() -- release the interpreter lock
PyEval_AcquireThread(tstate) -- acquire the lock and make the thread current
PyEval_ReleaseThread(tstate) -- release the lock and make NULL current
Other administrative changes:
- The header file bltinmodule.h is deleted.
- The init functions for Import, Sys and Builtin are now internal and
declared in pythonrun.h.
- Py_Setup() and Py_Cleanup() are no longer declared.
- The interpreter state and thread state structures are now linked
together in a chain (the chain of interpreters is a static variable
in pythonrun.c).
- Some members of the interpreter and thread structures have new,
shorter, more consistent, names.
- Added declarations for _PyImport_{Find,Fixup}Extension() to import.h.
Add a simple way to enable purify; now you can set the Make variable
PURIFY (e.g. in the Setup file or on the make command line) to point
to the purify command, to run purify.
Functionality enhancement: allow other threads to use Tk commands
while one thread is blocked in mainloop(). Also, handle signals (not
just interrupts) as soon as they happen.
Cleanup: remove support for Tcl/Tk versions 7.4/4.0. (I've confirmed
that it works for 7.5/4.1 and 7.6/4.2, as well as 8.0b2.)
Coding style change: instead of ``func (args)'', write ``func(args)''
everywhere.
Minor functionality change: use PyArg_ParseTuple everywhere. This
should only affect the errors reported for bad argument lists; in
particular, deletefilehandler() is much clearer about what's going
on.
Extend the "Don Beaudry hack" with "Guido's corollary" -- if the base
class has a __class__ attribute, call that to create the new class.
This allows us to write metaclasses purely in C!
Moved the special compile of getbuildno.o to ../Makefile.in.
A dummy getbuildno.o (with a number of 0) still gets built here,
to make the library complete.
Build getbuildno.o here, to adequately update it every time a new
python executable is built. (It still won't reflect builds of the
library only, but since the default make target builds the python
executable, that's alright.)
Tweaks by Lars Wirzenius to parse some more forms of illegal dates:
the comma after the day name is optional if it is a recognized day
name; and the date and month may be swapped. Thus, the following two
test dates will now be parsed correctly:
Thu, Feb 13 12:16:57 1992
Thu Feb 13 12:16:57 1992
Fred Drake [Fri, 25 Jul 1997 13:14:35 +0000 (13:14 +0000)]
Bump up the version number. The date still needs to be set (Guido will
have to do that one!), but at least the version will match the release
version. (1.5a2 got skipped for this.)
Fred Drake [Thu, 24 Jul 1997 15:39:16 +0000 (15:39 +0000)]
A couple of grammatical nits.
Re-sequenced the function descriptions so that the formatting is described
before the assumption is made that the reader has a clue about the formatting.
Moved examples to be closer to the functions being demonstrated.
Removed some variables that are used to exchange data between import.c and
importdl.c: the MAXSUFFIXSIZE macro is now defined in importdl.h, and
the modules dictionary is now passed using PyImport_GetModuleDict().
Also undefine USE_SHLIB for AIX -- in AIX 4.2 and up, dlfcn.h exists
but we don't want to use it.
projects / python / log