can cause it to get called by multiple threads simultaneously.
Ditto for PyInterpreterState_Delete.
Of the former, the docs say "The interpreter lock need not be held, but may
be held if it is necessary to serialize calls to this function". This
kinda implies it both is and isn't thread-safe.
Of the latter, the docs merely say "The interpreter lock need not be
held.", and the clause about serializing is absent.
I expect it was *believed* these are both thread-safe, and the bit about
serializing via the global lock was meant as a permission rather than a
caution.
I also expect we've never seen a problem here because the Python core
(prior to the _PyPclose fix) only calls these functions once per run.
The Py_NewInterpreter subsystem exposed by the C API (but not used by
Python itself) also calls them, but that subsystem appears to be very
rarely used.
Whatever, they're both thread-safe now.
newlines at the start or end. Fiddle test_popen2 and popen2._test() to
tolerate this. Also change all "assert"s in these tests to raise
explicit exceptions, so that python -O doesn't render them useless.
Also, in case of error, make the msg display the reprs of what we
wrote and what we read, so we can tell exactly why it's failing.
Fix import support to work with import as variant of Python 2.0. The
grammar for import changed, requiring changes in transformer and code
generator, even to handle compilation of imports with as.
very late in the process when running on Windows 2000 without admim
privs. Rearranged so that the admin check is done at the start
instead. Added words to the end of the blurb to make it very clear
how to abort the install (wasn't obvious to me that "Cancel" was
the right thing to click).
Py_FatalError() from module initialization functions. The importing
mechanism already checks for PyErr_Occurred() after module importation
and it Does The Right Thing.
Unfortunately, the following either were not compiled or tested by the
regression suite, due to issues with my development platform:
almodule.c
cdmodule.c
mpzmodule.c
puremodule.c
timingmodule.c
by Martin v. Loewis, proofed by Barry Warsaw for coding standards,
typos, and to make command line options compatible with GNU msgfmt
where they overlap.
Closes patch #101295.
Python test suite. Specifically,
- import time instead of strop in test_b1
- test for ClassType of exceptions using isinstance instead of
equality in test_exceptions
- remove __builtins__ from dir() output in test_pkg
test_pkg output needs to be regenerated.
implementation. You don't want to know. I've asked Guido to give this
a critical review (we agreed on the approach, but the implementation
proved more ... interesting ... than anticipated). This will almost
certainly be the highlight of Mark Hammond's day <wink>.
#101187, which some modifications. Specifically,
ntransfercmd(), transfercmd(), and retrbinary() all grow an optional
`rest' argument, which if not None, is used as the argument to an FTP
REST comman dbefore the socket is returned. Differences from the SF
patch:
- always compare against None with `is' or `is not' instead of == or !=
- no parens around conditional
- RFC 959 defines the argument to REST is a string containing any
ASCII characters in the range [33..126]. Therefore, we use the %s
format character instead of %f or %d as suggested in the patch's
comments. Note that we do /not/ sanity checkthe contents of the
rest argument (but we'll document this in the library reference
manual).
Guido van Rossum