Fred Drake [Wed, 15 Dec 1999 18:31:10 +0000 (18:31 +0000)]
Rip out the code to check the ordering of the tables used to map
strings to integers for the *conf*() functions.
Added code to sort the tables at module initialization. Three
dictionaries, confstr_names, sysconf_names, and pathconf_names, are
added to the module as well. These map known configuration setting
names to the numeric value which is used to represent the setting in
the system call. This code is always called.
Fred Drake [Wed, 15 Dec 1999 15:34:33 +0000 (15:34 +0000)]
Added table entries for Irix 6.5 names for confstr()/sysconf()/
pathconf() names, from Sjoerd.
Added code to verify that these tables are properly ordered, only
included and used when CHECK_CONFNAME_TABLES is defined. This is only
needed to test the tables, so I haven't enabled this by default.
Fred Drake [Tue, 14 Dec 1999 21:25:03 +0000 (21:25 +0000)]
Added support for getlogin(); does *not* use getlogin_r() where
available since the interface is poorly defined on at least one major
platform (Solaris).
Moved table of constant names for fpathconf() & pathconf() into the
conditional that defines the conv_path_confname() helper; Mark Hammond
reported that defining the table when none of the constants were
defined causes the compiler to complain (won't allow 0-length array,
imagine that!).
In posix_fpathconf(), use conv_path_confname() as the O& conversion
function, instead of the conv_confname() helper, which has the wrong
signature (posix_pathconf() already used the right thing).
Fred Drake [Tue, 14 Dec 1999 21:11:47 +0000 (21:11 +0000)]
Remove test for getlogin_r(); the interface is not clearly defined, at
least on Solaris (sometimes it's Unix98, sometimes it conforms to an
early draft).
Properly generate config.h.in using autoheader instead of editing it
manually; thanks, Guido!
Guido van Rossum [Mon, 13 Dec 1999 23:27:45 +0000 (23:27 +0000)]
V 2.16 from Piers:
I've changed the login command to force proper
quoting of the password argument. I've also added
some extra debugging code, which is removed when
__debug__ is false.
Jack Jansen [Mon, 13 Dec 1999 16:04:48 +0000 (16:04 +0000)]
Methods {Get,Set}ControlData that know about data types passed for the various
4-char codes. The table which maps codes to datatypes is still pretty empty,
I'll fill it as I need entries (or maybe someone wants to spend a nice day filling it?).
Jack Jansen [Sun, 12 Dec 1999 22:57:51 +0000 (22:57 +0000)]
Updated Message, question, YNC and progress dialogs for Appearance. Password
remains to be done, and maybe we should convert everything to movablemodal too
and make things like question look nicer.
Jack Jansen [Sun, 12 Dec 1999 22:57:29 +0000 (22:57 +0000)]
In places where a ResObj is expected for PyArg_Parse and the object passed in isn't but it does have an as_Resource method use that. This makes life a lot easier
for appearance portability (and was needed anyway).
Jack Jansen [Sun, 12 Dec 1999 21:56:56 +0000 (21:56 +0000)]
Regenerated toolbox modules with CW Pro 5.2, which has MacOS 8.6 and Appearance 1.1 support, and added workaround for non-existing calls for 68K.
Static 68K Python still has a tk problem.
Jack Jansen [Sun, 12 Dec 1999 21:38:29 +0000 (21:38 +0000)]
Dummy functions which return unimpErr for lots of toolbox calls that are unavailable for 68K. This way we don't have to disable thm in bgen, so we
can still use them on PPC.
Greg Ward [Sun, 12 Dec 1999 17:07:22 +0000 (17:07 +0000)]
Catch missing MANIFEST file and warn rather than blowing up.
Added 'nuke_release_tree()' method to blow away the directory from
which the archive file(s) are created, and call it (conditionally)
from 'make_distribution()'.
Added 'keep_tree' option (false by default) to disable the call to
'nuke_release_tree()'.
Greg Ward [Sun, 12 Dec 1999 17:03:59 +0000 (17:03 +0000)]
Fixed 'find_package_modules()' to ensure that we never build (and thus
install) the setup script itself.
Fixed 'build_module()' so we do *not* preserve file mode (which means
we can install read-only files, which makes the next installation
of this distribution fail -- at least under Unix); added a comment
explaining this.
Greg Ward [Sun, 12 Dec 1999 17:01:01 +0000 (17:01 +0000)]
Changed 'build_extensions()' so 'sources' can be a list or tuple; and
call CCompiler method 'compile()' with 'include_dirs' not 'includes'.
Fixed stupid typo in 'get_source_files()'.
Greg Ward [Sun, 12 Dec 1999 16:57:47 +0000 (16:57 +0000)]
In 'compile()' method, renamed 'includes' parameter to 'include_dirs' for
consistency with 'build_ext' command option.
Changed 'compile()' and 'link_shared_object()' so 'include_dirs',
'libraries', and 'library_dirs' can be lists or tuples.
Greg Ward [Sun, 12 Dec 1999 16:54:55 +0000 (16:54 +0000)]
Added support for printing out help text from option table: 'print_help()',
'generate_help()', 'wrap_text()' functions, and a little tiny test
of 'wrap_text()'.
Changed how caller states that one option is the boolean opposite of
another: added 'negative_opt' parameter to 'fancy_getopt()', and changed
to use it instead of parsing long option name.
Greg Ward [Sun, 12 Dec 1999 16:51:44 +0000 (16:51 +0000)]
Made "verbose" mode the default; now you have to supply --quiet if you
want no output. Still no option for a happy medium though.
Added "--help" global option.
Changed 'parse_command_line()' to recognize help options (both for the
whole distribution and per-command), and to distinguish "regular run"
and "user asked for help" by returning false in the latter case.
Also in 'parse_command_line()', detect invalid command name on command
line by catching DistutilsModuleError.
a 'negative_opt' class attribute right after 'global_options'; changed
how we call 'fancy_getopt()' accordingly.
Initialize 'maintainer' and 'maintainer_email' attributes to Distribution
to avoid AttributeError when 'author' and 'author_email' not defined.
Initialize 'help' attribute in Command constructor (to avoid
AttributeError when user *doesn't* ask for help).
In 'setup()':
* show usage message before dying when we catch DistutilsArgError
* only run commands if 'parse_command_line()' returned true (that
way, we exit immediately when a help option is found)
* catch KeyboardInterrupt and IOError from running commands
Bulked up usage message to show --help options.
Comment, docstring, and error message tweaks.
Fred Drake [Thu, 9 Dec 1999 21:13:07 +0000 (21:13 +0000)]
Added support for abort(), ctermid(), tmpfile(), tempnam(), tmpnam(),
and TMP_MAX.
Converted all functions that used PyArg_Parse() or PyArg_NoArgs() to
use PyArg_ParseTuple() and specified all function names using the
:name syntax in the format strings, to allow better error messages
when TypeError is raised for parameter type mismatches.
Jack Jansen [Tue, 7 Dec 1999 23:08:10 +0000 (23:08 +0000)]
First bits and pieces of appearance support: an init routine, a global flag PyMac_AppearanceCompliant (exported thru MacOS). If USE_APPEARANCE
is off the code is disabled (but the variables are still there, set to 0).
OpenSSL support. This is based on patches for a version of SSLeay by
Brian E Gallew, which were improved and adapted to OpenSSL 0.9.4 by
Laszlo Kovacs of HP. Both have kindly given permission to include
the patches in the Python distribution. Final formatting by GvR.
OpenSSL support. This is based on patches for a version of SSLeay by
Brian E Gallew, which were improved and adapted to OpenSSL 0.9.4 by
Laszlo Kovacs of HP. Both have kindly given permission to include
the patches in the Python distribution. Final formatting by GvR.
Jack Jansen [Fri, 3 Dec 1999 23:38:05 +0000 (23:38 +0000)]
Set the finder "is shared" bit, by request of Joe Strout. It seems this allows multiple simultaneous copies to be run from a server, and Applets shouldn't
write their datafork so it appears safe.
Greg Ward [Fri, 3 Dec 1999 16:18:56 +0000 (16:18 +0000)]
[from 1999-11-04]
Bunch of little bug fixes that appeared in building non-packagized
distributions. Mainly:
- brain-slip typo in 'get_package_dir()'
- don't try to os.path.join() an empty path tuple -- it doesn't like it
- more type-safety in 'build_module()'
Jack Jansen [Fri, 3 Dec 1999 16:08:50 +0000 (16:08 +0000)]
Default schedparams set to (0,0): no event-intervention by Python mainloop. This
was always meant to be the default, and the new example-2 was pretty critical
of this.
Jack Jansen [Thu, 2 Dec 1999 22:52:12 +0000 (22:52 +0000)]
Give a warning if system-wide sys.argv processing is off (because then we
may have missed an applet being dropped on us). This may fix the obscure bug
Tony Ingraldi was experiencing.