Changes by Eryk Sun :
--
dependencies: +os.walk always follows Windows junctions
___
Python tracker
<http://bugs.python.org/issue29250>
___
___
Python-bugs-list m
Eryk Sun added the comment:
I simply listed the tags that have the name-surrogate bit set out of those
defined in km\ntifs.h.
To keeps things simple it might be better to only include Microsoft tags (i.e.
bit 31 is set). That way we don't have to deal with REPARSE_GUID_DATA_BUFFER
s
Eryk Sun added the comment:
In 3.6, type_new in Objects/typeobject.c sets the __classcell__ in the dict if
it's a cell object. It happens that CreateSwappedType in
Modules/_ctypes/_ctypes.c re-uses the dict to create the swapped type (e.g. big
endian), which in turn updates the __classc
Eryk Sun added the comment:
Here's a patch that deletes __classcell__ from the dict before calling type_new.
--
keywords: +patch
Added file: http://bugs.python.org/file46285/issue_29270_01.patch
___
Python tracker
<http://bugs.python.org/is
Changes by Eryk Sun :
--
stage: -> patch review
___
Python tracker
<http://bugs.python.org/issue29270>
___
___
Python-bugs-list mailing list
Unsubscrib
Eryk Sun added the comment:
OK, this is completely broken and needs a more thoughtful solution than my
simpleminded hack. Here's a practical example of the problem, tested in 3.5.2:
class MyInt(ctypes.c_int):
def __repr__(self):
return super().__repr__()
Eryk Sun added the comment:
Resolving this would be straightforward if we could use a subclass for the
swapped type, but ctypes simple types behave differently for subclasses. A
simple subclass doesn't automatically call the getfunc to get a converted value
when returned as a function r
Eryk Sun added the comment:
> it should be replaced with sys.getfilesystemencodeerrors()
> to support UTF-8 Strict mode.
I did that in the patch for issue 28188. The focus of the patch is to add bytes
support on Windows for os.putenv and os.environb, but I also tried to maximize
consi
Eryk Sun added the comment:
Some WinSock functions are just dispatchers that call a provider function. The
dispatch table is set up when WinSock (i.e. ws2_32.dll) calls the WSPStartup
function [1] of a provider DLL (e.g. mswsock.dll). In the case of select(), it
calls the socket provider&#
Eryk Sun added the comment:
Craig, can you add a patch for issue 29248, including a test based on the "All
Users" link?
--
dependencies: +os.readlink fails on Windows
___
Python tracker
<http://bugs.python.o
Changes by Eryk Sun :
--
stage: -> needs patch
versions: +Python 3.6, Python 3.7 -Python 3.3, Python 3.4
___
Python tracker
<http://bugs.python.org/issu
Changes by Eryk Sun :
--
Removed message: http://bugs.python.org/msg226094
___
Python tracker
<http://bugs.python.org/issue22302>
___
___
Python-bugs-list mailin
Eryk Sun added the comment:
isabs also fails for device paths such as r"\\.\C:", which is an absolute path
for opening the C: volume. UNC and device paths (i.e. \\server, \\?, and \\.)
should always be considered absolute. Only logical drives (i.e. C:, D:, etc)
support drive-rela
Changes by Eryk Sun :
--
resolution: -> duplicate
stage: -> resolved
superseder: -> In 2.7.13, _ctypes.LoadLibrary no longer accepts Unicode objects
___
Python tracker
<http://bugs.python.or
Eryk Sun added the comment:
> they do not appear in the byte code files
It's simple to coonfirm that unassigned string literals get thrown away.:
>>> code = compile('"doc"\n"unused"\n"us"+"ed"', '', 'exec&#
Eryk Sun added the comment:
Python 3 uses __index__ for bin(), oct(), and hex(), but Python 2 only uses
__index__ for bin() and otherwise uses __oct__ and __hex__. Antoine overlooked
this when updating the 2.7 docs for hex() in issue 16665.
--
nosy: +eryksun
Eryk Sun added the comment:
Are you suggesting that the Helper class in Lib/pydoc.py should index
[sub-]topics by special names such as "__slots__"? Currently you can see the
__slots__ documentation via the topics "ATTRIBUTEMETHODS" and "SPECIALMETHODS".
The lis
Eryk Sun added the comment:
Change the last print to test the exit code, as follows:
if p.exitcode != 0:
print('Child failed with exit code:', p.exitcode)
else:
print(num.value)
print(arr[:])
Note that when you fail to limit creating a new process t
Eryk Sun added the comment:
>if p.exitcode != 0:
>print('Child failed with exit code:', p.exitcode)
Interestingly the exit code is always 1 when running under pythonw.exe. I just
inspected what's going on by setting sys.stderr to a file and discovered
Eryk Sun added the comment:
Python appends "\*.*" to the path, so in the case in which the leaf element is
a file, FindFirstFile naturally fails with ERROR_PATH_NOT_FOUND.
Python 3.5+ could maybe switch to calling CreateFile to open a handle and
GetFileInformationByHandleEx
Eryk Sun added the comment:
> FindFirstFile naturally fails with ERROR_PATH_NOT_FOUND
Getting this error actually depends on the file system. I don't see it with
NTFS, which returns STATUS_NOT_A_DIRECTORY, which gets translated to
ERROR_DIRECTORY. On the other hand, VboxSF (VirtualBo
Eryk Sun added the comment:
Python has no specific use for ERROR_PATH_NOT_FOUND (3) and just maps it to
FileNotFoundError -- like the CRT maps it to ENOENT. Even if we wanted
listdir() to specially handle this error, given that it's implemented via
FindFirstFile, there's nothing t
Eryk Sun added the comment:
os.symlink calls CreateSymbolicLink, which creates the reparse data buffer with
the print name stored first, so the offset is always 0. Otherwise we would have
noticed this problem already. For example:
>>> os.symlink('C:\\', 'link&#
Changes by Eryk Sun :
--
resolution: -> third party
stage: -> resolved
status: open -> closed
___
Python tracker
<http://bugs.python.org/issue29391>
___
___
Eryk Sun added the comment:
The old CRT doesn't do any parameter validation on the nbytes parameter. It
just passes it directly to Windows LockFile as follows:
LockFile((HANDLE)_get_osfhandle(fh), lockoffset, 0L, nbytes, 0L)
which is locking (DWORD)-1 bytes, i.e. 0x. This a
Changes by Eryk Sun :
--
stage: -> needs patch
versions: +Python 3.6, Python 3.7
___
Python tracker
<http://bugs.python.org/issue29392>
___
___
Python-bugs-lis
Eryk Sun added the comment:
I had reopened this issue with a suggestion for expanding the supported paths
in msg276864, but it's languished for a while now. I've attached a patch
implementing the change to _PyIO_get_console_type that I had suggested. Here
are some example paths
Eryk Sun added the comment:
I added some tests to ensure open() returns an instance of _WindowsConsoleIO
for a few console paths. It also checks that opening a non-console file raises
a ValueError, both for a path and an fd, as well for a negative file descriptor.
--
Added file: http
Changes by Eryk Sun :
Removed file: http://bugs.python.org/file46462/issue_28164_02.patch
___
Python tracker
<http://bugs.python.org/issue28164>
___
___
Python-bugs-list m
Changes by Eryk Sun :
Added file: http://bugs.python.org/file46463/issue_28164_02.patch
___
Python tracker
<http://bugs.python.org/issue28164>
___
___
Python-bugs-list m
Eryk Sun added the comment:
LoadLibraryEx with LOAD_WITH_ALTERED_SEARCH_PATH worked for me in Windows 10. I
tested loading 64-bit python3.dll, for both 3.5 and 3.6, and getting the
address of Py_Main. The DLLs were loaded from DLL35 and DLL36 subdirectories.
It worked using both a relative
Eryk Sun added the comment:
> Are you sure it really works on Windows 10?
Neither the DLL35 nor DLL36 directory was set in PATH, nor any directory
containing python3x.dll. Each of the latter two directories contained only the
minimal set python3.dll, python3x.dll, and vcruntime140.dll. I
Eryk Sun added the comment:
I can confirm that LoadLibraryEx w/ LOAD_WITH_ALTERED_SEARCH_PATH unfortunately
does not work in Windows 7. Actually I almost looked into this before on
python-list:
https://mail.python.org/pipermail/python-list/2016-September/thread.html#714622
but I got
New submission from Eryk Sun:
PEP 529 isn't implemented for io.FileIO, and I think it should be. If a UTF-8
path is passed to open(), it ends up calling C _open instead of decoding the
path and calling C _wopen. Also, if a pathlike object is passed to io.FileIO,
it calls PyUnicode_FSConv
Changes by Eryk Sun :
--
keywords: +patch
Added file: http://bugs.python.org/file46473/issue_29409_01.patch
___
Python tracker
<http://bugs.python.org/issue29
Changes by Eryk Sun :
--
stage: needs patch -> test needed
___
Python tracker
<http://bugs.python.org/issue29409>
___
___
Python-bugs-list mailing list
Un
Eryk Sun added the comment:
This case is similar to issue 29079. It should only attempt to make the parent
to handle ENOENT if self and self.parent aren't equal. Here's the snippet from
Path.mkdir:
try:
self._accessor.mkdir(self, mode)
except FileExistsError:
Eryk Sun added the comment:
I'm closing this as a third-party issue that's unrelated to the development of
CPython.
PyWin32 is a package of extension modules for accessing Windows C and COM APIs
in Windows Python. You can't install it in Linux Python. You should be able to
ac
Eryk Sun added the comment:
> difference from the behaviour of Posix's execvpe() was deliberate
POSIX doesn't define execvpe [1]. GNU glibc implemented it in 2009 [2]. On
Windows, MSC has had execvpe and spawnvpe since at least 5.0 [3], and I think
it arrived in 4.0 in 1986. G
Eryk Sun added the comment:
Sure, let's close this in favor of the older documentation issue 8557.
--
resolution: -> duplicate
stage: needs patch -> resolved
status: open -> closed
superseder: -> subprocess PATH semantics and portability
_
Changes by Eryk Sun :
--
versions: +Python 3.5, Python 3.6, Python 3.7 -Python 2.6, Python 3.1, Python
3.2
___
Python tracker
<http://bugs.python.org/issue8
Eryk Sun added the comment:
The Unix implementation of subprocess.Popen follows the behavior of os.execvpe,
which is an outlier. Other execvpe implementations, such as the one added to
glibc in 2009, search PATH in the current environment instead of the passed
environment. As such, and given
Eryk Sun added the comment:
The user is presented with 3 buttons: Back, Next, and Cancel. The "Next" button
is the only way forward to modify the installation. Did you assume "Next" would
start the modification rather than take you to another dialog with options? How
ab
Eryk Sun added the comment:
It's an ugly inconsistency that GetFullPathName fails for bare CONIN$ and
CONOUT$ prior to Windows 8, in that it gives a different result from simply
passing those names to CreateFile. Anyway, thanks for modifying it to work
correctly in this case.
We should
Eryk Sun added the comment:
As Martin said, you need to set the LC_TIME category using an empty string to
use the locale LC_* environment variables. Python 3 sets LC_CTYPE at startup
(on Unix platforms only), but LC_TIME is left in the initial C locale:
>>> locale.
Eryk Sun added the comment:
> it's not possible to tell by inspection the version of a Python
> interpreter.
If getting the version of python[w].exe is ever required, it should be simple
for 3.5+, for which python[w].exe has standard file version information with
the product v
Eryk Sun added the comment:
If the system doesn't have a "python3.exe" on PATH, then "env python3" will run
a registered version, exactly like it currently does. IMO, this is slightly
less confusing than the current behavior, which skips searching PATH in this
ca
Eryk Sun added the comment:
The macro is defined but not defined. If I insert the following line before the
#ifdef check:
#define IPPROTO_IPV6 IPPROTO_IPV6
then the constant gets added:
>>> import _socket
>>> _socket.IPPROTO_IPV6
41
The same applies t
Eryk Sun added the comment:
Unless someone has a better (more automated) way to handle the Winsock IPPROTO
enum, I suggest we either special-case the individual tests for MS_WINDOWS when
we know that Winsock defines the value; or just define macros for the values in
the enum:
#ifdef
Eryk Sun added the comment:
It doesn't appear to be a crash, but seems like the REPL is quitting because
there's no input (e.g. like stdin is redirected to NUL). Echo %errorlevel%;
it's probably 0.
readline (pyreadline) is probably the culprit. Try the following she
Eryk Sun added the comment:
For the tokenizer, a blank line is "[a] logical line that contains only spaces,
tabs, formfeeds and possibly a comment" [1]. A blank line is normally ignored,
except in the REPL an entirely blank line (i.e. no whitespace or comment) is
used to end a
Changes by Eryk Sun :
--
priority: normal -> high
versions: +Python 3.5, Python 3.6, Python 3.7
___
Python tracker
<http://bugs.python.org/issue22273>
___
_
Changes by Eryk Sun :
--
stage: -> needs patch
___
Python tracker
<http://bugs.python.org/issue22273>
___
___
Python-bugs-list mailing list
Unsubscrib
Eryk Sun added the comment:
Our support of passing structs and unions by value is really bad. Passing this
particular struct by value is also broken on 64-bit Unix, but instead because
it's small. It crashes Python due to an abort(). See issue 22273.
--
nosy: +eryksun
priority: n
Eryk Sun added the comment:
ctypes defines arrays as a pointer FFI type because they degenerate as pointers
in C calls. But it's generally wrong to set a pointer FFI type for an array in
the `elements` of a struct's FFI type. An integer array in a struct that's 16
bytes or
Eryk Sun added the comment:
gethostbyname_ex won't do a reverse lookup on an IP to get the fully-qualified
domain name, which seems pointless for a function named getfqdn. I think
calling gethostbyaddr is intentional here and goes back to the Python 1.x days.
Also, FYI, socket_gethostb
Eryk Sun added the comment:
To install KB3118401 [3], you may first need to install KB2919442 [1] and
KB2919355 [2]. The installation order would be as follows:
KB2919442:
Windows8.1-KB2919442-x64.msu
KB2919355:
clearcompressionflag.exe
Windows8.1-KB2919355-x64.msu
Windows8.1
Eryk Sun added the comment:
Please provide the complete traceback. Copy and paste it from the command
prompt. Also, include the fully-qualified path for what "python" runs, e.g. run
`where python` in the command prompt. Also, include the value of the
environment variables PYTH
Eryk Sun added the comment:
That error looks like __init__.py[c] was left in
"C:\python27\lib\site-packages\pip". Just to be sure, since I don't know what
you did to clear the value, please run `set PYTHONPATH=` to clear the value for
the current shell. Then try `python -m
Eryk Sun added the comment:
That's good news. As far as I can see, there isn't anything Python's installer
can do if the CRT update fails for reasons beyond our control, so I'm closing
this as a 3rd party issue. Thank you for the report.
--
resolution:
Eryk Sun added the comment:
pip should be installed in site-packages, which for you is
"C:\Python\Python35-32\Lib\site-packages". Try manually running ensurepip to
make sure pip is installed:
python -m ensurepip --verbose --upgrade --d
Eryk Sun added the comment:
If you have the 3.5 installation logs in your %TEMP% directory, please zip them
up and attach them to this issue.
--
___
Python tracker
<http://bugs.python.org/issue29
Eryk Sun added the comment:
This issue is more straight-forward than #22273. ISTM, we just have to copy
large values. For example, in ffi_call, we'd iterate over the arguments and
alloca and memcpy the large values prior to calling ffi_call_AMD64:
case FFI_SYSV:
/* If a s
Eryk Sun added the comment:
The original install failed with the status code STATUS_DLL_NOT_FOUND
(0xC135), as reported in "Python 3.5.3
(32-bit)_20170217102132_008_pip_JustForMe.log":
MSI (s) (88:D8) [10:22:59:574]: Executing op: ActionStart(Name=UpdatePip,,)
MSI (s) (
Eryk Sun added the comment:
Structs that are larger than 32 bytes get copied to the stack (see
classify_argument in ffi64.c), so we don't have to worry about classifying
their elements for register passing. Thus if a new field is added for this in
StgDictObject, then PyCArrayType_new s
Eryk Sun added the comment:
classify_argument is in Modules/_ctypes/libffi/src/x86/ffi64.c. This file is
for the 64-bit Unix ABI. libffi doesn't use it for 64-bit Windows.
Some (all?) Linux distros link ctypes to the system libffi. In my experience,
building 3.6 on Linux defaults t
Eryk Sun added the comment:
I see now why you couldn't find ffi64.c. I've been using a 3.6 worktree. The
libffi sources have been removed from master.
For the union and bitfield problem, also see the crash reported in #26628.
--
___
Pyth
Eryk Sun added the comment:
The 24-byte struct gets passed on the stack, as it should be. In this case
ffi_call doesn't abort() because examine_argument returns 0, which is due to
the following code in classify_argument:
if (words > 2)
{
/* When size > 16 bytes, i
Eryk Sun added the comment:
Notes on fix-22273-02.diff:
In the second pass over _fields_, you can (should) use dict->length and
dict->proto for array types instead of the _length_ and _type_ attributes.
When reassigning stgdict->ffi_type_pointer.elements, if
use_broken_old_ctypes_
Eryk Sun added the comment:
> Perhaps you mean 16 rather than 8?
Sorry, that was a misfire. It should be 16.
--
___
Python tracker
<http://bugs.python.org/issu
Eryk Sun added the comment:
I had suggested inheriting the TYPEFLAG_NONARGTYPE flag in StructUnionType_new.
It requires a minor change to get basedict unconditionally, and then assign
if (basedict)
dict->flags |= basedict->flags & TYPEFLAG_NONARGTYPE;
We need more f
Eryk Sun added the comment:
> I'm not sure using this flag impacts on consistency with CFFI
I meant consistency with respect to supported argument types. If someone
contributed a workaround to CFFI, then I would rather port it, but it looks
like they're also waiting for this to
Eryk Sun added the comment:
Perhaps it should instead use two specific flags, TYPEFLAG_HASBITFIELD and
TYPEFLAG_HASUNION, which are propagated unconditionally from the base class and
fields. As a base case, a union itself is flagged TYPEFLAG_HASUNION. Arrays are
unrolled on X86_64 only if
Eryk Sun added the comment:
Nick wrote:
> 1. Open the file in the current process
> 2. Write content to the file
> 3*. Close the file in the current process
In step 1, do you mean calling NamedTemporaryFile with delete=False? In that
case there's no immediate problem with op
Eryk Sun added the comment:
Richard wrote:
> while a handle is open with share mode X, you can only reopen
> the file if you also use share mode X
To clarify, the share mode is not a property of a handle. It's a property of a
File object. A handle is a generic reference to any kin
Eryk Sun added the comment:
The error message is misleading. It happens that WiX is trying to run an
elevated process (see the WiX functions CoreLaunchApprovedExe, CoreElevate,
ElevationElevate, and PipeLaunchChildProcess). However, the actual error code
has nothing to do with elevation
Eryk Sun added the comment:
All types are instances of `type`, so the single argument case makes sense to
me as a 'constructor'. It always returns an instance of `type`, just not a new
instance.
>>> X = type('X', (type,), {})
>>> type(X)
Eryk Sun added the comment:
> why write `metatype == &PyType_Type` rather than
> PyType_CheckExact(metatype)`?
If only `type` should implement this special case, then it needs to be
`metatype == &PyType_Type`. This was actually how it was implemented in 2.2a3:
https://hg.pyth
Eryk Sun added the comment:
> Please treat this as a new feature (just in case) and only
> apply it to 3.6.
How about changing PyType_CheckExact to PyType_Check for 2.7 and 3.5? It solves
the original problem by expanding the single-argument case to metaclasses that
aren't an exa
Changes by Eryk Sun :
--
Removed message: http://bugs.python.org/msg266710
___
Python tracker
<http://bugs.python.org/issue27157>
___
___
Python-bugs-list mailin
Eryk Sun added the comment:
There is no right encoding as far as I can see.
If it's attached to a console (i.e. conhost.exe), then cmd.exe uses the
console's output codepage when writing to a pipe or file, which is the scenario
that your patch attempts to address. But i
Eryk Sun added the comment:
> ``sizeof(long double) == sizeof(double)`` it is an alias to
> :class:`c_double`.
This should be "``sizeof(long) == sizeof(int)`` ... :class:`c_long`".
--
resolution: fixed ->
stage: resolved ->
New submission from Eryk Sun:
nt is the module name for posixmodule.c on Windows, so I'm changing the title
to reference ntpath instead.
Regarding nt, two of its built-in functions are exposed directly in ntpath.
_isdir is imported as isdir, so it needs a wrapper to support __fspath__.
Eryk Sun added the comment:
Sorry, I must have missed or misunderstood something. PEP 519 discusses
modifying os.path. For the os module it only discusses adding fspath and
updating fsencode and fsdecode. It also discusses a new PyOS_FSPath C API, but
without any discussion regarding its use
Eryk Sun added the comment:
> I would say almost all Windows console programs does use
> console's encoding for input/output because otherwise
> user wouldn't be able to read it.
While some programs do use the console codepage, even when writing to a pipe or
disk file
Changes by Eryk Sun :
--
Removed message: http://bugs.python.org/msg267090
___
Python tracker
<http://bugs.python.org/issue27179>
___
___
Python-bugs-list mailin
Eryk Sun added the comment:
> I would say almost all Windows console programs does use
> console's encoding for input/output because otherwise
> user wouldn't be able to read it.
While some programs do use the console codepage, even when writing to a pipe or
disk file
Eryk Sun added the comment:
>> so ANSI is the natural default for a detached process
>
> To clarify - ANSI is the natural default *for programs that
> don't support Unicode*.
By natural, I meant in the context of using GetConsoleOutputCP(), since
WideCharToMultiByte(0,
Eryk Sun added the comment:
Another set of counterexamples are the utilities in the GnuWin32 collection,
which use ANSI in a pipe:
>>> call('chcp.com')
Active code page: 437
0
>>> 'ยก'.encode('1252')
b'\xa1'
Eryk Sun added the comment:
> so if we default to UTF-8 it will be even worse than
> defaulting to ANSI because there aren't many programs
> on Windows which would use UTF-8
I didn't say subprocess should default to UTF-8. What I wish is for Python to
default to using UTF
Eryk Sun added the comment:
This affects all Tkinter applications on Windows. For example, in Python 2.7:
import Tkinter
import ctypes
libc = ctypes.CDLL('msvcr90')
libc.getenv.restype = ctypes.c_char_p
>>> libc.getenv(b"HOME&quo
Eryk Sun added the comment:
The universal CRT that's used by 3.5+ implements locales using Windows locale
names [1], which were introduced in Vista. Examples for Spanish include 'es',
'es-ES', and 'es-ES_tradnl' . The latter is the traditional sort
Eryk Sun added the comment:
> The message comes from Objects/floatobject.c:183
No, in this case the error is set in PyOS_string_to_double in
Python/pystrtod.c, because `fail_pos == s`, and it doesn't get replaced in
PyFloat_FromString because `end == last`. The format s
Eryk Sun added the comment:
Probably adding from_pointer is a good idea. That said, there's a simpler way
to go about getting a bytes copy for a pointer.
Say that you have a pointer p for the following array:
>>> a = (c_float * 3)(1, 2, 3)
>>> bytes(a)
b&#x
Changes by Eryk Sun :
--
resolution: -> third party
stage: -> resolved
status: open -> closed
___
Python tracker
<http://bugs.python.org/issue27264>
___
___
Changes by Eryk Sun :
--
Removed message: http://bugs.python.org/msg268100
___
Python tracker
<http://bugs.python.org/issue27284>
___
___
Python-bugs-list mailin
Changes by Eryk Sun :
--
resolution: -> not a bug
stage: -> resolved
status: open -> closed
___
Python tracker
<http://bugs.python.org/issue27284>
___
Eryk Sun added the comment:
I get an access violation due to calling the ASN1_ITEM pointer as a function:
> 810: ASN1_ITEM_ptr(method->it)));
_ssl!_get_peer_alt_names+0x12a:
7ffe`17b1225a ffd0call
Eryk Sun added the comment:
os.symlink needs the target_is_directory argument on Windows. Maybe extract()
can search for the link target in namelist() to determine if it's a directory
(i.e. ends in "/"). Note that the flag gets set automatically if the target
exists, so this is
Eryk Sun added the comment:
vcvarsall.bat is mostly `set` and `echo` statements, which print using UTF-16LE
with "/u". It also runs reg.exe, but with stdout and stderr redirected to nul,
so that's no problem. The final `set` command prints cmd's UTF-16LE
environment. U
1701 - 1800 of 2119 matches
Mail list logo
