[issue2833] __exit__ silences the active exception
Antoine Pitrou [EMAIL PROTECTED] added the comment: Small typo in the snippet above, this should obviously read: try: raise Exception(foo) except Exception as e: try: raise KeyError(caught) except KeyError: pass raise e __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2833 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2833] __exit__ silences the active exception
Antoine Pitrou [EMAIL PROTECTED] added the comment: As Amaury said, lexically nested exception handlers make re-raising behaviour buggy. In Py3k, a workaround is to instead write: try: raise Exception(foo) except Exception as : try: raise KeyError(caught) except KeyError: pass raise e With the slight inconvenience that the raise e line will be appended to the original traceback. If we want bare raise statements to work as expected after a nested exception handler, we'll need to add proper exception stacking, for example by adding the exception value as a member of PyTryBlock. The effect on performance should also be measured. __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2833 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue1467929] %-formatting and dicts
Marc-Andre Lemburg [EMAIL PROTECTED] added the comment: I guess the patch was just forgotten after the 2.5 release was out. I've added a 2.6 tag and assigned the patch to Sean. -- assignee: anthonybaxter - jafo versions: +Python 2.6 _ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue1467929 _ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2850] Augmenting the Windows build to support code signing.
New submission from Trent Nelson [EMAIL PROTECTED]: Now that we've finally got a VeriSign code-signing certificate (hurrah!), we can look at how we want to integrate the aspect of code signing into our build process. I'd like to propose augmenting the build process such that as a post- link step for .exe|.pyd targets, a little helper Python code signing script will be run (i.e. trunk/pcbuild/signfile.py). The first thing that this script will check is whether or not the environment variable PYTHON_CODESIGNING_CERT_NAME is defined. If it is not defined or is empty, the script will simply exit. This will be the case for the vast majority of users. However, if the environment variable exists, it indicates to the signfile.py script that the target .exe|.pyd should be signed, using the identifier specified by the environment variable, with Microsoft's code-signing facilities (signtool.exe etc). The net effect is that if a user has a valid code-signing certificate installed, they can get the standard Python build system to automatically sign all binaries by setting this environment variable. I think this approach is more useful than, say, having Martin manually sign each binary when it comes to release time, as it is automated (and as such, less error prone), and can be leveraged by anyone wishing to create signed Python distributions, not just the PSF (i.e. ActiveState, Enthought, etc). Assuming there are no objections to this proposal, a couple of other things I'll provide: a) documentation on both what's required in order to produce a signed distribution, as well as guidelines for other companies wishing to obtain code-signing certificates for the purposes of signing their custom Python distributions (perhaps better suited to a wiki entry) b) a helper script in pcbuild/ that can a) augment existing .vcprojs and add the post-link signing step, and b) be run in such a way that it checks all existing .vcprojs for .exe|.pyd targets have the post-link event correctly configured Comments/objections? Trent. -- assignee: Trent.Nelson components: Build messages: 66814 nosy: Trent.Nelson, loewis, tmick priority: normal severity: normal status: open title: Augmenting the Windows build to support code signing. type: feature request versions: Python 2.6, Python 3.0 __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2850 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2775] Implement PEP 3108
Quentin Gallet-Gilles [EMAIL PROTECTED] added the comment: I'm working on renaming the ConfigParser module. -- nosy: +quentin.gallet-gilles __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2775 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2775] Implement PEP 3108
Jesse Noller [EMAIL PROTECTED] added the comment: I've somehow lost my xmlrpc* changes on my local machine - if someone else gets to it before me, feel free to work on it. __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2775 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2851] Eliminate Perl legacy in re flag names
New submission from Mark Summerfield [EMAIL PROTECTED]: The re module has the following flags (amongst others): re.X == re.VERBOSE re.S == re.DOTALL The short forms of both these flags are clearly taken from Perl, but they don't seem necessary for Python and are confusing since all the other short names start with the same letter as the long name, e.g., re.I == re.IGNORECASE and re.M == re.MULTILINE. Why not add re.V for re.VERBOSE and re.D for re.DOTALL and kill re.X and re.S and say a final farewell to Perl? -- components: Library (Lib) messages: 66817 nosy: mark severity: normal status: open title: Eliminate Perl legacy in re flag names type: feature request versions: Python 3.0 __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2851 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2845] shutil.copy2() copies permission bits
Alexander Belopolsky [EMAIL PROTECTED] added the comment: This seems to be a documentation issue. shutil.copy2(src, dst) Similar to copy(), but last access time and last modification time are copied as well. This is similar to the Unix command cp -p. -- http://docs.python.org/dev/library/shutil.html#shutil.copy2 Does not mention permissions while $ pydoc shutil.copy2 shutil.copy2 = copy2(src, dst) Copy data and all stat info (cp -p src dst). implies that the permissions should be copied. Maybe the documentation should simply recite the implementation: copy2 is copyfile follwed by copystat. -- nosy: +belopolsky __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2845 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2851] Eliminate Perl legacy in re flag names
Guido van Rossum [EMAIL PROTECTED] added the comment: Please keep them. They correspond to the (?x) and (?s) syntax that is supported inside the regex. Perl compatibility is a feature, not a bug for the re module. -- nosy: +gvanrossum resolution: - wont fix status: open - closed __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2851 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2838] Verify callback for SSL
Bill Janssen [EMAIL PROTECTED] added the comment: On the client side, are you passing a ca_certs file with the self-signed certificate in it? If not, the library won't be able to validate the certificate enough to be able to see the data in it. But if you do that, you should be able to see the bits of the certificate. There's no point to seeing the bits of an unvalidated certificate, because they may be forged. So the library doesn't allow you to see the bits of an unvalidated certificate from the other side of the connection. -- assignee: - janssen nosy: +janssen __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2838 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2852] sidebar directive fails
New submission from Noah Kantrowitz [EMAIL PROTECTED]: The normal ReST sidebar directive creates a div with the same class as the HTML sidebar. This makes it not render its contents correctly. Perhaps the default template/CSS should match on div#sphinxsidebar or similar? -- assignee: georg.brandl components: Documentation tools (Sphinx) messages: 66821 nosy: coderanger, georg.brandl severity: normal status: open title: sidebar directive fails type: feature request __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2852 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2775] Implement PEP 3108
Juracy Filho [EMAIL PROTECTED] added the comment: I almost finished the http package patch, but I'm doubt about how I would do it with Docs. There are various doc files: basehttpserver.rst, cgihttpserver.rst, httplib.rst and so on. Would I join them into one's or only update their information (imports) ? __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2775 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2778] set_swap_bodies is unsafe
Adam Olsen [EMAIL PROTECTED] added the comment: Revised again. sets are only hashed after PyObject_Hash raises a TypeError. This also fixes a regression in test_subclass_with_custom_hash. Oddly, it doesn't show up in trunk, but does when my previous patch is applied to py3k. Added file: http://bugs.python.org/file10321/python-setswap-3.diff __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2778 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2853] *** glibc detected *** python: double free or corruption
New submission from Michael Lang [EMAIL PROTECTED]:
Hi,
i am trying to solve some problems we encounter, when locking files on a
NFS Storage using fcntl.
since this is a security related problem i just add some pseudo code
here that was used to create the problem
fh = os.open('filename')
fcntl.lockf(fh, fcntl.LOCK_EX)
fhw = os.fdopen(fh)
fhw
fcntl.lockf(fh, fcntl.LOCK_UN)
...
when using threads, it's possible to create following problems when
using a Solaris (openSolaris) NFS server:
*** glibc detected *** python: double free or corruption (!prev):
0x1bdbfb20 ***
=== Backtrace: =
/lib64/libc.so.6[0x32b086f4f4]
/lib64/libc.so.6(cfree+0x8c)[0x32b0872b1c]
/lib64/libc.so.6(fclose+0x14b)[0x32b085e75b]
/usr/lib64/libpython2.4.so.1.0[0x32c3e447ce]
/usr/lib64/libpython2.4.so.1.0(PyEval_EvalFrame+0x47c7)[0x32c3e947a7]
/usr/lib64/libpython2.4.so.1.0(PyEval_EvalFrame+0x44a6)[0x32c3e94486]
/usr/lib64/libpython2.4.so.1.0(PyEval_EvalCodeEx+0x925)[0x32c3e95905]
/usr/lib64/libpython2.4.so.1.0[0x32c3e4c263]
/usr/lib64/libpython2.4.so.1.0(PyObject_Call+0x10)[0x32c3e35f90]
/usr/lib64/libpython2.4.so.1.0[0x32c3e3c01f]
/usr/lib64/libpython2.4.so.1.0(PyObject_Call+0x10)[0x32c3e35f90]
/usr/lib64/libpython2.4.so.1.0(PyEval_CallObjectWithKeywords+0x6d)[0x32c3e8f55d]
/usr/lib64/libpython2.4.so.1.0[0x32c3ebb33d]
/lib64/libpthread.so.0[0x32b14062f7]
/lib64/libc.so.6(clone+0x6d)[0x32b08ce85d]
=== Memory map:
0040-00401000 r-xp fd:01 845448
/usr/bin/python
0060-00601000 rw-p fd:01 845448
/usr/bin/python
1bd4d000-1bdd rw-p 1bd4d000 00:00 0
4000-40001000 ---p 4000 00:00 0
40001000-40a01000 rw-p 40001000 00:00 0
40a01000-40a02000 ---p 40a01000 00:00 0
40a02000-41402000 rw-p 40a02000 00:00 0
41402000-41403000 ---p 41402000 00:00 0
41403000-41e03000 rw-p 41403000 00:00 0
41e03000-41e04000 ---p 41e03000 00:00 0
41e04000-42804000 rw-p 41e04000 00:00 0
42804000-42805000 ---p 42804000 00:00 0
42805000-43205000 rw-p 42805000 00:00 0
43205000-43206000 ---p 43205000 00:00 0
43206000-43c06000 rw-p 43206000 00:00 0
43c06000-43c07000 ---p 43c06000 00:00 0
43c07000-44607000 rw-p 43c07000 00:00 0
44607000-44608000 ---p 44607000 00:00 0
44608000-45008000 rw-p 44608000 00:00 0
45008000-45009000 ---p 45008000 00:00 0
45009000-45a09000 rw-p 45009000 00:00 0
45a09000-45a0a000 ---p 45a09000 00:00 0
45a0a000-4640a000 rw-p 45a0a000 00:00 0
32b040-32b041a000 r-xp fd:00 127400
/lib64/ld-2.5.so
32b0619000-32b061a000 r--p 00019000 fd:00 127400
/lib64/ld-2.5.so
32b061a000-32b061b000 rw-p 0001a000 fd:00 127400
/lib64/ld-2.5.so
32b080-32b0946000 r-xp fd:00 127417
/lib64/libc-2.5.so
32b0946000-32b0b46000 ---p 00146000 fd:00 127417
/lib64/libc-2.5.so
32b0b46000-32b0b4a000 r--p 00146000 fd:00 127417
/lib64/libc-2.5.so
32b0b4a000-32b0b4b000 rw-p 0014a000 fd:00 127417
/lib64/libc-2.5.so
32b0b4b000-32b0b5 rw-p 32b0b4b000 00:00 0
32b0c0-32b0c82000 r-xp fd:00 127423
/lib64/libm-2.5.so
32b0c82000-32b0e81000 ---p 00082000 fd:00 127423
/lib64/libm-2.5.so
32b0e81000-32b0e82000 r--p 00081000 fd:00 127423
/lib64/libm-2.5.so
32b0e82000-32b0e83000 rw-p 00082000 fd:00 127423
/lib64/libm-2.5.so
32b100-32b1002000 r-xp fd:00 127455
/lib64/libdl-2.5.so
32b1002000-32b1202000 ---p 2000 fd:00 127455
/lib64/libdl-2.5.so
32b1202000-32b1203000 r--p 2000 fd:00 127455
/lib64/libdl-2.5.so
32b1203000-32b1204000 rw-p 3000 fd:00 127455
/lib64/libdl-2.5.so
32b140-32b1415000 r-xp fd:00 127463
/lib64/libpthread-2.5.so
32b1415000-32b1614000 ---p 00015000 fd:00 127463
/lib64/libpthread-2.5.so
32b1614000-32b1615000 r--p 00014000 fd:00 127463
/lib64/libpthread-2.5.so
32b1615000-32b1616000 rw-p 00015000 fd:00 127463
/lib64/libpthread-2.5.so
32b1616000-32b161a000 rw-p 32b1616000 00:00 0
32b640-32b640d000 r-xp fd:00 127465
/lib64/libgcc_s-4.1.2-20070626.so.1
32b640d000-32b660d000 ---p d000 fd:00 127465
/lib64/libgcc_s-4.1.2-2Segmentation fault
python imported modules/functions
from threading import Thread
import fcntl
from os import O_APPEND, O_CREAT, O_EXCL, O_LARGEFILE, O_NDELAY, ...
from time import asctime, localtime, sleep
from os import open as oopen
from os import fdopen
import sys
is this a python bug ? or am i doing something wrong ... the real code
will be available to troubleshoot the problem on request
regards
import sys
sys.version
'2.4.3 (#1, Mar 14 2007, 19:01:42)
[issue2850] Augmenting the Windows build to support code signing.
Martin v. Löwis [EMAIL PROTECTED] added the comment: Objection. I don't see the point of signing the binaries; signing the MSI files should be enough. __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2850 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2838] Verify callback for SSL
Ruben Kerkhof [EMAIL PROTECTED] added the comment: Hi Bill, When I include the server certificate in ca_certs, verification succeeds, and I can view the peer certificate dict with getpeercert(False) When I set ca_certs to none and ssl.CERT_NONE, I can still call getpeercert(True) and call DER_cert_to_PEM_cert to get the same PEM certificate. SSL is all new to me, so forgive me if I talk nonsense, but what I'm trying to do is the following: I receive a key from Bob which is a digest of his servers certificate. To make sure I'm really talking to Bob I need to decrypt his servers signature with his public key and check the resulting digest against my key. So I have to ignore failures like X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT and X509_V_ERR_SELF_SIGNED_CERT_IN_CHAIN, but detect things like X509_V_ERR_CERT_SIGNATURE_FAILURE. The idea is based on what foolscap is doing with FURLS (http://foolscap.lothar.com/trac) Am I making sense? __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2838 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2854] Add gestalt back into Python 3.0
New submission from Brett Cannon [EMAIL PROTECTED]: gestalt was removed as part of the stdlib cleanup for Mac code. But it turns out that gestalt is one of those modules that is just needed. Either the original version needs to be added back in or a ctypes version needs to be implemented. Either way, something to get system info needs to be added back into the stdlib (and probably go into plat-darwin). -- components: Library (Lib) messages: 66828 nosy: brett.cannon priority: release blocker severity: normal status: open title: Add gestalt back into Python 3.0 type: behavior versions: Python 3.0 __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2854 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2855] lookkey should INCREF/DECREF startkey around PyObject_RichCompareBool
New submission from Adam Olsen [EMAIL PROTECTED]: sets are based on dicts' code, so they have the same problem as bug 1517. Patch attached. -- files: python-lookkeycompare.diff keywords: patch messages: 66829 nosy: Rhamphoryncus severity: normal status: open title: lookkey should INCREF/DECREF startkey around PyObject_RichCompareBool Added file: http://bugs.python.org/file10322/python-lookkeycompare.diff __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2855 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2775] Implement PEP 3108
Changes by Brett Cannon [EMAIL PROTECTED]: -- dependencies: +Add gestalt back into Python 3.0 __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2775 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2855] lookkey should INCREF/DECREF startkey around PyObject_RichCompareBool
Raymond Hettinger [EMAIL PROTECTED] added the comment: FWIW, I periodically update setobject.c based on patches made to dictobject.c so there is usually no need to post separate bug reports. -- assignee: - rhettinger nosy: +rhettinger __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2855 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2854] Add gestalt back into Python 3.0
Brett Cannon [EMAIL PROTECTED] added the comment: Actually, the plat-darwin comment is not right if the original C version is kept since that would just end up in Modules. __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2854 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2854] Add gestalt back into Python 3.0
Benjamin Peterson [EMAIL PROTECTED] added the comment: Does it have to be public? How about _gestalt? -- nosy: +benjamin.peterson __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2854 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2778] set_swap_bodies is unsafe
Raymond Hettinger [EMAIL PROTECTED] added the comment: By replacing temporary immutability with temporary hashability, does this approach create the possibility that someone could mutate the key- set during a search? Is it possible to get the __eq__ check out-of- sync with the __hash__ value? Also, after a quick look at the patch, I'm not too keen on any modifications to set_swap_bodies() nor with changing the signature of set_contains_key(). __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2778 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2838] Verify callback for SSL
Bill Janssen [EMAIL PROTECTED] added the comment: Yep, it looks like you're on the right track. I'll close this bug. Bill On Wed, May 14, 2008 at 12:51 PM, Ruben Kerkhof [EMAIL PROTECTED] wrote: Ruben Kerkhof [EMAIL PROTECTED] added the comment: Hi Bill, When I include the server certificate in ca_certs, verification succeeds, and I can view the peer certificate dict with getpeercert(False) When I set ca_certs to none and ssl.CERT_NONE, I can still call getpeercert(True) and call DER_cert_to_PEM_cert to get the same PEM certificate. SSL is all new to me, so forgive me if I talk nonsense, but what I'm trying to do is the following: I receive a key from Bob which is a digest of his servers certificate. To make sure I'm really talking to Bob I need to decrypt his servers signature with his public key and check the resulting digest against my key. So I have to ignore failures like X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT and X509_V_ERR_SELF_SIGNED_CERT_IN_CHAIN, but detect things like X509_V_ERR_CERT_SIGNATURE_FAILURE. The idea is based on what foolscap is doing with FURLS (http://foolscap.lothar.com/trac) Am I making sense? __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2838 __ Added file: http://bugs.python.org/file10323/unnamed __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2838 __Yep, it looks like you#39;re on the right track.nbsp; I#39;ll close this bug.brbrBillbrbrdiv class=gmail_quoteOn Wed, May 14, 2008 at 12:51 PM, Ruben Kerkhof lt;a href=mailto:[EMAIL PROTECTED][EMAIL PROTECTED]/agt; wrote:br blockquote class=gmail_quote style=border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;br Ruben Kerkhof lt;a href=mailto:[EMAIL PROTECTED][EMAIL PROTECTED]/agt; added the comment:br br Hi Bill,br br When I include the server certificate in ca_certs, verificationbr succeeds, and I can view the peer certificate dict with getpeercert(False)br br When I set ca_certs to none and ssl.CERT_NONE, I can still callbr getpeercert(True) and call DER_cert_to_PEM_cert to get the same PEMbr certificate.br br SSL is all new to me, so forgive me if I talk nonsense, but what I#39;mbr trying to do is the following:br br I receive a key from Bob which is a digest of his servers certificate.br To make sure I#39;m really talking to Bob I need to decrypt his serversbr signature with his public key and check the resulting digest against mybr key. So I have to ignore failures likebr X509_V_ERR_DEPTH_ZERO_SELF_SIGNED_CERT andbr X509_V_ERR_SELF_SIGNED_CERT_IN_CHAIN, but detect things likebr X509_V_ERR_CERT_SIGNATURE_FAILURE.br br The idea is based on what foolscap is doing with FURLSbr (a href=http://foolscap.lothar.com/trac; target=_blankhttp://foolscap.lothar.com/trac/a)br br Am I making sense?br divdiv/divdiv class=Wj3C7cbr __br Tracker lt;a href=mailto:[EMAIL PROTECTED][EMAIL PROTECTED]/agt;br lt;a href=http://bugs.python.org/issue2838; target=_blankhttp://bugs.python.org/issue2838/agt;br __br /div/div/blockquote/divbr ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2838] Verify callback for SSL
Changes by Bill Janssen [EMAIL PROTECTED]: -- resolution: - works for me status: open - closed __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2838 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2778] set_swap_bodies is unsafe
Adam Olsen [EMAIL PROTECTED] added the comment: There is no temporary hashability. The hash value is calculated, but never stored in the set's hash field, so it will never become out of sync. Modification while __hash__ or __eq__ is running is possible, but for __eq__ that applies to any mutable type. set_contains_key only has two callers, one for each value of the treat_set_key_as_frozen argument, so I could inline it if you'd prefer that? set_swap_bodies has only one remaining caller, which uses a normal set, not a frozenset. Using set_swap_bodies on a frozenset would be visible except in a few special circumstances (ie it only contains builtin types), so a sanity check against that seems appropriate. The old code reset -hash to -1 in case one of the arguments was a frozenset - impossible now, so I sanity check that it's always -1. __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2778 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2632] performance problem in socket._fileobject.read
Changes by Georg Brandl [EMAIL PROTECTED]: -- priority: critical - release blocker __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2632 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2854] Add gestalt back into Python 3.0
Brett Cannon [EMAIL PROTECTED] added the comment: On Wed, May 14, 2008 at 1:37 PM, Benjamin Peterson [EMAIL PROTECTED] wrote: Benjamin Peterson [EMAIL PROTECTED] added the comment: Does it have to be public? How about _gestalt? That's fine by me. If someone really cares enough they can document it themselves. __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2854 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2854] Add gestalt back into Python 3.0
Benjamin Peterson [EMAIL PROTECTED] added the comment: I feel kinda bad about removing the module without noting how it was used in platform.py, so I'll deal with this. -- assignee: - benjamin.peterson __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2854 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2854] Add gestalt back into Python 3.0
Brett Cannon [EMAIL PROTECTED] added the comment: On Wed, May 14, 2008 at 6:01 PM, Benjamin Peterson [EMAIL PROTECTED] wrote: Benjamin Peterson [EMAIL PROTECTED] added the comment: I feel kinda bad about removing the module without noting how it was used in platform.py, so I'll deal with this. While I am not going to stop you dealing with it, you shouldn't feel bad. It would have been best had a test case in test_platform picked this up (if you can it would be great to toss a test in). __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2854 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2856] os.listdir doc should mention that Unicode decoding can fail
New submission from Brodie Rao [EMAIL PROTECTED]: The documentation for os.listdir should mention that there's a possibility that it can fail to decode paths to unicode objects and that it returns str objects for those paths it couldn't decode. The documentation should also explain when this might happen, and perhaps how to deal with it. The cases that this could happen or how to work around it are beyond my knowledge, but #683592 does mention something about resetting the locale. I don't know if that comment is still relevant, or if it applies to all cases of decoding failure, however. I don't know if this behavior still exists in Python 2.6. If it does, the documentation for 2.6 should be amended as well. -- assignee: georg.brandl components: Documentation messages: 66839 nosy: brodierao, georg.brandl severity: normal status: open title: os.listdir doc should mention that Unicode decoding can fail type: behavior versions: Python 2.3, Python 2.4, Python 2.5 __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2856 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2775] Implement PEP 3108
Juracy Filho [EMAIL PROTECTED] added the comment: I've finished a patch for http package, but I've doubts about how to make a patch. I've used a svn diff and svn status to make the patch and status file respectively. Added file: http://bugs.python.org/file10324/http_package_on_py3k.diff __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2775 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2775] Implement PEP 3108
Juracy Filho [EMAIL PROTECTED] added the comment: Output for svn status of http package patch. Added file: http://bugs.python.org/file10325/http_package_on_py3k.status __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2775 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2775] Implement PEP 3108
Brett Cannon [EMAIL PROTECTED] added the comment: On Wed, May 14, 2008 at 6:33 PM, Juracy Filho [EMAIL PROTECTED] wrote: Juracy Filho [EMAIL PROTECTED] added the comment: I've finished a patch for http package, but I've doubts about how to make a patch. I've used a svn diff and svn status to make the patch and status file respectively. That should be enough. Thanks, Juracy! __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2775 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2857] add coded for java modified utf-8
New submission from paul rubin [EMAIL PROTECTED]: For object serialization and some other purposes, Java encodes unicode strings with a modified version of utf-8: http://en.wikipedia.org/wiki/UTF-8#Java http://java.sun.com/javase/6/docs/api/java/io/DataInput.html#modified-utf-8 It is used in Lucene index files among other places. It would be useful if Python had a codec for this, maybe called UTF-8J or something like that. -- components: Library (Lib) messages: 66843 nosy: phr severity: normal status: open title: add coded for java modified utf-8 versions: Python 2.5 __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2857 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2819] Full precision summation
Raymond Hettinger [EMAIL PROTECTED] added the comment: When you need full precision, the Kahan approach helps but doesn't make guarantees and can sometimes hurt (it makes some assumptions about the data). One use case in is computing stats like a mean where many of the larger magnitude entries tend to cancel out but only after clipping bits off of the lower magnitude components. __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2819 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
[issue2858] bsddb.db.DBEnv.lock_get test_lock.test03_set_timeout crashes
New submission from Gregory P. Smith [EMAIL PROTECTED]: I disabled the Lib/bsddb/test/test_lock.py test03_set_timeout test as it crashes the interpreter when compiled in debug mode with an UNREF test. It appears to happen on all platforms according to the buildbots. This is not a new problem, it exists in 2.5 if a backport the recent updates to this test. I did a little bit of debugging and it looks like the UNREF failure occurs when the call to DBEnv.lock_get is has exited and Python is freeing the internal callargs tuple passed as arguments to the method. A pointer in the callargs._ob_prev structure had the 0xdbdbdbdb value which I believe is a sign of BerkeleyDB overwriting the memory. Jesus, can you look into this one? -- components: Extension Modules messages: 66845 nosy: gregory.p.smith, jcea priority: high severity: normal status: open title: bsddb.db.DBEnv.lock_get test_lock.test03_set_timeout crashes type: crash versions: Python 2.5, Python 2.6 __ Tracker [EMAIL PROTECTED] http://bugs.python.org/issue2858 __ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
