[issue16145] Abort in _csv module

R. David Murray added the comment: It won't be too much longer. -- ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscrib

[issue16145] Abort in _csv module

Roger Binns added the comment: So is 3.3.1 with the fix ever going to be released? Georg did predict mid-November and we are 4 months after that. -- ___ Python tracker ___

[issue16145] Abort in _csv module

Stefan Krah added the comment: Andrew and Serhiy, thanks for the reviews. Should be fixed now. -- assignee: -> skrah resolution: -> fixed stage: needs patch -> committed/rejected status: open -> closed ___ Python tracker

[issue16145] Abort in _csv module

Roundup Robot added the comment: New changeset 0b5f00ccf907 by Stefan Krah in branch '3.3': Issue #16145: Support legacy strings in the _csv module. http://hg.python.org/cpython/rev/0b5f00ccf907 -- nosy: +python-dev ___ Python tracker

[issue16145] Abort in _csv module

Larry Hastings added the comment: Roundup knows I'm the release manager for 3.4? It's well-informed! ;-) -- ___ Python tracker ___ __

[issue16145] Abort in _csv module

Arfrever Frehtes Taifersar Arahesis added the comment: Probably Versions="Python 3.4" + Priority="release blocker" results in addition of 3.4 Release Manager (Larry Hastings) to nosy list. -- ___ Python tracker __

[issue16145] Abort in _csv module

R. David Murray added the comment: I don't know. I was able to remove him. (I have javascript turned off, don't know if that makes any difference). -- ___ Python tracker ___ _

[issue16145] Abort in _csv module

Changes by R. David Murray : ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/pyth

[issue16145] Abort in _csv module

Serhiy Storchaka added the comment: Hmm. Is it a bugtracker bug? I don't know how I added Larry to the nosy list and now I can not remove him. -- ___ Python tracker ___ __

[issue16145] Abort in _csv module

Changes by Serhiy Storchaka : ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/pyt

[issue16145] Abort in _csv module

Changes by Serhiy Storchaka : ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.python.org/mailman/options/pyt

[issue16145] Abort in _csv module

Serhiy Storchaka added the comment: LGTM. But I left some minor comments in Rietveld. -- nosy: +larry ___ Python tracker ___ ___ Pytho

[issue16145] Abort in _csv module

Andrew Svetlov added the comment: LGTM -- nosy: +asvetlov ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: http:

[issue16145] Abort in _csv module

Georg Brandl added the comment: 3.3.1 will not be too long: maybe 4 weeks from now. -- nosy: +georg.brandl ___ Python tracker ___ ___

[issue16145] Abort in _csv module

Changes by Arfrever Frehtes Taifersar Arahesis : -- components: +Extension Modules priority: normal -> release blocker ___ Python tracker ___

[issue16145] Abort in _csv module

Roger Binns added the comment: Roughly how long will it be before Python 3.3.1 comes out? This issue means my users will get garbage or crashes, so I'll need to work around it if it will be quite a while till 3.3.1. -- ___ Python tracker

[issue16145] Abort in _csv module

Stefan Krah added the comment: > The PEPs and doc said the C API would remain backwards compatible. Roger: Yes, you don't need to change anything. I was misled by the PyUnicode_AsUnicode() docs, thinking that the responsibility to call PyUnicode_READY() also applies if the generated Unicode Obje

[issue16145] Abort in _csv module

Stefan Krah added the comment: Here's a patch. Victor, I guess I have a feature request for fusil: It would be nice if fusil also generated legacy strings (or does it already do so?). -- keywords: +patch Added file: http://bugs.python.org/file27454/issue16145.diff

[issue16145] Abort in _csv module

Roger Binns added the comment: (APSW author here). I haven't ported to the new Python 3.3 Unicode handling yet but it is on my todo list. The PEPs and doc said the C API would remain backwards compatible. The APSW code supports Python 2.3 onwards. SQLite APIs support both UTF-8 and UTF-16.

[issue16145] Abort in _csv module

STINNER Victor added the comment: "In general APSW does not look ready for the Python 3.3 Unicode implementation." I bet that most Python modules implemented in C use the "legacy" (old) Unicode API (Py_UNICODE*). Python 3.3 must call PyUnicode_READY() everywhere (where the new API is used to g

[issue16145] Abort in _csv module

STINNER Victor added the comment: "I think we need to add a couple of PyUnicode_READY() to some functions in Modules/_csv.c." Oh yes, the _csv module was in my TODO list. I forgot to fix it :-/ -- nosy: +haypo ___ Python tracker

[issue16145] Abort in _csv module

Stefan Krah added the comment: David is right, the existing (APSW) code should continue to work. I think we need to add a couple of PyUnicode_READY() to some functions in Modules/_csv.c. -- ___ Python tracker _

[issue16145] Abort in _csv module

Stefan Krah added the comment: R. David Murray wrote: > My understanding was that code that calls the public APIs should continue to > work. Is this a new requirement in 3.3, or is it that it has always been a > requirement but code could get away without the ready before? It's a new requirem

[issue16145] Abort in _csv module

R. David Murray added the comment: I'm going to reopen this at least until we have more information. -- resolution: invalid -> stage: committed/rejected -> needs patch status: closed -> open type: -> behavior ___ Python tracker

[issue16145] Abort in _csv module

Arfrever Frehtes Taifersar Arahesis added the comment: Apparently http://docs.python.org/py3k/whatsnew/3.3.html#pep-393-flexible-string-representation says: "On the C API side, PEP 393 is fully backward compatible." -- ___ Python tracker

[issue16145] Abort in _csv module

R. David Murray added the comment: My understanding was that code that calls the public APIs should continue to work. Is this a new requirement in 3.3, or is it that it has always been a requirement but code could get away without the ready before? Either way we need to add an note to the P

[issue16145] Abort in _csv module

Serhiy Storchaka added the comment: Oh, I see the issue has already been solved. -- ___ Python tracker ___ ___ Python-bugs-list mailin

[issue16145] Abort in _csv module

Serhiy Storchaka added the comment: How do you get values to write? Most likely, the problem has nothing to do with csv module, but written values are broken. Possible error in sqlite module, or in any conversion function, or in you APSW C-code. Can you dump this values before sending to write

[issue16145] Abort in _csv module

Stefan Krah added the comment: In general APSW does not look ready for the Python 3.3 Unicode implementation. Example: src/statementcache.c:217 Py_UNICODE *out; PyObject *res=PyUnicode_FromUnicode(NULL, size); Python 3.3 docs: "If the buffer is NULL, PyUnicode_READY() must be called o

[issue16145] Abort in _csv module

Roger Binns added the comment: I'm the APSW author. You do not need SQLite installed - APSW's setup can fetch the current SQLite and use it privately not affecting the rest of the system. An easier way of testing is: python3 setup.py fetch --sqlite --version 3.7.14 build_ext --inplace --for

[issue16145] Abort in _csv module

Antoine Pitrou added the comment: Well, can you try to enable core dumps and get a complete stack trace? -- ___ Python tracker ___ ___

[issue16145] Abort in _csv module

New submission from Arfrever Frehtes Taifersar Arahesis: Commit f2adbb1065eb introduced abort in _csv module in debug builds in Python 3.3, when using APSW shell. (You need to have SQLite >=3.7.14 installed.) $ cd /tmp $ wget http://apsw.googlecode.com/files/apsw-3.7.14-r2.zip $ unzip apsw-3.7