On 01/06/2011 11:08 AM, Victor Stinner wrote:
Le jeudi 06 janvier 2011 à 10:47 -0500, R. David Murray a écrit :
On Thu, 06 Jan 2011 12:55:24 +0100, Victor
Stinner<victor.stin...@haypocalc.com> wrote:
Le jeudi 06 janvier 2011 à 00:10 -0500, Alexander Belopolsky a écrit :
If calling specific system functions such as strftime with tm_year<
0 is deemed unsafe, we can move the check to where the system function
is called.
What do you mean by "unsafe"? Does it crash? On my Linux box,
strftime("%Y") is able to format integers in [-2^31-1900; 2^31-1-1900]
(full range of the int type).
I believe that we have had several cases where Windows "crashed" when
out-of-range values were passed to the CRT that other platforms
accepted.
If there are only issues on Windows, we can add a #ifdef _MSC_VER and
raise a ValueError("Stupid OS, install Linux or recompile with Cygwin")
for year< 1900.
Is strftime really so complex that we shouldn't just write our own? I'd
be willing to do it. Over the years the platform strftime has caused any
number of problems. The last time I looked at it we already have to do
some work pre-parsing the format string and passing it off to platform
strftime, so it's not like it's not already a maintenance hassle.
I understand strptime is probably more complex and there's some value to
having strptime/strftime coming from the same library. But I'd be
willing to look at it, too.
Eric.
_______________________________________________
Python-Dev mailing list
Python-Dev@python.org
http://mail.python.org/mailman/listinfo/python-dev
Unsubscribe:
http://mail.python.org/mailman/options/python-dev/archive%40mail-archive.com
