[python-win32] Oldest Python version for pywin32?

[Tim Golden]

Really addressed to Mark / Roger, but in case anyone else wants to chip
in anyway..

What's the oldest Python version still supported by pywin32? I ask
because I'm slightly tweaking the [somewhat gnarly] hunt-the-SDK code in
setup.py, and I don't want to introduce something which won't build on
an older version.

FWIW the changes I'm talking about are here:

https://bitbucket.org/TJG/pywin32/commits/227e8c794af76512832e5f79f8544c4a1e96915b

TJG

[Mark Hammond]

Hi Tim,
I still build for 2.5 and 3.1, but really only because they do still
build. If there's a reasonable reason to drop support for some I doubt
it will hurt many people - the sourceforge page should show you download
stats, but last I looked 2.5 was rarely used then, and that was some
time ago!

FTR, last time I built I used 6.0 of the SDK for 2.5 and 7.1 for the
rest. Dropping 2.5 for the next build sounds OK to me (and would
probably happen anyway if I had to recreate the build environment from
scratch - I don't think I'd bother hunting 6.0 down :)

Cheers,

Mark

Post by Tim Golden
Really addressed to Mark / Roger, but in case anyone else wants to chip
in anyway..
What's the oldest Python version still supported by pywin32? I ask
because I'm slightly tweaking the [somewhat gnarly] hunt-the-SDK code in
setup.py, and I don't want to introduce something which won't build on
an older version.
https://bitbucket.org/TJG/pywin32/commits/227e8c794af76512832e5f79f8544c4a1e96915b
TJG
_______________________________________________
python-win32 mailing list
https://mail.python.org/mailman/listinfo/python-win32

[Tim Golden]

Post by Mark Hammond
Hi Tim,
I still build for 2.5 and 3.1, but really only because they do still
build. If there's a reasonable reason to drop support for some I doubt
it will hurt many people - the sourceforge page should show you download
stats, but last I looked 2.5 was rarely used then, and that was some
time ago!

FWIW the sf page for build 219 shows less than 100 downloads for 2.5/6
and 3.1/2.

Unsuprisingly 2.7 dominates with 3,000 downloads at 32-bit (2,000 at
64-bit).

3.3 comes in just over 100 and 3.4/5 each a few hundreds.

So I feel no particular compunction about dropping forward support for
2.6 and lower and 3.2 and lower. (Could say 2.5/3.1 but it's the same
SDK level I think).

TJG

[Vernon D. Cole]

I concur. I am one of the 100 downloads of 2.5 -- and the only reason I
download it is to test it, not to use it in production. How many of the
other downloaders are like me? Most, I would bet. Dropping 2.5 would allow
use of many Python3 features, since 2.6 has the backports for them (print
function, "{}".format(), byte literals, and especially "except ... as"). It
would simplify the Python code in the library.

Post by Tim Golden

Post by Mark Hammond
Hi Tim,
I still build for 2.5 and 3.1, but really only because they do still
build. If there's a reasonable reason to drop support for some I doubt
it will hurt many people - the sourceforge page should show you download
stats, but last I looked 2.5 was rarely used then, and that was some
time ago!

FWIW the sf page for build 219 shows less than 100 downloads for 2.5/6
and 3.1/2.
Unsuprisingly 2.7 dominates with 3,000 downloads at 32-bit (2,000 at
64-bit).
3.3 comes in just over 100 and 3.4/5 each a few hundreds.
So I feel no particular compunction about dropping forward support for
2.6 and lower and 3.2 and lower. (Could say 2.5/3.1 but it's the same
SDK level I think).
TJG
_______________________________________________
python-win32 mailing list
https://mail.python.org/mailman/listinfo/python-win32

[Andreas Holtz]

I am one 2.5 user, too.
Reason is, that thanks to 4Suite (which is a fantastic XML framework) we are bound to 2.5.
I didn't find a better XML framework for Python but it's not gonna developed further and will not be
ported to 3.x (reminds me of contacting Jeremy Kloth for a port to a newer Python version).

But on the other hand, I don't download PyWin regulary and stick with the version we are currently
using (214, I think). If a piece of software works fine (enough), there is no need to downloading a
newer version. So see it as a complement that people are happy with the software :-)

I concur. I am one of the 100 downloads of 2.5 -- and the only reason I download it is to test it,
not to use it in production. How many of the other downloaders are like me? Most, I would bet.
Dropping 2.5 would allow use of many Python3 features, since 2.6 has the backports for them (print
function, "{}".format(), byte literals, and especially "except ... as"). It would simplify the
Python code in the library.

Post by Mark Hammond
Hi Tim,
  I still build for 2.5 and 3.1, but really only because they do still
build. If there's a reasonable reason to drop support for some I doubt
it will hurt many people - the sourceforge page should show you download
stats, but last I looked 2.5 was rarely used then, and that was some
time ago!

FWIW the sf page for build 219 shows less than 100 downloads for 2.5/6
and 3.1/2.
Unsuprisingly 2.7 dominates with 3,000 downloads at 32-bit (2,000 at
64-bit).
3.3 comes in just over 100 and 3.4/5 each a few hundreds.
So I feel no particular compunction about dropping forward support for
2.6 and lower and 3.2 and lower. (Could say 2.5/3.1 but it's the same
SDK level I think).
TJG
_______________________________________________
python-win32 mailing list
https://mail.python.org/mailman/listinfo/python-win32
_______________________________________________
python-win32 mailing list
https://mail.python.org/mailman/listinfo/python-win32