Everything's in place and ready to go. Georg, can you update the
documentation links on the web site? As soon as that's ready I'll
make the announcement.
Thanks everyone. -Barry
Hello. HtmlHelp(windows) in Python2.6.2 seems to be broken. I didn't notice this. :-(
http://bugs.python.org/issue5764
This happens because r71380 was not backported. (Documents under Doc/ require Sphinx 0.6.1 now)
I've merged this into release26-maint in r71639. But in order to update Python2.6.2 binary, I think Martin's help is needed. Could you?
On Apr 16, 2009, at 12:34 AM, Martin v. Löwis wrote:
I've merged this into release26-maint in r71639. But in order to
update Python2.6.2 binary, I think Martin's help is needed. Could you?I won't be re-releasing 2.6.2. We should make a 2.6.3 release instead.
Agreed. If necessary, and when we resolve the problem, we can do a
2.6.3 release. This can happen soon if the current situation is a
problem for Windows users.
-Barry
On 16 Apr, 2009, at 14:33, Barry Warsaw wrote:
On Apr 16, 2009, at 12:34 AM, Martin v. Löwis wrote:
I've merged this into release26-maint in r71639. But in order to
update Python2.6.2 binary, I think Martin's help is needed. Could you?I won't be re-releasing 2.6.2. We should make a 2.6.3 release
instead.Agreed. If necessary, and when we resolve the problem, we can do a
2.6.3 release. This can happen soon if the current situation is a
problem for Windows users.
Could "soon" at least be after this weekend? We found a small
configure issue at PyCon that forces me to build the installers on
10.4 and I'd like to fix that before a next release and don't have
time to seriously look into that before sunday.
FWIW. The issue is that printf("%zd") works correctly on OSX 10.5, but
fails to print negative numbers correctly on 10.4. If you build the
installer on 10.4 the resulting binary works fine on all OSX systems,
if you build the installer on 10.5 the repr of itertools.count(-1) is
wrong on 10.4 systems.
Ronald
On Apr 16, 2009, at 8:46 AM, Ronald Oussoren wrote:
On 16 Apr, 2009, at 14:33, Barry Warsaw wrote:
On Apr 16, 2009, at 12:34 AM, Martin v. Löwis wrote:
I've merged this into release26-maint in r71639. But in order to
update Python2.6.2 binary, I think Martin's help is needed. Could you?I won't be re-releasing 2.6.2. We should make a 2.6.3 release
instead.Agreed. If necessary, and when we resolve the problem, we can do a
2.6.3 release. This can happen soon if the current situation is a
problem for Windows users.Could "soon" at least be after this weekend? We found a small
configure issue at PyCon that forces me to build the installers on
10.4 and I'd like to fix that before a next release and don't have
time to seriously look into that before sunday.
Yes definitely. I won't have time until next week at the earliest
anyway.
FWIW. The issue is that printf("%zd") works correctly on OSX 10.5,
but fails to print negative numbers correctly on 10.4. If you build
the installer on 10.4 the resulting binary works fine on all OSX
systems, if you build the installer on 10.5 the repr of
itertools.count(-1) is wrong on 10.4 systems.
Do you have an issue number for this? Either way, please make sure
there's a release blocker issue on 2.6 for this problem.
-Barry
Barry Warsaw wrote:
On Apr 16, 2009, at 12:34 AM, Martin v. Löwis wrote:
I've merged this into release26-maint in r71639. But in order to update Python2.6.2 binary, I think Martin's help is needed. Could you?
I won't be re-releasing 2.6.2. We should make a 2.6.3 release instead.
Agreed. If necessary, and when we resolve the problem, we can do a 2.6.3 release. This can happen soon if the current situation is a problem for Windows users.
Meanwhile, I thought that it might be sufficient to upload a fixed version of the CHM file separately. Maybe that is good enough for people running into the problem.
Regards, Martin
This happens because r71380 was not backported.
This can't possibly be the problem. I don't use make.bat to build the html documentation.
To build the documentation, I had checked out the trunk of sphinx, r68598.
Since 2.6.2, I never did "make checkout", but only "make update"; I guess this was incorrect.
Regards, Martin
participants (4)
-
"Martin v. Löwis"
-
Barry Warsaw
-
Hirokazu Yamamoto
-
Ronald Oussoren