On one of my roundup instance, I have a crawler that follow the calendar
and it reach the year 9999.
But the calendar display a link for next year but this link fails
because Python has a limit on datetime.year.
Here is a patch that prevents to show links that will fail.

Cédric,
thanks for the patch.

What is a good way to reach all the code paths that you have added?
(Best probably is an automatic test case? >:) )

Hi Cédric,

tested your patch now:
It solved 3 our of 4 cases for me.
When being on Nov 999, the next link for the month is shown
and leads to exceptions.OverflowError'>: date value out of range
in line 
        for week in calendar.monthcalendar(display.year, display.month):

If you happen to do a new patch, could you add a CHANGES.txt
entry as well? :)

Thanks,
Bernhard

On 29/07/12 20:04 +0000, Bernhard Reiter wrote:
> tested your patch now:
> It solved 3 our of 4 cases for me.
> When being on Nov 999, the next link for the month is shown
> and leads to exceptions.OverflowError'>: date value out of range
> in line 
>         for week in calendar.monthcalendar(display.year, display.month):

I know but it is a Python bug:

    http://bugs.python.org/issue15421

> If you happen to do a new patch, could you add a CHANGES.txt
> entry as well? :)

Ok, but I was looking how to add testcase for this.

-- 
Cédric Krier

B2CK SPRL
Rue de Rotterdam, 4
4000 Liège
Belgium
Tel: +32 472 54 46 59
Email/Jabber: cedric.krier@b2ck.com
Website: http://www.b2ck.com/

Ah, nice to know this is a python defect, would have been cool to know before 
testing. 

If creating an automated testcase is too cumbersome, we can do without.
The code for the test should not exceed the tested code too much in this case I 
believe.

Adding testcase is too cumbersome.
So here is the patch with the CHANGES entry.

Thanks!
hg4651:beb8d43f4d9d

While fixing the error when the year 9999 is reached is a good thing,
have you considered adding a <meta name="robots" content="noindex,
nofollow"> so that the search engines* will avoid wasting time and
resources to go through ~8000 years of links needlessly?

If you think this is reasonable I can add the meta tag to the
_generic.calendar.html page.

* Note that this won't prevent all the robots to follow the links, but
it's still better than nothing.

Ezio,
sounds like a good idea to me. Feel free to go head.

I added <meta name="robots" content="noindex, nofollow"> to the calendar
pages in a099ff2ceff3, and fixed the upstream Python issue too.

Ezio: Cool, thanks!