Caching problems on main page

edit

We are about to solve this by merging the development branches to master.

We've started using .timeago() jQuery plugin to recalculate the displayed timestamp renderings in the browsers.

The main reason for this change is to allow much longer caching of the template fragments. We have also moved personalization-related code from .html templates to javascript on more popular pages - also to improve caching.

Thanks @Evgeny, looking forward to your solution!

Meanwhile, I tried to have a look into the code, although i'm not familiar with Python at all. I noticed that on /models/question.py file there is get_summary_html() function calling get_cached_summary_html() and if it doesn't return anything (probably meaning that it didn't find the data it seeks at the cache), it then calls update_summary_html():

def get_summary_html(self, search_state):
    html = self.get_cached_summary_html()
    if not html:
        html = self.update_summary_html()
    ...

So, if I comment out the first call and force it to always call update_summary_html(), problem disappears. Of course that's a stupid workaround that doesn't take advantage of the cache, so system may respond slower. But it works and gives the impression to users that everything is fine, so until team gives the perpanent solution, this may fit you.

Last, I notice that in the same file at line 456 of invalidate_cached_data() you have lately commented out the call of invalidate_cached_thread_content_fragment() which deletes stuff from the cache, and I was wondering if that change has anything to do with the problem we discuss here.