I think you may be missing my original point, dates aside, these are two different images for two different boards. Of course the paths are different. The dated folder structure is built-in with Wordpress, it's fundamental to how it handles file uploads.
And luckily for
@Robert, there's Wordpress/WooCommerce. Setting up what you're talking about (if both images were the same, and in this case, they're not) is possible but I don't know why you would. Permalink modules are likely readily available. Web has gotten far
more user-friendly than it was 25 years ago. Props to the guys that did all this manually back then.
As a programmer I've written programs that generate HTML as well as FTP code. I am not a user, WordPress is more of a packaged user interface. I don't think you understand what I'm saying? Or perhaps I was not clear enough....
After the selection criterion and sort pattern is selected all the data is copied in that order to a sequential access file, so that initial page display, as well as followup display during scrolling is in the same file and accessible by an internal NEXT command to the sequential file.
This file name would not change because the NEXT statement would not have a starting pointer.
On server testing during upgrades, temporary files often have release or date information in the name. This is so a global variable can be flipped and you're testing using the new version of apps and data. These are flipped back after testing. When testing is complete, the switch is flipped and it stays with the new version.
This appears to be a 'upgrade artifact' where the path name is hard coded to the now empty file. So initial page display works fine since it's pointed at the data, but when it hits the statement with the old name there is no data to display.
Probably. I suspect that WordPress is generating this on the fly and somewhere in the 'real code' a hard coded path name exists instead of the indirect pathname obtained from the global switch variable.
I was involved in writing AI applications when I retired in 2010, I've written word processors, command interpreters, O/S indexing routines, used machine code, assembly code, and four different advanced 3d database models.
Started programming in 1974 when we used flint tools to carve the code into rock!!!!