Quote from: sharkbait at Nov 27, 2012, 04:10 PM
would it be possible to recreate an existing blog with PubKit, circumventing the cache bottleneckt with custom tables?
More room here than in a tweet! Starting from an extreme and working backwards, you could replicate the structure of the site_content table in your own table, and delete whatever columns are superfluous (among which would probably be "parent"). I think that would give you item storage outside the cache mechanism. Then you'd need custom snippets to do what Ditto does so well; the ListSections in my demo could be a starting point, or the old loopDbChunk extra, which has some templating. Or even a customized version of Ditto (?Blitto?), set to use your table instead of site_content as its data source.
You'd lose TVs as tags, but now you don't need an open set of TVs, you can create your own fields for tagging as required. (NB you can still use TVs to provide things like shared drop-down lists, but not store the selections in the TV tables per item as you would in a MODX resource).
and in your demo-setup, what's the difference between the news section and the blog
They demonstrate use of different selections of heading fields, also the two-level tagging in the News section (category + tags).
btw. I get "Check configuration. Set &root as -1 for entire tree" on the "ExportPages" module, any ideas?
Ha! Me too! The export works for sets of pages starting with an actual page, but not as yet for the whole MODX tree. Need to work on that one a bit more, also on making the Import a bit friendlier (it fails if there's no database table to import at the moment, and setting the module ID is a bit cacky). See attached pic of configuration of the Export module for sending 27 & its children to a file called allPages.csv in the assets/export directory.
I made progress on some other aspects of PubKit in October, but ran out of time to tie it all up; I hope to be retreating to my coding cave again this weekend to make something more of it.
:) KP