My property list is mainly empty - what happened and how do I fix it?

From semantic-mediawiki.org

Ah, new options to the script. Hmm ..., did I document this? No I would do the classic php rebuildData.php -fv once and when php rebuildData.php -v once. I do not think that what you are doing will work / is meant to work. If you want to rebuild specific tables thereafter you may use all these "cpt" options.

14:15, 20 August 2014

I'd really like to know how to understand what happened - so it doesn't happen again.... as well, of course, fixing it now.

The -v option is, as I understand it, supposed to visit each page, check if there are any properties on it, and, if there are, then recreate them. This should bring the properties back.

The fact that it hasn't, after running it a number of times, is a concern - where should I be looking to see what it's not doing?

The properties themselves are there, and not showing red any more, so it seems to have sorted out the table of properties itself, just not the pointers to where the properties are used.

that's why I thought that the -cpt options, particularly the -p option, might help.

Is there an error log somewhere that might help show what's not working?

14:22, 20 August 2014

The -v option just tells the script to prompt what it is doing while it is doing something. So basically you are deleting the store rebuild specific tables with -cpt and then rebuild the rest of the tables. To me this appears like building a car by adding the seats, trunk and lights etc. and when try to somehow add the chassis at the very end of the process. I doubt that this will work.

12:18, 21 August 2014

That certainly wouldn't work!

Looking at the documentation, it suggested that the standard, with no option, or a -v if you want to see how it is going, rebuilds the properties and so forth by looking at each page in turn from pageID 0000 on to 4000 or however many there are.

The -cpt, on the other hand, seem to rebuild the tables from properties, categories and other indices that it gets from elsewhere, an existing table or log, presumably - because it is much quicker and only looks at the ~500 items.

So the logic, to me, was to make sure the tables for the indices were in place with the -cpt option. Then get it to go through each page adding links to the indices.

The logic behind running it more than once was that there might be links that are not apparent the first time, concealed by parse expressions or similar that it might pickup on another pass.

I was hoping that one of the authors of the script could enlighten me on whether I've misunderstood.

12:45, 21 August 2014