Call to undefined function drupal_get_path_alias() in url()

I've notice the same error using the drush rr command on some sites but not others. The patch in #2 resolved this issue for me.

I hit this issue on a postgres D7 install after I restore a database backup and try to do a registry rebuild.

The patch from #5 works but *only* if I disable redis caching. If I leave redis enabled (configured via settings.php) then drush fails with:

Drush command terminated abnormally due to an unrecoverable error.       [error]
Error: Cannot redeclare drupal_path_initialize() (previously declared
in /srv/www/html/sites/all/modules/contrib/redis/redis.path.inc:36)
in /srv/www/html/includes/path.inc, line 22

Looking at the redis module, it does indeed define a drupal_path_initialize() function, so either this include needs to be conditional *or* the redis config needs to be conditional and check if it's run via drush or not.

Dunno, if you make that require conditional (eg: if (!function_exists('drupal_get_path_alias')) { } ) it should work in both cases. That logic is used elsewhere in common.inc as well.

I did a little bit more looking and found what looks like a nice solution in #2474499: Conflict with Redis module.

Updated patch attached - this one does work for me locally with redis configured and without it on PHP7/PgSQL 9.5.

Hi all, and thanks for sleuthing out this problem. I had been checking my contrib modules for calls to url() in hook_init() or hook_boot() but didn't think about potentially fixing url() itself!

I'm going to upgrade this ticket's Priority to Critical as these php fatals were causing intermittent downtime for my site.

I've applied the patch in #10 and it seems to be working so far, but I'd like to let the site run for a whole day to see if any additional errors appear.

Update: The problem I ran into was actually caused by the Acquia purge module which registered a shutdown function that happened to call core functions. Since Drupal isn't always fully bootstrapped during the shutdown runtime phase I was getting PHP fatals that caused site downtime.

Applying the patch here fixed part of the problem, and enabled me to find the rest of the contrib code that was causing my problems.

I'm lowering the priority of this ticket from Critical to Major, because even though this problem was causing downtime for my site - it wasn't core's fault. Core could still prevent the downtime in some cases, however, so I'm leaving the status at Major.

Poked the patch so it applies to the current 7.x HEAD.

Patch in #14 also fixed this for me. Thanks!

I can haz test(s)?

I mean, the patch doesn't break core Path alias tests?

Path alias functionality 135 passes, 0 fails, and 0 exceptions

Can't really make the core tests depend on a contrib module, unless you mean make a failing test when path_inc is set?

Just a heads-up, if you're using (or have inherited) path_alias_xt module with the patch to core path.inc applied, you will need to patch your redis.path.inc as well.

Patch in #14 still applies cleanly to 7.70, 7.71, 7.72, 7.73, 7.74 and 7.75.

Was added to #3179845: [meta] Priorities for 2020-12-02 bugfix release of Drupal 7.76 / 7.77 on weekend to raise it's visibility, did not make the cut for next release this week but might in 3 months.

Thanks for the update @izmeez! Patch in #14 still applies cleanly to 7.77.

Patch in #14 still applies to 7.78.

Is this problem always caused by contrib (or custom) code?

I see several modules mentioned; I believe the problem @jenlampton mentioned in #12 has since been fixed in Acquia Purge. Redis has also been mentioned a few times.

Is this core's problem to solve or are we talking about "babysitting" broken code from elsewhere?

Thanks all!