Details
-
Improvement
-
Resolution: Obsolete
-
Medium
-
None
-
4.1.4, 4.2.0, 4.3.0, 4.4.0beta1
-
None
Description
in myslq.php:
function _startCacheGeneration()
line 1405: return array( 'result' => 'error'); would be more consistent
line 1415: since we use _selectonerow(), this test will always fail
line 1437: if the update fails because of 0 rows update, we should not log an error msg, as it is because either the cache generation by other process was ended (same as line 1415) or taken over by another process.
All of these 3 cases should be unlikely to occurr, so they should imho:
. return an 'error' code so that the caller startCacheGeneration() goes back to start of loop to check if the genreated file is now ok
. output a debug msg of high concurrency
If the update fails because of other sql errors though, an error should be reported.
ezdbfilehandler.php:
function processcache: always log file name in every message, it is helpful (eg. line 393)
also there might be a need to force the reload of file metadata when startcachegeneration abends, to see if it is beacuse another process just finished generating the same cache file (eg. around line 542)
In general:
. the messages needed to debug cache issues are split beteen error, debug and cluster, plus the ezdebugsetting one. This makes it hard to debug the code. Concentrating messages in fewer places would be good.