Support for Drupal 7 is ending on 5 January 2025—it’s time to migrate to Drupal 10! Learn about the many benefits of Drupal 10 and find migration tools in our resource center.
Don't mind me, this is just a helper issue to run testbot without clobbering #1040790: _field_info_collate_fields() memory usage.
Comment | File | Size | Author |
---|---|---|---|
#57 | field-info-cache-1040790-196-D7.patch | 59.64 KB | yched |
#56 | field-info-cache-1040790-199-D7.patch | 59.18 KB | yched |
#51 | field-info-cache-1040790-196-D7.patch | 58.95 KB | yched |
#47 | field-info-cache-1040790-89-D7-4.patch | 46.83 KB | yched |
#42 | field-info-cache-1040790-89-D7-3.patch | 46.8 KB | yched |
Comments
Comment #1
yched CreditAttribution: yched commentedGah.
Comment #3
yched CreditAttribution: yched commentedReally ?
Comment #5
yched CreditAttribution: yched commentedComment #7
yched CreditAttribution: yched commentedComment #9
yched CreditAttribution: yched commentedComment #11
yched CreditAttribution: yched commentedComment #13
yched CreditAttribution: yched commentedComment #15
yched CreditAttribution: yched commentedComment #17
yched CreditAttribution: yched commentedComment #19
yched CreditAttribution: yched commentedThat one should be green.
Comment #20
yched CreditAttribution: yched commentedLet's start refactoring, then.
Comment #21
andypostDo we need tests for this?
Comment #22
yched CreditAttribution: yched commentedThe values returned by the various field_info_*() functions are used all over the place and are thus fairly extensively tested, as the # of fails / exceptions above show.
Tests for the caching behavior (static cache in the FieldInfoFieldCache properties and persistent cache in the cache_field table) could probably be useful though.
Comment #23
yched CreditAttribution: yched commentedClosing, back to the parent issue.
Comment #24
yched CreditAttribution: yched commentedCrap, failures are back. Reopening the testing playground.
Comment #26
yched CreditAttribution: yched commentedComment #27
yched CreditAttribution: yched commentedRefactoring.
Comment #29
yched CreditAttribution: yched commentedComment #31
yched CreditAttribution: yched commentedComment #33
yched CreditAttribution: yched commentedComment #34
yched CreditAttribution: yched commentedMore refactoring
Comment #35
yched CreditAttribution: yched commentedSome more refactoring.
Comment #36
yched CreditAttribution: yched commentedJust trying out how the DIC plays with testbot right now.
Comment #38
yched CreditAttribution: yched commentedD7 backport ?
Comment #40
yched CreditAttribution: yched commentedComment #42
yched CreditAttribution: yched commentedComment #44
geerlingguy CreditAttribution: geerlingguy commentedIs the testbot having some troubles tonight?
Comment #45
geerlingguy CreditAttribution: geerlingguy commented#42: field-info-cache-1040790-89-D7-3.patch queued for re-testing.
Comment #47
yched CreditAttribution: yched commentedComment #48
jthorson CreditAttribution: jthorson commentedLooks like #45 either ran into or caused a cron issue on the testbot ... got an 'attempt to start review when already running', followed by an hour of 'attempting to re-run cron while it is already running' watchdog logs; suggesting that the cron process hung for the duration of that test.
Comment #49
yched CreditAttribution: yched commented@jthorson: sorry about that. #45 was an attempt at a D7 backport of #1040790: _field_info_collate_fields() memory usage. The problem was an improper conversion between D8 and D7's cache APIs. the patch contained cache_[get|set|clear_all]() calls with $bin = 'field' instead of 'cache_field'.
Not sure why that would lead to the effects you mentioned, but the patch is fixed now, won't happen again :-).
Comment #50
swentel CreditAttribution: swentel commentedLet's close this one.
Comment #51
yched CreditAttribution: yched commentedRebooting work on the D7 backport of #1040790: _field_info_collate_fields() memory usage. Let's see how this behaves.
Comment #52
geerlingguy CreditAttribution: geerlingguy commented(Just fyi, I don't use this particular patch on the production site anymore, as we've necessarily had to switch to a different solution outside the Entity/Field API for our profile system—we passed 600 entity types and 20,000 field instances, and Drupal was killing itself on cache rebuilds—with or without this patch).
I'll still help review the patch as best I can, of course, but I won't be able to test on such an outrageous installation :)
Comment #54
yched CreditAttribution: yched commented#51: field-info-cache-1040790-196-D7.patch queued for re-testing.
Comment #56
yched CreditAttribution: yched commentedComment #57
yched CreditAttribution: yched commented