Commit Graph

20542 Commits

Author SHA1 Message Date
Daniel Hensby
5f6eec6403
Merge pull request #8269 from open-sausages/pulls/4/session-lazy
This is a cherry-picked PR from 4.2 -> 4.2.0
2018-07-23 14:10:16 +01:00
Ingo Schommer
73026292bf
Test coverage for session data change 2018-07-23 14:09:43 +01:00
Ingo Schommer
74b655d3fc
Fix tests on unset session data
Thanks Robbie!
2018-07-23 14:09:42 +01:00
Ingo Schommer
76ac8465de
BUG Lazy session state (fixes #8267)
Fixes regression from 3.x, where sessions where lazy started as required:
Either because an existing session identifier was sent through with the request,
or because new session data needed to be persisted as part of the request execution.

Without this lazy starting, *every* request will get a session,
which makes all those responses uncacheable by HTTP layers.

Note that 4.x also changed the $data vs. $changedData payloads:
In 3.x, they both contained key/value pairs.
In 4.x, $data contains key/value, while $changedData contains key/boolean to declare isChanged.
While this reduces duplication in the class, it also surfaced a bug which was latent in 3.x:
When an existing session is lazily resumed via start(), $data is set back to an empty array.
In 3.x, any changed data before this point was *also* retained in $changedData,
ensuring it gets merged into existing $_SESSION data.
In 4.x, this clears out data - hence the need for a more complex merge logic.

Since isset($this->data) is no longer an accurate indicator of a started session,
we introduce a separate $this->started flag.

Note that I've chosen not to make lazy an opt-in (e.g. via start($request, $lazy=false)).
We already have a distinction between lazy starting via init(), and force starting via start().
2018-07-23 14:09:42 +01:00
Daniel Hensby
ad850a187d
Merge pull request #8263 from dhensby/4.1/mask-backtrace
This is a cherry-picked PR from 4.2 -> 4.2.0
2018-07-23 14:08:44 +01:00
Daniel Hensby
e37b3b95f4
FIX updateValidatePassword calls need to be masked from backtraces 2018-07-23 14:08:26 +01:00
Daniel Hensby
ab942c9290
Merge pull request #8245 from dhensby/pulls/4.2/http-cache-fixes
This is a cherry-picked PR from 4.2 -> 4.2.0
2018-07-23 14:07:25 +01:00
Daniel Hensby
540c9aba69
TEST Add tests for (im)mutablity of states 2018-07-23 14:07:11 +01:00
Daniel Hensby
a3687147fe
State default should be state enabled (no-cache is an enabled state) 2018-07-23 14:07:10 +01:00
Daniel Hensby
9f1471332d
Make augmentState method more efficient 2018-07-23 14:07:10 +01:00
Daniel Hensby
cebed776ab
FIX If theres a max-age set remove no-cache and no-store 2018-07-23 14:07:09 +01:00
Daniel Hensby
2b1c55bc4e
FIX Allow setNoCache(false) to remove no-cache directive 2018-07-23 14:07:09 +01:00
Daniel Hensby
842b39e988
FIX Add must-revalidate to default state so its common on all our core states 2018-07-23 14:07:09 +01:00
Daniel Hensby
997730aa7f
FIX Allow cache control changes to affect default state 2018-07-23 14:07:08 +01:00
Daniel Hensby
d46410c49b
Merge pull request #8233 from dhensby/pulls/4.2/caching-fixes
This is a cherry-picked PR from 4.2 -> 4.2.0
2018-07-23 14:05:41 +01:00
Daniel Hensby
c52be7fe09
Consolidate disabling cache logic 2018-07-23 14:05:12 +01:00
Daniel Hensby
f7f567a12e
Make config private (notation isnt working) 2018-07-23 14:05:12 +01:00
Daniel Hensby
b78a89a76c
FIX Default cache state should be no-cache 2018-07-23 14:05:12 +01:00
Daniel Hensby
5ec0fafe43
Merge pull request #8193 from open-sausages/pulls/4.2/reset-on-ddl
This is a cherry-picked PR from 4.2 -> 4.2.0
2018-07-23 14:04:15 +01:00
Daniel Hensby
793aafae91
FIX Transaction depth should error if not implemented by child classes 2018-07-23 14:03:20 +01:00
Robbie Averill
1048520fbe
Restore check for zero or negative transaction nesting 2018-07-23 14:01:22 +01:00
Daniel Hensby
cbdf547c1b
Address feedback 2018-07-23 14:01:22 +01:00
Damian Mooyman
1d06dcdd28
Use 2.1 of postgres 2018-07-23 14:01:22 +01:00
Damian Mooyman
8ea3bb36a0
Maybe fix it? 2018-07-23 14:01:21 +01:00
Damian Mooyman
225e61dc67
BUG FIx manual resetDBSchema() calls breaking the database 2018-07-23 14:01:21 +01:00
Damian Mooyman
11e0a3de43
BUG Ensure that build includes extra classes 2018-07-23 14:01:21 +01:00
Damian Mooyman
95bcac796a
ENHANCEMENT Ensure test DB is flushed on either DDL or transaction-disabled tests
Fixes #8182
2018-07-23 14:01:20 +01:00
Daniel Hensby
b984959170
Merge branch '4.1' into 4.2 2018-07-23 13:30:49 +01:00
Ingo Schommer
389cc0d5fd
Merge pull request #8269 from open-sausages/pulls/4/session-lazy
BUG Lazy session state (fixes #8267)
2018-07-20 15:28:18 +12:00
Ingo Schommer
c541283093 Test coverage for session data change 2018-07-20 15:13:26 +12:00
Ingo Schommer
e415bcb44a Fix tests on unset session data
Thanks Robbie!
2018-07-19 13:32:08 +12:00
Ingo Schommer
93b0884e19 BUG Lazy session state (fixes #8267)
Fixes regression from 3.x, where sessions where lazy started as required:
Either because an existing session identifier was sent through with the request,
or because new session data needed to be persisted as part of the request execution.

Without this lazy starting, *every* request will get a session,
which makes all those responses uncacheable by HTTP layers.

Note that 4.x also changed the $data vs. $changedData payloads:
In 3.x, they both contained key/value pairs.
In 4.x, $data contains key/value, while $changedData contains key/boolean to declare isChanged.
While this reduces duplication in the class, it also surfaced a bug which was latent in 3.x:
When an existing session is lazily resumed via start(), $data is set back to an empty array.
In 3.x, any changed data before this point was *also* retained in $changedData,
ensuring it gets merged into existing $_SESSION data.
In 4.x, this clears out data - hence the need for a more complex merge logic.

Since isset($this->data) is no longer an accurate indicator of a started session,
we introduce a separate $this->started flag.

Note that I've chosen not to make lazy an opt-in (e.g. via start($request, $lazy=false)).
We already have a distinction between lazy starting via init(), and force starting via start().
2018-07-19 13:32:04 +12:00
Daniel Hensby
b83281fa5e
Merge pull request #8265 from gbaumeister/4.2
Added install-requirement-check for php-intl
2018-07-16 17:31:32 +01:00
Gerald Baumeister
df5395b101 Added check for php-intl requirement 2018-07-16 18:08:25 +02:00
Loz Calver
b317bf163c
Merge pull request #8213 from creative-commoners/pulls/4.2/falsy-forms
FIX FormField::Link works when no form is currently set
2018-07-16 13:53:58 +01:00
Robbie Averill
b93e94c0c3
FIX FormField::Link now throws a LogicException if no form is set yet 2018-07-16 13:02:58 +01:00
Loz Calver
106ca6643a
Merge pull request #8263 from dhensby/pulls/4.1/mask-backtrace
FIX updateValidatePassword calls need to be masked from backtraces
2018-07-16 10:30:22 +01:00
Damian Mooyman
c6952b46c4
Merge pull request #8245 from dhensby/pulls/4.2/http-cache-fixes
HTTPCacheControlMiddleware fixes
2018-07-16 13:13:12 +12:00
Robbie Averill
d122995652 FIX Duplicate config values for cascade_duplicates no longer duplicate their duplicates
Previously you could define identical values for this config prop via a DataExtension and on the base
class, resulting in double duplication
2018-07-16 12:04:56 +12:00
Daniel Hensby
8703839eb1
FIX updateValidatePassword calls need to be masked from backtraces 2018-07-15 01:06:45 +01:00
Daniel Hensby
bd2abc7a12
Merge branch '4.1' into 4.2 2018-07-13 16:44:23 +01:00
Daniel Hensby
ec9281ee02
Merge branch '4.0' into 4.1 2018-07-13 16:42:00 +01:00
Loz Calver
f0e545ada8
Merge pull request #7971 from dhensby/pulls/4.0/fix-extension-config
FIX Config from extension should not override
2018-07-13 16:15:18 +01:00
Daniel Hensby
d80ef3d9e6
DOCS Update docs to reflect true config merge priorities 2018-07-13 15:43:56 +01:00
Daniel Hensby
4acec33562
FIX Fixed bug in config merging priorities so that config values set by extensions are now least important instead of most important 2018-07-12 00:55:39 +01:00
Daniel Hensby
28f2ceac6f
TEST Add tests for (im)mutablity of states 2018-07-10 16:58:08 +01:00
Daniel Hensby
fd8448889c
State default should be state enabled (no-cache is an enabled state) 2018-07-10 16:58:08 +01:00
Daniel Hensby
601bb4d768
Make augmentState method more efficient 2018-07-10 16:58:07 +01:00
Daniel Hensby
399ebd0031
FIX If theres a max-age set remove no-cache and no-store 2018-07-10 16:58:07 +01:00
Ingo Schommer
2e6f29fea0
Merge pull request #8223 from dhensby/pulls/4.0/remove-email-from-pswdrecovery
FIX remove personal information from password reset confirmation screen
2018-07-06 11:27:03 +12:00