32 Commits (8da6b7e649438dda06be2716ba16ba3602594bde)

Author	SHA1	Message	Date
Oliver Günther	7cf11bb689	Update copyright notice	8 years ago
Jens Ulferts	8065c66256	eager loading associated objects, the dirty way	9 years ago
Jens Ulferts	3734b3a232	limit aggregation to one journable ... With only .where("#{table_name}.id = ?", notes_id) appended, the DB would aggregated all (from all journables of all types) and then select the one that is necessary. This takes a considerable amount of time. The commit will limit the aggregation to one journalizable.	9 years ago
Jan Sandbrink	26574521d3	Prefill predecessors for results of multiple journables too	9 years ago
Jan Sandbrink	0343e12c83	Prefill predecessor when it is already known ... Getting the predeccessor of a journal is a common operation. When computing the journals for a single journable, it is easily possible to prefill the predecessor (because we know all the journals). Doing this avoids an N+1 query later on (in case the journals are used to display diffs).	9 years ago
Jan Sandbrink	dfb9a5cd20	remove remaining AR find_by_x finders ... - most of them were removed by @myabc in the rails4 branch - some where reintroduced due to parallel development to that branch - still some occurences left that will hit custom implementations	9 years ago
Jan Sandbrink	b322fafc25	remove workaround for type conversion in PSQL ... as of Rails 4 this should no longer be neccessary	9 years ago
Jan Sandbrink	3fe8b33f78	enforce disabled aggregations ... due to limited time resolution we can't assume that a setting of 0 will implicitly disable aggregation	9 years ago
Jan Sandbrink	d47d6685ed	filter for right journal on SQL level	9 years ago
Jan Sandbrink	f9051bba3e	display diff of containing aggregated journal ... ... instead of displaying the diff for the requested journal only - this should provide the best forward compatibility of links - even when aggregation timeframes are changed intermittently	9 years ago
Jan Sandbrink	e405605343	fix wrongful notificaiton delivery	9 years ago
Jan Sandbrink	285b7f5437	move on-behalf-delivery into notification mailer	9 years ago
Jan Sandbrink	974523c370	partly address an edge case that can prevent notification delivery ... "partly" because the current implementation is still "vulnerable" if the succeeding journal gets outdated in the same way as the preceeding one. The edge-case of the edge-case....	9 years ago
Jan Sandbrink	6c1e4dbf4f	send aggregated notifications ... - only sending the last mail of an aggregation chain - displaying the aggregated state of the change - replaces old classes for mail notifications - old special error handling for missing users is obsoleted, since recipients are determined at jobs runtime - special cases for missing author and journal are considered	9 years ago
Jan Sandbrink	1ff8f15fbd	explicitly define order of groups in PSQL ... Always remember kids: SQL databases do not guarantee an order if you don't ask for one! That is why PostgreSQL sometimes found a predecessor for an aggregated journal that was not supposed to have one: Based on internal decisions the order of rough groups was not ascending (as expected implicitly) and master and addition would switch places. ALWAYS SPECIFY AN ORDER IF YOU EXPECT ONE!	9 years ago
Jan Sandbrink	a2b952f549	improve exception message on missing journable.id ... previously an invalid MySQL statement would have been indicated, which is correct but not very helpful. On PSQL this will also lead to a DB lockup (at least on environments known to me)	9 years ago
Jan Sandbrink	cefa25b189	fix predecessor implementation	9 years ago
Jan Sandbrink	563cb3c5c7	migrate APIv3 to aggregated journals ... - turns out agrgegated journals needed more features - turns out specs were incomplete and bogus	9 years ago
Jan Sandbrink	f65f9e3e2d	remove the AR portion of aggregated journals ... - this prohibits doing stupid things like saving aggregated journals - should make clear that aggregated journals are not real journals...	9 years ago
Jan Sandbrink	3a87b269fa	allow to fetch aggregated journals for a journable ... - this should be faster (or at least easier to optimize) than applying a WHERE to the whole scope - also starting to diverge from using classic ActiveRecord in aggregated journals	9 years ago
Jan Sandbrink	a8ab1024b5	clarify explanations ... typos, wording and stuff	9 years ago
Jan Sandbrink	5f06a8900d	better PSQL compatibility for aggregated journals	9 years ago
Jan Sandbrink	f02b32faad	override :initial? for AggregatedJournal ... the simple approach of looking at the version number does not work, because aggregated journals can have gaps in their versions (and there might be no version 1)	9 years ago
Jan Sandbrink	a5152e9875	use the actual setting for the aggregation time	9 years ago
Jan Sandbrink	88169a46de	expose id of "master" as notes_id ... - make spec assertions more detailed by comparing all the attributes	9 years ago
Jan Sandbrink	3c3b03bbcc	make SQL compliant for both MySQL and PSQL	9 years ago
Jan Sandbrink	405aa2d524	add PostgreSQL-compatible group numbering	9 years ago
Jan Sandbrink	3b7b0ec289	Enhance AggregatedJournals to correctly merge journals with notes ... Aggregation should now follow the rules previously defined for merging journals (see code comments). Jens (@ulferts) and I developed the SQL for this. I hope the explanatory comments make it a quite readable SQL experience.	9 years ago
Jan Sandbrink	ddbc8a4ba7	remove obsoleted helper	9 years ago
Jan Sandbrink	885d8d734f	add basic spec for aggregated journals	9 years ago
Jan Sandbrink	ec720a0533	make aggregation capable for MySQL and PSQL ... - also give the time threshold a useful meaning (number of seconds)	9 years ago
Jan Sandbrink	d6533b100c	first spike for aggregatable journals	9 years ago
Alex Coles	e0191e759c	Update year in copyright header to 2015 ... [ci skip]	10 years ago
Alex Coles	bb0e6e6aa5	Fix syntax (w/Rubocop) in (Rails) controllers ... Signed-off-by: Alex Coles <alex@alexbcoles.com>	10 years ago
Jens Ulferts	1513a19b84	readd experimental controllers and add authorization to them	10 years ago
Alex Coles	7ff6f4fe94	Fully qualify parent class in API v2 Controllers ... This fixes the following issue: Circular dependency detected while autoloading constant Api::V2::ProjectsController Signed-off-by: Alex Coles <alex@alexbcoles.com>	10 years ago
Alex Coles	1900e9b55d	Fix whitespace in (Rails) controllers ... Signed-off-by: Alex Coles <alex@alexbcoles.com>	11 years ago
Johannes Wollert	8b096975fe	updates copyright headers ... updates more copyright more copyright headers	11 years ago
Philipp Tessenow	a1e67dd460	new copyright header #1903	11 years ago
Martin Czuchra	5cba510b1f	Updates copyright notice in timelines sources. ... This also fixes some of the broken license headers after the previous merge.	12 years ago
Martin Czuchra	46ffcfc195	Refactors current timelines API into api/v2.	12 years ago