2a6412d5ae
Merge remote-tracking branch 'origin/release/17.3' into release/17.4
2026-06-02 16:22:05 +02:00
bef69b6aaf
Parse query params as a separate options hash, not kwargs
2026-06-02 16:21:14 +02:00
cc475e6c2e
[ #71645 ] Improve interface of WP identifier setting controller
2026-04-16 20:35:34 +02:00
f9d8bc6614
Introduce SubclassResponsibility error
...
This error is intended for cases when a method is
intentionally not implemented, because the module/class defining
it expects a subclass (or class including the module) to implement
the method.
This is intended to distinguish it from other cases, such as:
* feature not implemented yet
* edge case of a method call not yet supported
Notably it avoids the misuse of the Ruby-defined NotImplementedError,
which is only intended for much more specific scenarios:
> Raised when a feature is not implemented on the current platform. For example, methods depending on the fsync or fork system calls may raise this exception [...]
Also see https://docs.ruby-lang.org/en/master/NotImplementedError.html
2026-03-27 08:14:56 +01:00
45b61771a7
Document single-job assumption on ApplyHandlesJob
...
The status polling UI assumes at most one active instance of this job
at a time. Add a FIXME to enforce this with GoodJob concurrency
control once the real migration body is implemented.
2026-03-07 12:54:27 +03:00
a5ea92db04
Unify identifier settings form into single state-machine component
...
Replace the two-component approach (IdentifierSettingsFormComponent +
IdentifierChangeInProgressComponent) with a single component cycling
through three lifecycle states: :edit, :change_in_progress, :completed.
2026-03-06 09:21:17 +03:00
80e6aa4657
Do not update the attachment on initial journals
...
https://community.openproject.org/work_packages/71403
2026-02-10 14:39:23 +01:00
3aa5275567
Merge pull request #21214 from opf/feat/workflow-work-package
...
Recreate artifact on status change
2025-12-08 14:08:55 +01:00
8c0241f567
Aggregate migration 20241120095318_update_scheduling_mode_and_lags
2025-11-24 14:19:33 +01:00
7b2fde2eb1
remove job no longer referenced
2025-07-22 14:54:30 +02:00
2b68eb08ad
squash migrations below 15.0
2025-07-22 14:54:28 +02:00
bc693e02ea
Fix missing copyright notices in app/workers
2025-07-18 18:10:42 +01:00
03572f4444
Freeze string literals in app/workers
...
rubocop -A --only Style/FrozenStringLiteralComment,Layout/EmptyLineAfterMagicComment,Style/RedundantFreeze app/workers
2025-07-18 17:42:42 +01:00
4b6056b333
refactor: Extract methods to deal with % complete modes
...
To keep all future usages of "work_weighted_average" and
"simple_average" magic strings at the same place.
2025-07-08 18:20:05 +02:00
a637335fe2
rename covering_dates_and_days_of_week to covering_dates_or_days_of_week
2025-06-04 18:33:32 +02:00
cde4bba693
extract common parts of ApplyWorkingDaysChangeJobs
2025-06-04 18:33:24 +02:00
54921b00d4
cache changed_days and changed_non_working_dates
2025-06-04 18:33:23 +02:00
3e8763173b
refactor method for producing a hash with differenc between two lists
...
Choosing combination of speed and clarity of intent:
```rb
require 'benchmark/ips'
require 'active_support/core_ext/enumerable'
def change_between(list_a, list_b)
set_a = Set.new(list_a)
set_b = Set.new(list_b)
(set_a ^ set_b).index_with { set_b.include?(it) }
end
def change_between1(list_a, list_b)
was = list_a.to_h { [it, false] }
now = list_b.to_h { [it, true] }
now.merge(was){ nil }.compact
end
def change_between2(list_a, list_b)
deleted = (list_a - list_b).to_h { [it, false] }
added = (list_b - list_a).to_h { [it, true] }
deleted.merge(added)
end
def change_between3(list_a, list_b)
set_a = Set.new(list_a)
set_b = Set.new(list_b)
deleted = (set_a - set_b).to_h { [it, false] }
added = (set_b - set_a).to_h { [it, true] }
deleted.merge(added)
end
def change_between4(list_a, list_b)
was = list_a.index_with(false)
now = list_b.index_with(true)
now.merge(was){ nil }.compact
end
def change_between5(list_a, list_b)
deleted = (list_a - list_b).index_with(false)
added = (list_b - list_a).index_with(true)
deleted.merge(added)
end
def change_between6(list_a, list_b)
set_a = Set.new(list_a)
set_b = Set.new(list_b)
deleted = (set_a - set_b).index_with(false)
added = (set_b - set_a).index_with(true)
deleted.merge(added)
end
def compare(list)
as_and_bs = 1000.times.map do
[list.sample(rand(list.length)), list.sample(rand(list.length))]
end
change_between_methods = private_methods.grep(/^change_between\d?$/)
as_and_bs.each do |a, b|
d = change_between(a, b)
change_between_methods.each do |change_between_method|
fail "#{change_between_method} returned different result" unless d
== send(change_between_method, a, b)
end
end
Benchmark.ips do |x|
change_between_methods.each do |change_between_method|
x.report(change_between_method) { as_and_bs.map{
send(change_between_method, _1, _2) } }
end
x.compare!
end
end
compare(7.times.to_a)
compare(100.times.to_a)
```
```
ruby 3.4.2 (2025-02-15 revision d2930f8e7a) +PRISM [arm64-darwin23]
Warming up --------------------------------------
change_between 40.000 i/100ms
change_between1 158.000 i/100ms
change_between2 170.000 i/100ms
change_between3 32.000 i/100ms
change_between4 152.000 i/100ms
change_between5 148.000 i/100ms
change_between6 33.000 i/100ms
Calculating -------------------------------------
change_between 408.732 (± 2.7%) i/s (2.45 ms/i) - 2.080k in 5.093365s
change_between1 1.588k (± 0.7%) i/s (629.69 μs/i) - 8.058k in 5.074251s
change_between2 1.645k (±12.2%) i/s (607.90 μs/i) - 8.160k in 5.048118s
change_between3 318.922 (± 2.8%) i/s (3.14 ms/i) - 1.600k in 5.021512s
change_between4 1.477k (± 1.3%) i/s (676.94 μs/i) - 7.448k in 5.042753s
change_between5 1.608k (± 8.5%) i/s (621.98 μs/i) - 7.992k in 5.012654s
change_between6 331.757 (± 1.8%) i/s (3.01 ms/i) - 1.683k in 5.074711s
Comparison:
change_between2: 1645.0 i/s
change_between5: 1607.8 i/s - same-ish: difference falls within error
change_between1: 1588.1 i/s - same-ish: difference falls within error
change_between4: 1477.2 i/s - same-ish: difference falls within error
change_between: 408.7 i/s - 4.02x slower
change_between6: 331.8 i/s - 4.96x slower
change_between3: 318.9 i/s - 5.16x slower
ruby 3.4.2 (2025-02-15 revision d2930f8e7a) +PRISM [arm64-darwin23]
Warming up --------------------------------------
change_between 4.000 i/100ms
change_between1 10.000 i/100ms
change_between2 10.000 i/100ms
change_between3 4.000 i/100ms
change_between4 9.000 i/100ms
change_between5 11.000 i/100ms
change_between6 4.000 i/100ms
Calculating -------------------------------------
change_between 45.431 (± 4.4%) i/s (22.01 ms/i) - 228.000 in 5.029300s
change_between1 112.354 (± 7.1%) i/s (8.90 ms/i) - 560.000 in 5.013557s
change_between2 133.466 (± 4.5%) i/s (7.49 ms/i) - 670.000 in 5.031395s
change_between3 44.247 (± 2.3%) i/s (22.60 ms/i) - 224.000 in 5.065967s
change_between4 104.443 (± 4.8%) i/s (9.57 ms/i) - 522.000 in 5.011400s
change_between5 114.941 (± 6.1%) i/s (8.70 ms/i) - 583.000 in 5.094324s
change_between6 46.083 (± 4.3%) i/s (21.70 ms/i) - 232.000 in 5.047312s
Comparison:
change_between2: 133.5 i/s
change_between5: 114.9 i/s - 1.16x slower
change_between1: 112.4 i/s - 1.19x slower
change_between4: 104.4 i/s - 1.28x slower
change_between6: 46.1 i/s - 2.90x slower
change_between: 45.4 i/s - 2.94x slower
change_between3: 44.2 i/s - 3.02x slower
```
2025-06-04 18:33:23 +02:00
c8bd1678b0
extract commond code for changed_days and changed_non_working_dates
2025-06-04 18:33:22 +02:00
3d673294c3
attr readers instead of accessors in WorkPackages::ApplyWorkingDaysChangeJob
2025-05-05 12:10:28 +02:00
6daf60e70e
Add missing magic comment # frozen_string_literal: true
2025-02-25 09:12:27 +01:00
b4d2f5dc9c
[59539] Correctly reschedule predecessors needing relations rescheduling
...
Edge case not detected by the previous algorithm: relation has no lag
but covers non-working days changed into working days. The successors
need to be rescheduled, which is done by rescheduling from the
predecessor.
2025-02-25 09:12:24 +01:00
3b5562fc5e
Simplify method calls
...
State is stored in instance variables instead of being passed around as parameters and return values.
2025-02-25 09:12:23 +01:00
4034896420
[59539] Keep manual scheduling mode in migration
...
The rule is: it never switches from manual to automatic scheduling mode.
It only switches from automatic to manual, and only if keeping automatic
is not possible because it would mean losing the dates.
The code and specs have been updated to reflect this.
A materialized view is used to reuse the data in multiple different
queries.
2025-02-25 09:12:21 +01:00
afdf4308b5
[59539] Only set lag for the closest relation in the migration
2025-02-25 09:12:21 +01:00
04d3f6651b
[59539] Respect non-working days when computing lags in migration
...
`lag` is the number of _working_ days between predecessor and successor
dates.
2025-02-25 09:12:21 +01:00
35f463f4bd
[59539] Migration: set lag for follows relations
...
To preserve dates, a lag is set for follows relations when both the
predecessor and the follower have dates.
2025-02-25 09:12:20 +01:00
dcaf5aa892
[59539] Migration: followers and parents have automatic scheduling mode
...
https://community.openproject.org/wp/59539
2025-02-25 09:12:20 +01:00
06f596cdf3
[42388] Migrate scheduling mode and lags
...
https://community.openproject.org/wp/42388
Scheduling mode is now manual by default. Only successors will be in
automatic mode.
WIP
2025-02-25 09:12:19 +01:00
381e2f4b09
[57990] Make migration work with database from before 14.0 version
...
https://community.openproject.org/wp/57990
The code to create the temporary table in common between the migration
and the job to update status and % complete values. It has recently
changed to get the `excluded_from_totals` value from the `statuses`
table, but this field has been created in 14.2, and does not exist yet
in 14.0 when this migration is run.
This commit adds the previous version of the code to create the
temporary table, that does not require the `excluded_from_totals` field.
It also avoids creating the depths table which is not necessary in this
case.
2024-09-26 15:11:51 +02:00
de86221361
Extract common operations to sql_commands
2024-09-25 13:54:31 -05:00
95cda6de29
Remove unnecessary update
2024-09-25 13:49:09 -05:00
946b57d9e5
Merge branch 'dev' into implementation/57529-options-for-total-calculation-sum-mode
2024-09-25 13:20:38 -05:00
3250f5104a
Compute total % complete according to mode in update status job
...
Reuse and duplicate a lot of logic from the other job. It needs to be
reworked, but it works
2024-09-25 17:13:35 +02:00
3f97171b79
Use work_package_hierarchies to create depth table
...
Simpler.
2024-09-25 17:12:34 +02:00
4ac1b1ca7a
Remove :percent_complete_edition feature flag
...
% complete edition will be released in 14.6
2024-09-25 15:31:11 +02:00
1f12393880
Account for status exclusion from total calculations in simple average mode
2024-09-24 16:56:36 -05:00
2f2561c08f
Add in simple average calculation mode job
...
* Adds temporary depth table - Could be cleaned up down thie line with
the `work_package_hierarchies` table.
2024-09-24 16:35:58 -05:00
6cbfdc8e78
Trigger the recalculation job when changing to work weighted mode
...
* Triggers job
* Adds journal entries with specs and translations
2024-09-24 11:10:22 -05:00
503ad154ca
Add recalculation in job when switching to work weighted mode
...
* Missing to add journal entries
2024-09-24 11:10:21 -05:00
0368f40b66
Fix rubocop Metrics/AbcSize warning
2024-09-23 16:38:04 +02:00
35b583bc35
refactor: rename methods to avoid abbreviations
2024-09-23 14:34:03 +02:00
eb9283507f
[57990] Preserve % Complete over Remaining work when migrating
...
https://community.openproject.org/wp/57990
In 14.0, Work and Remaining work were deemed more important than
% Complete. Now in 14.6, as % Complete becomes editable and more
important, we want to preserve it over Remaining work for users having
not yet migrated to version 14.0.
2024-09-23 14:34:03 +02:00
f0f1e9eca6
Clear % complete when work is 0h when switching to work-based mode
...
That prevent having something like W=0h RW=0h %C=60% which leads to an
error being displayed when the popover is open.
2024-08-23 15:49:16 +02:00
4647786916
Fix missed work derivation when switching progress calculation modes
...
If a work package has only remaining work being set and progress
calculation mode is changed from work-based to status-based, then work
package gets a % complete value.
Remaining work and % complete being both present, work has to be
derived.
2024-08-23 15:26:04 +02:00
4911b8a149
remove years from copyrights (except for COPYRIGHT file)
2024-07-31 15:02:49 +02:00
8c82acdccd
Update app/workers/work_packages/progress/apply_statuses_change_job.rb
...
Co-authored-by: Aaron Contreras <61627014+aaron-contreras@users.noreply.github.com >
2024-06-03 17:59:10 +02:00
5e4766f072
Merge branch 'dev' into feature/49409-status-exclude-from-totals
2024-06-03 17:56:41 +02:00
36108c86c0
Apply @ulferts feedback
2024-05-27 11:24:07 +02:00
42dce573e2
Makes sure jobs are eventually discarded when concurrency happens
2024-05-27 11:24:05 +02:00
Oliver Günther