english (au): 1 change by krysclarke
english (us): 1 change by 2TallTyler
chinese (simplified): 9 changes by WenSimEHRP
korean: 1 change by telk5093
russian: 1 change by Ln-Wolf
finnish: 1 change by hpiirai
danish: 2 changes by bscargo
french: 2 changes by ottdfevr
portuguese: 1 change by azulcosta
chinese (simplified): 141 changes by WenSimEHRP
russian: 3 changes by Ln-Wolf
danish: 6 changes by bscargo
latvian: 10 changes by lexuslatvia
dutch: 16 changes by Afoklala
french: 2 changes by ottdfevr
chinese (simplified): 16 changes by WenSimEHRP
arabic (egypt): 23 changes by AviationGamerX
korean: 1 change by telk5093
portuguese (brazilian): 10 changes by pasantoro
polish: 4 changes by pAter-exe
english (au): 7 changes by krysclarke
english (us): 7 changes by 2TallTyler
chinese (simplified): 6 changes by WenSimEHRP
serbian: 39 changes by DoLoop216
russian: 3 changes by Ln-Wolf
finnish: 4 changes by hpiirai
portuguese: 10 changes by azulcosta
polish: 34 changes by pAter-exe
english (au): 3 changes by krysclarke
english (us): 3 changes by 2TallTyler
chinese (simplified): 5 changes by WenSimEHRP
korean: 3 changes by telk5093
russian: 3 changes by Ln-Wolf
finnish: 3 changes by hpiirai
french: 5 changes by Lishouuu
Delay the nearest depot order search for a day if the vehicle can't find its destination, which happens when it has already attempted to do so and failed to find a valid destination.
The brings some performance advantages:
* No need to iterate all vehicles and check for primary vehicle as only vehicles that can have orders are listed.
* Shared orders only need to be tested once instead of for each vehicle sharing them.
* Vehicle tests only need to be performed on the first shared vehicle instead of all.
There could be a callback in _new_http_callbacks that is not
processed yet. All callbacks in _http_callbacks were cancelled,
but not the ones in _new_http_callbacks
english (au): 2 changes by krysclarke
chinese (simplified): 6 changes by WenSimEHRP
russian: 2 changes by Ln-Wolf
finnish: 2 changes by hpiirai
dutch: 6 changes by Afoklala
portuguese: 5 changes by azulcosta
portuguese (brazilian): 2 changes by ericandradex
polish: 4 changes by pAter-exe
This avoids needing to iterate the complete vehicle pool. Company group statistics are maintained elsewhere already.
The vehicle pool is still iterated later to find the nth random road vehicle.
If not enough parameters are supplied for a string, then a value of 0 was used, which could result in incorrect information being displayed.
Instead, throw an exception and include an error in the string.
This is specifically for s-cedilla and t-cedilla to their comma variants.
These variants, especially in smaller font sizes, look almost identical but
they are different. Currently the translation uses a mix of the cedilla and
comma variants, where the cedilla ones are often in the older strings.
Replace reinnoi with innoi, as it is the correct form.
* Fix #11644: Off by one error in StrMakeValid UTF-8 decode overrun detection
* Fix #11644: Off by one error in StrMakeValid buffer last character
* Fix: Unnecessary string duplication at StrMakeValid call sites
galician: 13 changes by Xocko12
catalan: 47 changes by J0anJosep
latvian: 14 changes by lexuslatvia
french: 11 changes by ottdfevr
portuguese: 52 changes by azulcosta
Engine age in months was calculated as the difference in days / 32, instead of the actually difference in months. This would result in engines being artificially younger if a game was started at a later date.
NWidgetMatrix modifies its child widget's index to indicate which element
is to be drawn, which now causes issues with code that does not know about
stuffing extra data into the index.
Instead, let NWidgetMatrix store the currently processing element, and
retrieve this information from the matrix widget while child widgets are
being drawn.
This means only widgets that are children of NWidgetMatrix need to know
anything about their extra data.
Orders window has different widget layouts depending on vehicle type
which don't all have the same widgets, and therefore it tries to disable
widgets that might not exist.
Restore the old behaviour of ignoring such requests, instead of crashing.
Empty area at the top of some textfile windows due to calling
SetDisplayedPlane() after calling FinishInitNested(), and/or changing
the displayed plane and not calling ReInit() after.
This was previously hidden by CheckForMissingGlyphs() reinitialising
all windows anyway.
This changes from naming storage-type to naming functionality.
* `FillNestedArray` is renamed to `FillWidgetLookup`.
* `Window::nested_array` is renamed to `Window::widget_lookup`.
* `array` parameter renamed as well.
Having two ways (`FillNestedArray` and `SetupSmallestSize`) to initialize
`Window::nested_array` introduces confusion.
Instead, make `FillNestedArray` the canonical way, always call it, and remove
init_array from `SetupSmallestSize`.
english (au): 2 changes by krysclarke
chinese (simplified): 37 changes by WenSimEHRP
romanian: 28 changes by bnegrut
russian: 13 changes by Ln-Wolf
finnish: 36 changes by hpiirai
danish: 16 changes by bscargo
dutch: 17 changes by Afoklala
Peter Nelson
translators
Loïc Guilloux
Jonathan G Rennison
Tyler Trahan
Rubidium
Patric Stout
Kuhnovic
Kaspars Putniņš
SamuXarick
glx22
EmperorJake
George Abraham
Krys Clarke
Charles Pigott