Change: Store station blocked/wires/pylons flags in map. (#11337)

This stores three flags in unused map bits, and avoids having to look up
station graphics and custom station specs to determine blocked/wires/pylons
status.

This potentially affects rail pathfinding performance.

Savegame version is not bumped, as the flags can just be updated every
time.

Update: Translations from eints
korean: 1 change by telk5093

Update: Translations from eints
italian: 25 changes by Rivarossi

Update: Translations from eints
finnish: 1 change by hpiirai

Update: Translations from eints
chinese (simplified): 139 changes by BYXxiaonanhui
catalan: 23 changes by J0anJosep
portuguese (brazilian): 1 change by pasantoro

Update: Translations from eints
english (au): 1 change by krysclarke
russian: 1 change by Ln-Wolf
danish: 1 change by bscargo
dutch: 1 change by Afoklala
french: 1 change by ottdfevr
portuguese: 1 change by azulcosta
polish: 68 changes by pAter-exe

Fix: #11329: Don't assert vehicle list length is non-zero when only asked to set string parameter. (#11330)

Since #11321, the vehicle list is not yet initialized when SetStringParameters is called, so a test that the vehicle list size is zero is triggered. However, doing this check in the SetStringParameters function is a bit out of its remit, so just remove it.

Codechange: Use std::bitset instead of bool array.

This avoids use of C-style pointers and memset.

Codechange: List fallback station names by ID instead of magic number.

This improves readability and intention of the station name generator.

Update: Translations from eints
danish: 20 changes by bscargo

Codechange: Don't access cargo filter before it is initialized. (#11321)

The cargo filter list is initialized during window OnInit, but the first
build of the filtered list occurred before this.

Codechange: Don't pass GUI cargo filter constants to command system. (#11320)

CMD_BUILD_VEHICLE expects CT_INVALID to indicate a new vehicle should not
be refitted on build, but the code passed CF_NONE instead. CF_NONE has the
same numeric value at CT_INVALID, so the test passed, but...

Codechange: All ParagraphLayoutFactory::AppendToBuffer assume that the buffer has at least some space.
Assert on that.

Fix: only count distance traveled in vehicles for cargo payment (#11283)

No longer you can utilize the free (and instant) labour of station
workers, transporting your cargo from one part of the station to
the other. No more!

Based on patch by dP.

Update: Translations from eints
korean: 1 change by telk5093

Fix #11315: Sort industries and cargoes by name* in industry chain window.

*Cargo types are sorted by the normal method so it's not strictly alphabetical.

Fix: Don't produce invalid cargo. (#11314)

Production for secondary industries did not always check that the produced cargo type isn't valid.

Codechange: use better location for the "invalid" action sentinel value

It used to be a random sentinel for end-of-(widget-)list that was used to tell
that no action has taken place yet. Since the last action is practically the
widget that was pressed, add the sentinel to that enumeration.

Fix #11307: Incorrect GroupStatistics after selling leading wagon

When this results in a countable consist

Codechange: Rename INVALID_CARGO to INVALID_CARGO_BITNUM

... to avoid future ambiguity.

Update: Translations from eints
english (us): 20 changes by 2TallTyler

Update: Translations from eints
romanian: 18 changes by bnegrut
turkish: 22 changes by densxd
dutch: 20 changes by Afoklala
portuguese: 21 changes by azulcosta

Codechange: Rename RoadType parameter of MakeRoadDepot

See: #9642

Fix: compilation failed on gcc 10.2 due to missing include. (#11300)

'_exit' is defined in 'unistd.h'.

Update: Translations from eints
english (au): 20 changes by krysclarke
korean: 25 changes by telk5093
russian: 20 changes by Ln-Wolf
finnish: 20 changes by hpiirai
french: 20 changes by glx22
portuguese (brazilian): 20 changes by pasantoro

Add: Install additional documentation files with the game

Also include it in Emscripten packages

Update: Translations from eints
swedish: 1 change by joeax910

Doc: Add more hyperlinks within the included documentation (#11288)

Co-authored-by: Niels Martin Hansen <nielsm@indvikleren.dk>

Update: Translations from eints
swedish: 2 changes by optiedev
vietnamese: 11 changes by KhoiCanDev
romanian: 5 changes by bnegrut
danish: 2 changes by bscargo
dutch: 2 changes by Afoklala
portuguese: 2 changes by azulcosta
polish: 15 changes by pAter-exe

Codechange: Use std::copy/fill pattern to initialize rail and road specs. (#11285)

This avoids use of lengthof and array indices.

Fix: don't compare next_station when trying to merge CargoPackets

For vehicle packets they shouldn't be compared, and for station
packets they are already in a bucket per next_station.

Feature: [NewGRF] Related Act2 objects for airports and airport tiles.

Airports are similar two stations and industries, both of which have the town as related object.
Airport tiles are similar to industry tiles, which have the industry as related object.
This seems a sensible structure, so let's make it Airport Tile -> Airport -> Town.

Add: [NewGRF] Inspection window for airports.

As as the station window combines all station types, accessing the
debug view is via the parent of the airport tile only.

Fix b0e73277: cargodist information got lost when splitting of cargo (#11280)

During b0e73277 we removed loaded_at_xy, but I kinda forgot that
it was a union with next_station. Now next_station wasn't copied
anymore, or checked in AreMergable.

Codechange: cleanup CargoPacket in terms of variable/function names (#11278)

Over the years, things got reused and changed, making the current
names somewhat unclear in what they actually mean and do.

Codechange: Reduce size of TextEffectMode and TextEffectID.

TextEffectID was promoted to size_t in #9235, when it was used in loops.
However since then the relevant code uses range-for or iterators instead.

Codechange: Make INVALID_TE_ID a TextEffectID instead of a TextEffectMode.

Type-correctness?

Codechange: Don't allocate a text effect with INVALID_TE_ID.

Previously, despite INVALID_TE_ID existing, it was not checked during allocation and there was no limit to the number of text effects.