Replay folders were named YYYY-MM-DD_NNNN, a per-machine sequence
counter. The same multiplayer match got a different folder name per
participant, and unrelated matches could share a name, making it hard
to correlate replays across players (rating aggregation, replay
sharing, bug reports) without parsing commands.txt first.
The engine already generates a matchID once per match, synced to all
participants before the match starts. Replay folders are now named
YYYY-MM-DD_<matchID>, so the same match produces the same folder name
everywhere.
The date comes from a new "startTime" attribute, set once by the host
(StartTime.js, alongside MatchID.js) and synced like matchID, instead
of each machine reading its own local clock. createDateIndexSubdirectory
is untouched (still used by OOS logs and serialization tests); a new
createReplaySubdirectory is added for replays only, following the same
collision-handling pattern (incrementing suffix) as the existing
function. This also means resuming a match (network rejoin, or a
singleplayer game reloaded via StartSavedGame, both of which reuse the
same matchID) saves an additional replay instead of overwriting the
original.
Fixes#8922
UV coordinates are constant so we don't need to upload them each frame.
Ideally we shouldn't have UV coordinates at all as we could use
gl_VertexID but it's not available on all supported GLSL versions.
The ogg library itself doesn't use off_t, mostly just int and long.
With off_t being problematic as Windows doesn't support _FILE_OFFSET_BITS
replacing this removes a potential headache.
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
GTK static box doesn't have padding, so add a style variable to allow
adding a custom padding.
Replace implicit padding with explicit padding where needed.
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
We need to check for storage support because we outpout result into a
storage. Note: it seems we might have a bug inside GL backend or inside
Mesa drivers which prevents us from running resolve on compute.
The new system allows to register all function objects as Input::Handler
not only function pointers.
To not get dangling references a handler is unsubscribed on it's
destruction.
The order in with the handlers are executed has to be specified by the
slot-argument, instead of the reverse order of registration.
The LOS override for all players isn't serialized as such we have to
restore it when resetting a game.
Fixes: #8926
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
Sending default setting on first view of a section/tool to the engine
triggers the "you have changes" mechanism. Sending them just in case
isn't needed unless there is a bug in a different place.
Drop this preemptive measure.
Fixes: #3029
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
We don't use neither opendir nor readdir nor closedir. Further there are
no callers for init_libc().
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
Replace the wxBoxSizer with a wxFlexGridSizer so the gaps between main
elements, typically wxStaticBoxSizer, remain uniform across the tools.
Add a style variable to control the wxFlexGridSizer on platform basis.
Also replace the obsolete wxScrolledWindow by inheriting from
wxScrolled<wxPanel> instead.
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
Use `<filesystem>` instead of `truncate` and remove `truncate`
portability wrapper.
Add a helper function `StatusFromSystemError` to convert error_code to
`Status`.
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
Drop the use of lowlevel interfaces and consitently use the cpp
interface decupling it from systm dependent code.
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
`this.events` is only valid during one turn. When it is used in the next
turnn it might be outdated. So it makes no sense to keep it alive untill
the next turn.
There is a workaround for an old bug with the use of wxCollapsiblePane.
As the widget itself is no longer in use, just remove the now dead code.
Signed-off-by: Ralph Sennhauser <ralph.sennhauser@gmail.com>
