In particular this crash can be observed when using the
bootstrap GUI to download the base graphics.
In ClientNetworkContentSocketHandler::OnReceiveContentInfo
ClientNetworkContentSocketHandler::callbacks is iterated, using an iterator
cb->OnReceiveContentInfo() is called (cb is of type BootstrapAskForDownloadWindow)
This calls new BootstrapContentDownloadStatusWindow()
This inherits from BaseNetworkContentDownloadStatusWindow
The constructor of which calls _network_content_client.AddCallback(this)
This reallocates the std::vector which is being iterated in ClientNetworkContentSocketHandler::OnReceiveContentInfo
This results in iter being invalid, and an assertion failure occurs shortly
afterwards due to its use in the next iteration of cb->OnReceiveContentInfo()
Adjust all locations where ClientNetworkContentSocketHandler::callbacks
is iterated to avoid problematic behaviour
The last use was for storing a list of memory blocks. As the way these lists are accessed is very
specific, it is easier to just write an explicit destructor instead of trying to exactly match the behaviour.
DropDownListItem are strongly managed using std::unique_ptr to ensure leak-free handling. Appropriate use
of move-semantics make intent a lot clearer than parameter comments and allows the compiler to generate
copy-free code for most situations.
A conforming compiler with a valid <mutex>-header is expected.
Most parts of the code assume that locking a mutex will never fail unexpectedly,
which is generally true on all common platforms that don't just pretend to
be C++11. The use of condition variables in driver code is checked.
It is skipped when NPF is in use.
It is trivial to work around by adding and removing dummy orders.
It is mostly alleviated by the ship path cache in YAPF.
In the autoreplace window, the rail type drop down is for choosing engines
of the given time. Many rail types do not have engines specifically designed for them,
and are merely compatible with other rail types. This list is thus unwieldy and many
options have no engines available.
As this drop down is for choosing _engine_ rail type rather than compatible rail types,
we can list just the rail types explicitly listed by engines.
Linkgraph nodes require a specific order that was maintained by swapping just the last
element for the node to be removed. std::vector::erase() changed this to removing the
node is then shuffling the remain items down, which upsets other references to this
indices.
This is fixed by switching back to the original swap & pop method.
Display of industry production around tiles (as shown when placing a station)
did not take account of the station catchment changes, so still showed production
from an industry even if it was not covered by a tile.
This is fixed by making a set of nearby industries that are covered, instead of
looping over all possible industries.
finnish: 2 changes by hpiirai
english (us): 17 changes by Supercheese
luxembourgish: 10 changes by Phreeze
dutch: 20 changes by JanWillem
norwegian (bokmal): 2 changes by Leifbk
latin: 16 changes by Supercheese
portuguese: 3 changes by JayCity
The use of std::none_of in network/core/host.cpp is driven by the non-const
comparison operator use by NetworkAddress. A future commit should address
the const_casts in that class to ensure const-correctness.
The public and protected interface to SmallVector are unchanged
SmallVector now requires that items be default constructible
This isn't an issue since some contained items were previously created
uninitialized.
Temporary default constructors are added to the following structs
- SmallPair
- SmallStackItem
- GRFPresence
Where vector<bool> is required, transition immediately to std::vector
to avoid returning proxy object references.
Sign width was only updated when the text was changed. This seems to work for player-placed
signs as there is always a rename operation, however AIs can create a sign with text in one
go, in which case the width was never set.
finnish: 12 changes by hpiirai
greek: 40 changes by Jubilee
luxembourgish: 2 changes by Phreeze
dutch: 12 changes by JanWillem
french: 4 changes by romazoon
norwegian (bokmal): 12 changes by Leifbk
hungarian: 14 changes by Brumi
portuguese: 51 changes by JayCity, 11 changes by vesgo
This is a C++11 feature that allows the compiler to check that a virtual
member declaration overrides a base-class member with the same signature.
Also src/blitter/32bpp_anim_sse4.hpp +38 is no longer erroneously marked
as virtual despite being a template.
Measurement tooltip was auto-closed as the hover/right-click test for tooltips was not
satisfied in this case. This is fixed by keeping the tooltip visible and instead explicitly
closing the tooltip when the PlaceObject is cancelled/completed.
This switch has been a pain for years. Often disabling broke
compilation, as no developer compiles OpenTTD without, neither do
any of our official binaries.
Additionaly, it has grown so hugely in our codebase, that it
clearly shows that the current solution was a poor one. 350+
instances of "#ifdef ENABLE_NETWORK" were in the code, of which
only ~30 in the networking code itself. The rest were all around
the code to do the right thing, from GUI to NewGRF.
A more proper solution would be to stub all the functions, and
make sure the rest of the code can simply assume network is
available. This was also partially done, and most variables were
correct if networking was disabled. Despite that, often the #ifdefs
were still used.
With the recent removal of DOS, there is also no platform anymore
which we support where networking isn't working out-of-the-box.
All in all, it is time to remove the ENABLE_NETWORK switch. No
replacement is planned, but if you feel we really need this option,
we welcome any Pull Request which implements this in a way that
doesn't crawl through the code like this diff shows we used to.
In 10 years there was no active development on DOS. Although it
turned out to still work, the FPS was very bad. There is little
interest in the current community to look into this.
Further more, we like to switch to c++11 functions for threads,
which are not implemented by DJGPP, the only current compiler
for DOS.
Additionally, DOS is the only platform which does not support
networking. It is the reason we have tons of #ifdefs to support
disabling networking.
By removing DOS support, we can both use c++11 functions for threads,
and remove all the code related to disabling network. Sadly, this
means we have to see DOS go.
Of course, if you feel up for the task, simply revert this commit,
and implement stub c++11 functions for threads and stub functions
for networking. We are more than happy to accept such Pull Request.
GetAircraftFlightLevel<Aircraft> is only used in static functions
inside aircraft_cmd.cpp. With GCC, Clang and MSVC this is not an
issue, but on ICC fails linking, because it doesn't find this
version of this template. Possibly these two pieces of information
are linked.
Explicit defining the function fixes the issue.
It is only an error if the invalid result is actually used. This will be silently ignored at the moment.
It is still an error if a duplicate cargo type is returned.
Currently this can only be triggered by NewGRF house tiles querying for cargo acceptance history
of nearby stations (var 0x64) with a tile offset, and providing an offset that happens to point
to an industry tile. This serves no useful purpose.
_current_company is not currently logged anywhere in the crashlog.
_local_company is logged, despite being much less useful than
_current_company.
This change logs _current_company alongside _local_company.
If it was compiled with MingW, both / and \ were accepted as
path separator. On MSVC, only \ was. This is an unexpected
difference between binaries for the same platform. Remove this
discrepancy by accepting both / and \ on all platforms.
