I could honestly not make much sense of the timer implementation
that was here. Instead I've implemented the type of timer algorithm
that I've used before, and tested it instead.
The concept is extremely simple: all timers are put in an ordered
list. We check every server tick if any of the timers have
elapsed, and execute the function associated with this timer.
We know that many timers by themselves cause new timers to be
added to this list, so we iterate *backwards* over the timer
list. This means that new timers being added while timers are
being executed, can never be executed in the same function pass,
as they are always appended to the table *after* the end of
the table, which we will never reach in the current pass over
all the table elements.
We switch time keeping to minetest.get_us_time(). dtime is
likely unreliable and we have our own high-res timer that we
can fix if it is indeed broken. This removes the need to do
any sort of time keeping.
clock_gettime() is a far better clock than gettimeofday().
Even better than clock_gettime() is that you can select either
CLOCK_MONOTONIC, or even CLOCK_MONOTONIC_RAW. These clocks offer
high precision time. And the _RAW variant will never roll back
due to NTP drift or daylight savings, or otherwise.
I've adjusted this code to select the right clock method auto-
matically based on what's available in the OS. This means that
if you're running a very old linux version, MacOS or other,
you will automatically get the best clocksource available.
I've tested all Linux clocksources by selectively compiling and
running a 10k+ timer test suite. In all cases I confirmed that
the 3 POSIX Linux clocksources worked properly, and were
selected properly.
I've modified the OS X compile path to use the high-res clock
source for all time functions, but I can't confirm it works or
that it compiles.
As for WIN32, I confirmed that the used clocksource is indeed
a Monotonic clocksource, so good news: that code section appears
to be exactly what it should be.
`errorstream` must not be overly verbose as clientside it is directly printed
onto the ingame chat window. These days, the serverlist can contain > 200k bytes,
so better print it to warningstream if the data buffer is too long.
Outdated servers are always sending tiledefs with culling
enabled no matter what, as the value was previously entirely
ignored.
To compensate, we must (1) detect that we're running against
an old server with a new client, and (2) disable culling for
mesh, plantlike, firelike and liquid draw types no matter what
the server is telling us.
In order to achieve this, we need to bump the protocol version
since we cannot rely on the tiledef version, and test for it
being older. I've bumped the protocol version, although that
should have likely happened in the actual change that introduced
the new backface_culling PR #3578. Fortunately that's only 2
commits back at this point.
We also explicitly test for the drawtype to assure we are not
changing the culling value for other nodes, where it should
remain enabled.
This was tested against various pub servers, including 0.4.13 and
0.4.12.
Fixes#3598
Backface culling is enabled by default for all tiles, as this
is how the lua parser initializes each tiledef. We revert to
always using the value from the tiledef since it is always
read and serialized.
Mods that wish to enable culling for e.g. mesh nodes, now can
specify the following to enable backface culling:
tiles = {{ name = "tex.png", backface_culling = true }},
Note the double '{' and use of 'name' key here! In the same
fashion, backface_culling can be disabled for any node now.
I've tested this against the new door models and this properly
allows me to disable culling per node. I've also tested this
against my crops mod which uses mesh nodes where culling needs
to be disabled, and tested also with plantlike drawtype nodes
where we want this to continue to be disabled.
No default setting has changed. The defaults are just migrated
from nodedef.cpp to c_content.cpp.
The save button is now fully functional again when an error message
is shown.
After an invalid value is entered in the settings tab dialog, the GUI
label for the error message that is shown was partly overlapping the
'save' button, so that the top half of the button could not be clicked
on.
cloneMesh() has to use a switch in order to create a different
mesh buffer type depending on vertex type. (Credit: the new cloneMesh
was written by RealBadAngel.)
To avoid repetitive code, all other methods use getVertexPitchFromType()
to automatically adapt the indexing to the vertex type at runtime.
Changes:
- Accept setting an empty flags-type value in the settings tab
if the variable specification permits it
- Don't accept substrings of flag values
E.g. with values: 'one,two,three', 'hree', 'w', etc. used to
be accepted. Not any more
- Don't accept flags with random pattern-matching special characters
E.g. with values: 'one,two,three', 'on.', '(o)[n]e*' etc. used
to be accepted. Not any more.
- Accept numbers prefixed with '+'
- Accept multiple spaces instead of just a single one where spaces are expected
- Allow flags to have an empty default value
The pageflip mode requires a stereo quadbuffer, and a modern graphic
card. Patch tested with NVidia 3D Vision.
The mini-map is not drawn, but that's what is done for topbottom and
sidebyside modes as well.
Also most of the time the user would prefer the HUD to be off. That's
for the user to decide though, and toggle it manually.
Finally, the interocular distance (aka eye separation) is twice as much
as the "3d_paralax_strength" settings. I find this a strange design
decision. I didn't want to chance this though, since it's how the other
3d modes interpret this settings.
Previously, we assumed that reply->str was NULL
terminated. However, this turned out to be not true,
as users reported crashes in strlen connected to
where reply->str was appended to an std::string.
Use the method recomended by the docs, to read the
length separately.
* Add main page.
* Organize into sections.
* Add threading sources.
* Include SpatialAreaStore, LevelDB/Redis, sound, FreeType, and cURL in output.
* Add logo.
* Fix project name hardcoding.
* Remove PAPER_TYPE (only used when GENERATE_LATEX is enabled).
* Have dot render graphs as SVG (smaller, and works even if dot's text drawing functionality is broken).
* Enable built-in STL support.
* Enable search bar.
* Switch from header-bar based navigation to treeview based navigation.
* Enable dynamic HTML (collapses graphs).
* Enable generation timestamp.
