Also Disallow linking against non-C++ system Lua, unless configuring with -Dworkaround_noncpp_lua=true, add -Dworkaround_elusive_bzip2 and friends, and get rid of the -image_base hack for macos.
I really don't like how the only way to return with an error from ParseFloatProperty is via an exception >_>
Also do a range check on airTemp only if isValid is true, otherwise it's uninitialized.
This website API was created to enable TPTMP to prove the identity of connecting users, and while TPTMP works fine without explicit support for this from the game, it has to resort to parsing powder.pref. This is not only ugly but also likely to be disallowed by the next version of the script manager. This new script manager will probably come after 97.0, so it's okay for it to rely on a game feature that won't be available until 97.0.
get_normal_interp is given a PHOT and scans a line section starting from the PHOT, extended in the direction of its velocity, to determine the surface normal of the surface that is reflecting the PHOT. It uses is_boundary to detect "boundaries" on this line section, defined as one "blocking" cell with at least one "non-blocking" non-diagonal neighbour. It never actually makes sure the position it passes to is_boundary is within the simulation area, so I assume is_boundary is expected to handle this correctly.
Plot twist: it does not. It delegates checking whether a cell is "blocking" (defined as something PHOT would normally fail to move into as per eval_move, but GLAS and BLGA are handled specially) to is_blocking. is_blocking returns true for cells beyond the simulation area (as eval_move returns "no move" for such cells). Once is_boundary sees that the cell it's been given is blocking, it then proceeds to check whether any of its non-diagonal neighbours might be non-blocking. In most cases, non-diagonal neighbours of an out-of-bounds cell are also out of bounds and are thus blocking, but if the cell is_boundary is given is in the innermost layer of out-of-bounds cells, just beyond the simulation area, then it has a neighbour that is in bounds, and that one may not block PHOT. The takeaway is that out of bounds cells can indeed be boundaries, as far as is_boundary is concerned.
This is a problem because get_normal_interp's line section can easily reach beyond the simulation area if the PHOT's velocity is high enough, and if it finds a boundary along this line section, it immediately stops looking and passes its position to photoelectric_effect, which then uses this potentially out-of-bounds position to index pmap. A position with y = -1 causes photoelectric_effect to read from the last few slots of parts data that it then interprets as pmap entries, which then may direct it to particles to spark that are beyond parts. This eventually crashes.
This commit doesn't fix is_boundary's definition of boundaries, but it stops get_normal_interp looking at cells beyond the simulation area.
The crash is difficult to reproduce because there have to be many particles in the simulation for the very last slots of parts to be in use, and for them to point to memory that isn't accessible. PHOT also has to survive a try_move beyond the simulation area first (otherwise the reflection code isn't even run), which requires it to start from EHOLE. I have no idea why this is so. Reproduce the out of bounds read in photoelectric_effect with
break Simulation.cpp:2934 if nx < 0
in gdb and executing the following Lua code:
sim.clearSim()
tpt.set_wallmap(55, 44, 12)
local i = sim.partCreate(-1, 223, 178, 31)
sim.partProperty(i, "vx", -300)
sim.partProperty(i, "vy", 0)
sim.framerender(1)
This is how we'll handle systems where the cert bundle and cert directory is stored where mbedtls doesn't expect it.
Also update tpt-libs to get new curl and mbedtls.
Also Factor out app constants that mods might change into Meson options and clean up format::URLEncode in the process, convert app and document icon data in arrays to actual images, actualize AppStream data for possible future packaging, add alternative command line format for opening filesystem saves and ptsave URLs, fix a memory leak in Platform::GetCwd, and add format::URLDecode.
This made it possible to get rid of two GameSave constructors.
Also clean up Client::LoadSaveFile, Client::ReadFile, and Client::WriteFile in the process, and remove unused SaveRenderer::Render
Also restructure meson.build and the ghactions workflow a bit, and enable -ffunction-sections and -fdata-sections.
Note that starcatcher uploads have not been tested and most likely don't work.
Also fix a few bugs and other weirdness in Platform::DirectorySearch. Empty string paths would crash and filenames with 4 or fewer characters wouldn't register.
This fixes bugs like "type\0hello mom" being a property name sim.partProperty accepts and half-fixes bugs like text formatting codes making gfx.drawText exit prematurely.
Currently it detects presence of v incorrectly (via if (ID(type))) and thus rewrites v = 0 to v = -1, even if 0 is actually what you want. Especially problematic if you're trying to spawn GOL (so LIFE(ctype=0)).
Also fix the first mouse click not being detected on windows with sdl 2.0.20. Apparenlty, we need SDL_HINT_MOUSE_FOCUS_CLICKTHROUGH set to 1 for it to be detected.
If the built-in update function is allowed to run, it can change the particle's type. The code path assumes that there is a Lua update function to call on the particle, but this type change may break this assumption and cause the code to call the update function of an element that doesn't even have one, producing a weird error message with no line number.
Empty stamps are those whose block width or height is 0. While they are technically valid and certain parts of the game are prepared to handle them, others aren't, so it's safest to just adjust the definition of valid stamps to exclude empty ones.
Also clear SaveRenderer graphics cache along with the main Renderer's when needed, and revert to built-in element callbacks rather than nothing at all when assigning nil to a callback slot in Lua.
Rather than hacking velocity, do it directly through can_move. Add a special case to make it slowly float upwards, rather than immediate like most weight differences in TPT.
This is a follow up to this crash fix - 0ed8d0a0be
This may possibly fix other crashes users occasionally experience, especially in the case of high velocity particles with loop edge mode on. Even so, there are other bugs at play, as a crash here can't be triggered if pmap is in a correct state
Very similar in nature to the problem fixed by 0fcad65d. Again, it'd be massive help if we didn't destroy the Lua state explicitly in LSI's dtor. But this is not worth refactoring LSI for.
All these smart pointers have to be cleared before the Lua state is closed. Ordinarily, we'd have a smart pointer to the Lua state defined earlier in LSI than these smart pointers, which would take care of destruction in the correct order, but tfw technical debt.
Ignoring broken pipes led to libcurl crashing when connections were dropped, rather than re-attempting. Now, libcurl is not set to ignore broken pipes. Closes#823.
Repurposing the comment I added because I'm lazy: pavg[1] used to be saved as a u16, which PressureInTmp3 elements then treated as an i16. tmp3 is now saved as a u32, or as a u16 if it's small enough. PressureInTmp3 elements will never use the upper 16 bits, and should still treat the lower 16 bits as an i16, so they need sign extension.
The code that was originally responsible for this somehow got lost in the tmp3/tmp4 migration.
Also add GameSave::PressureInTmp3 to check for elements with pressure memory and fix TUNG not sampling pressure on creation. This does not in itself fix#822 because tmp3 and tmp4 are still saved in 16 bits each, so full ctypes still don't fit in tmp3.
This fixes a crash with water equalization on, with loop mode enabled, when there are high velocity particles near the edges. This can be occasionally reproduced with id:2800901
The water was moved to a new position and pmap updated, but the movement code continued, assuming water was at its old position. pmap for the WATR's old position won't be cleared once it moves, leaving a stale entry. If a particle then looks up the water in that location and tries to swap positions, this can cause a crash at the end of try_move
lua_close wasn't being called at all before due to the delete commandInterface being missing. With it there, the lua gc could delete the LuaComponent before the LuaWindow. Make sure if that happens, it tells the parent LuaWindow it's already been gced
Becuase *of course* whether it succeds or not depends on whether you've rebuilt menus.
With all seriousness, the real culprit here is GameModel::GetToolFromIdentifier, which looks for tools in the menu section tool lists, plus another list with tools that aren't in any menu section. This is absolutely blaphemous, but I don't feel like refactoring this right now. It also wouldn't be a problem in itself, but allocating an element also doesn't rebuild these lists, only changing the MenuSection property and a few other obscure operations do. This makes allocating elements also rebuild these lists.
The underlying problem was that the spreading step in SimulateGOL would record activity concerning a cell to builtinGol even if said cell already housed a non-GOL particle. The culling step handles these records and purges them once it's done (thus builtinGol only ever has non-zero values inside SimulateGOL), except in this case, it saw the non-GOL particle and skipped the cell without purging the corresponding records. This would later let GOL spread seemingly out of nowhere.
LITH: show .tmp2 in HUD, remove outdated comment
SLCN: update description
ROCK: form under pressure from both normal LAVA and LAVA(STNE), as these are basically the same thing
Summary of migrated files will be shown to user in a popup, and a log file with every moved file will be left in the original directory
stamps, saves, scripts, screenshots, and powder.pref will be migrated. Recordings are not.
Username will be fixed automatically if not in the right casing
Logins are always sent over https
Logins are stored and salted using bcrypt on the server-side (wraps around original md5), fixes#294
Count goes negative if an element is killed due to .life=0 on the same frame an elementRecount is scheduled
element count skyrockets when recount is scheduled during game pause
Element count increases by a static amount when undoing, because it schedules a recount without clearing the old counts
Flushing user changes to powder.pref was in fact the reason why I began this pref flushing thing at all, but of course Client handles user changes by modifying preferences directly, not through SetPref, so it skipped WritePrefs.
Allows finer control over whether you want none included or not.
Fixes some invalid things being allowed for element 0 in legacy lua api
Fixes {ctype} signs showing 0 instead of NONE
Improve efficiency in FIRE logic, it no longer runs the check multiple times for each surrounding particle. RNG chances reduced accordingly.
Remove unnecessary code in init_can_move
Set Weight to 100, which is the proper max for weight
These got lost in the migration to Meson; they used to be called
with atexit, but this caused more problems than it should have.
Anyway, it's fine to call these only when we're exiting normally,
since otherwise we have bigger problems than not quitting SDL.
Standardizes the code to run on all mods and betas, not just snapshots and debug builds
Ensures mods and betas can load their own saves
Blocks publishing saves from all non-release versions if they use features not present in the previous release (currently only GoL elements)
Ctrl+A no longer selects all saves if any of the textboxes in
the view are in focus, as a ctrl+A in that case is expected to
select everything in the textbox, not in the save browser.
This change also makes the shortcut deselect all saves if all
saves are selected.
And no, I'm not making events cancellable just for this.
This is very similar to the SPRK-PROT bug (fixed by 4aa58d6c)
in that the cause for a later crash is the free list of particles
being corrupted by messing with the .life property of a dead
particle.
The code in the loop body following this kill_part call has no
effect on other particles if the call were to happen. The only
thing this continue skips is increasing pressure under the now
dead particle.
Credit goes to @QuanTech for finding the problem, see id:2547788
(originally his save, I just saved to my alt account for future
reference).
This fixes the only valid bug on the "Glitches" wiki page. All the rest were already fixed (multi-head stkm) or are not bugs (lava ice, "destroying" dmnd with dray)
Also fix a bug with Requests where any connection that took
longer to finish than 15 seconds would be killed. Should have
used CURLOPT_CONNECTTIMEOUT instead of CURLOPT_TIMEOUT when
specifying the timeout, oops.
Used in create_part, kill_part, and part_change_type, allows us to remove element-specific stuff in those functions
Note: difference between Create and ChangeType is that Create is called when we want to initialize default element properties, but ChangeType is called every time a particle is changed to that type, even if it doesn't need default properties set. ChangeType is mainly used for things that need accurate state tracking, like the stkm spawn status.
ChangeType is called every time a particle is changed to or from its type, which includes if the particle is deleted, and also Lua.
Neither of these functions are called when loading saves, that's probably an oversight, will fix later
Most of the switch statement in create_part is gone. There's a few others that I will get rid of in future commits. There will also be a CreateAllowed function, and a ChangeType. ChangeType will handle stuff that is duplicated in both create_part and part_change_type. Considering making a Destroy function instead of ChangeType, though.
Later on, Lua events will be made for all 3
Credit to jacksonmj for the original design of all of this, I copied it into my mod years ago
The rest will will in a commit soon that adds the Create event and some other stuff
Lua events will come after that
PLSM default temp lowered to MAX_TEMP
Not sure why TPT's ellptical brush currently looks like that,
when you type "pixel circle" on Google every result is the
smooth version.
Got the idea from this post:
https://powdertoy.co.uk/Browse/View.html?ID=2464991
Signed-off-by: Nick Renieris <velocityra@gmail.com>
Currently only affects the smudge tool. I'd have to look
into the others to decide if they need chaging. They probably
do though, they're not exactly intuitive.
I also fixed a bug with DropDowns where their popup would
show up in some random place if the parent window was a
ScrollPanel, and changed a few alignments here and there.
Notably, DropDowns now align the popup so that the currently
selected item is vertically centered and doesn't move when
the popup is opened.
Not a perfect solution as it doesn't remember the modifications
made to previous commands, but it's more common to be concerned
about the command being entered than about the changes made to
the previous commands.
That wrap_if_needed call would sometimes cause empty newlines
to appear under otherwise already well-wrapped blocks of text.
The idea is that whenever a new character is about to be appended,
the current line is checked for being too long with that character
included, and the character is only actually appended after the
wrapper makes sure that appending it won't overflow the line by
inserting a newline. This means that wrap_if_needed is only ever
called before pushing the current character to records, never
after, as was the case here.
This adds an exception from brush-based sparking behaviour for
elements that have a ctypeDraw function. A functionally equivalent
exception has been removed in de1fc0f, thus making it impossible to
ctype-draw SPRK on a few elements that otherwise aren't sparkable,
such as CONV.
The metrics of the rectangles comprising the selection background
were calculated based on the selected range and the unformatted
text, so these rectangles would be off when the formatted text
was different (i.e. had any formatting).
This feature was removed temporarily in a407aba.
Whether pressure is included when saving or loading is
determined as follows:
* load-like operations (which invoke Simulation::Load) always
include pressure;
* save-like operations (which invoke Simulation::Save) include
pressure if the Simulation.IncludePressure preference node
is true;
* finally the state of the shift key inverts the decision.
Though I'm quite sure nobody actually cares about that table. It's been broken for a
very long time.
The PROP tool now properly calls part_change_type when type is being set.
I don't know how it ended up inside, I wanted it outside. For the record,
signalling the CV while the associated mutex is locked is a pessimisation,
it just yields predictable scheduling.
I didn't use 'fixes' because this doesn't strictly fix that specific
bug (I think?). Anyway, it does remove strdup which is mentioned in the
issue.
Also fix elem.free not allowing DEFAULTFOO_PT_STUFF and similar to
be freed.
