This changes ParamNode to use UTF8 values internally (XMB files are UTF8
since cb9d0733ef).
This removes the need for a lot of conversions, speeding things up and
allows cleaning up the validator interface & a few other callsites.
ConstructJSVal could be a tad slower because of UTF8->16 conversions
within Spidermonkey; but the difference is unlikely to be noticeable in
practica.
Also:
- Changes `ToXML` to `ToXMLString` for clarity.
- Add a simple "op" test & show a particular behaviour of merge nodes
that I intend to change somewhat in D3830.
- Remove Component.h from simulation2 PCH - brought in too much.
Tested by: langbart
Differential Revision: https://code.wildfiregames.com/D3834
This was SVN commit r25228.
NaN values could not be serialised safely because of the multiple
possible NaN numbers.
Since NaN values are usually the result of bugs or dangerous code, it
seems simpler to refuse to serialise them.
(D3205 was a safe-serialization alternative, should the need arise).
Fixes#1879
Differential Revision: https://code.wildfiregames.com/D3729
This was SVN commit r25151.
This is the same commit as 4bb31f084e, reverted in 5d1899785a. Noted
issues were fixed.
Partial revert/fix of 2f19cf86d3 and 2567fee329.
Before this diff, it was possible for a mod containing a cache/ or a
config/ folder to get written to incorrectly.
The issue is VFS can map multiple directories to one 'virtual' path, for
reading mods. However, writing data is problematic: which path to
choose?
The only viable solution is to use a path relative to the highest
priority directory encountered in the VFS path, or write paths could be
'hijacked' by lower-priority mods.
This fixes these issues by:
- Adding a new lookup mode ('Real-path') that explicitly picks the real
path relative to the highest-priority subdirectory in the VFS Path.
- Preventing overwriting a real directory with a lower priority one in
general.
- Revert c0c8132dd4's GetRealPath change, re-introducing the function as
GetOriginalPath.
This also cleans up some duplication that led to empty mod folders in
the user mod path, and cleans up loading the 'user' mod.
It also makes it explicit that a directory must be passed to Mount().
Note that the new 'realpath' lookup can still be somewhat complex with
many mount points at various hierarchy levels, but it is at least
predictable/deterministic without having to be careful about populating
order.
Fixes#2553
Differential Revision: https://code.wildfiregames.com/D3728
This was SVN commit r25107.
Partial revert/fix of 2f19cf86d3 and 2567fee329.
Before this diff, it was possible for a mod containing a cache/ or a
config/ folder to get written to incorrectly.
The issue is VFS can map multiple directories to one 'virtual' path, for
reading mods. However, writing data is problematic: which path to
choose?
The only viable solution is to use a path relative to the highest
priority directory encountered in the VFS path, or write paths could be
'hijacked' by lower-priority mods.
This fixes these issues by:
- Adding a new lookup mode ('Real-path') that explicitly picks the real
path relative to the highest-priority subdirectory in the VFS Path.
- Preventing overwriting a real directory with a lower priority one in
general.
- Revert c0c8132dd4's GetRealPath change, re-introducing the function as
GetOriginalPath.
This also cleans up some duplication that led to empty mod folders in
the user mod path, and cleans up loading the 'user' mod.
Note that the new 'realpath' lookup can still be somewhat complex with
many mount points at various hierarchy levels, but it is at least
predictable/deterministic without having to be careful about populating
order.
Fixes#2553
Differential Revision: https://code.wildfiregames.com/D3217
This was SVN commit r25104.
Following 9fc6c3c897, the OOS debugSerializer would try to serialize AI
objects entirely. This is extremely slow (might be an infinite loop in
some cases).
Instead, it should print the result of Serialize() in those cases.
(The original idea was that it would print more debugging information,
but in practice it seems to mostly print things that show up when
diffing but aren't actually sources of OOS, so essentially garbage).
Fixes#5917
Differential Revision: https://code.wildfiregames.com/D3361
This was SVN commit r24641.
This reverts dad2857538. That diff had two issues:
- It modifies the JS objects, which means subsequent serialization in
quicksave are 'dirty'.
- It doesn't work with non-extensible objects. That's rather annoying,
and has already caused problems.
It also revert f0faab7a42, which was necessary because of the second
issue.
Fixes#5908
Differential Revision: https://code.wildfiregames.com/D3336
This was SVN commit r24563.
Generalize component/AI serialization system to any user-defined JS
object. This includes Vector2D/3D, fixing an old issue.
As with components/AI, JS Objects may implement a Serialize/Deserialize
function to store custom data instead of the default, which attemps to
serialize all enumerable properties.
Fixes#4698
Differential Revision: https://code.wildfiregames.com/D2746
This was SVN commit r24462.
By using templates appripriately we can remove the need for explicit
specification of serializers, making it easier to serialize container
types and to write new serialization helpers.
Direct serialization calls haven't been replaced in this diff.
Comments by: vladislavbelov
Differential Revision: https://code.wildfiregames.com/D3207
This was SVN commit r24427.
C++14 deprecates and C++17 removes functions we use for random sounds.
This makes FastRand() fit the URG interface, and uses the appropriate
function.
Refs #5862
Differential Revision: https://code.wildfiregames.com/D3107
This was SVN commit r24302.
No particularly noteworthy changes, as most complex API changes were
already supported in SM45 and done.
The addition of JSStructuredCloneData allows to remove our custom class.
Changes:
- InformalValueTypeName is back in the API, so remove our
implementation.
- Stop using JSRuntime entirely in favour of JSContext*
- JSPropertyDescriptor is renamed.
- CompartmentOptions are tweaked slightly (no functional changes)
- JS::Construct - API update.
- JSClass split - API update.
- A js.msg error message was removed, so we had to use a different one.
- Tests fix: fix comparison of union instances
- Disable warning in spidermonkey Vector.h
- Update error reporting to SM52 (minor API updates)
- Ignore warnings about unused return values (would come from OOM, which
isn't recoverable)
Most of the patching was done by Itms.
Tested by: Stan, Freagarach
Fixes#4893
Differential Revision: https://code.wildfiregames.com/D3095
This was SVN commit r24203.
- Check for pending exceptions after function calls and script
executions.
- Call LOGERROR instead of JS_ReportError when there is a conversion
error in FromJSVal, since that can only be called from C++ (where JS
errors don't really make sense). Instead, C++ callers of FromJSVal
should handle the failure and, themselves, either report an error or
simply do something else.
- Wrap JS_ReportError since that makes updating it later easier.
This isn't a systematical fix since ToJSVal also ought return a boolean
for failures, and we probably should trigger errors instead of warnings
on 'implicit' conversions, rather a preparation diff.
Part of the SM52 migration, stage: SM45 compatible (actually SM52
incompatible, too).
Based on a patch by: Itms
Comments by: Vladislavbelov, Stan`
Refs #742, #4893
Differential Revision: https://code.wildfiregames.com/D3093
This was SVN commit r24187.
SM52 essentially replaces JSRuntime with JSContext (though JSContext
itself was replaced with JSCompartment).
To prepare for this migration, rename all Runtime-related things to
Context.
Part of the SM52 migration, stage: SM45 compatible.
Patch by: Itms
Refs #4893
Differential Revision: https://code.wildfiregames.com/D3091
This was SVN commit r24181.
JSAutoRequest is required before calling into most JSAPI methods, for GC
reasons.
Calling it is required and fragile as one must not forget.
Further, SM52 and later make manipulating JSContext* dangerous as that
can cross Compartment(Realm in SM68) barriers (and ScriptInterface now
matches a Compartment).
The solution to both problems is to avoid using JSContext* in 0 A.D.
itself. To achieve this, a Request class is introduced, and must be used
to access a JSContext* from a scriptInterface. Further, Request is
passed to other ScriptInterface functions isntead of JSContext*, making
it obvious that the caller has already called it, reducing errors and
redundant JSAutoRequest calls.
Only JSNative functions now get a naked JSContext* without protection,
but the likelihood of forgetting a request is lower since many
ScriptInterface functions now expect it.
JSContext* is directly passed to JSAPI functions only.
Part of the SM52 migration, stage: SM45 compatible
Based on a patch by: Itms
Tested By: Freagarach
Refs #4893
Differential Revision: https://code.wildfiregames.com/D3088
This was SVN commit r24176.
When serialising JS objects, we keep track of any encountered object,
and serialize it only once. Any further serialisation instead stores an
ID referring to the original object (essentially an opaque pointer).
The trouble of course is to have a unique, persistent identifier for
such an object.
svn uses an ObjectIDCache, essentially a "JS Object -> ID" map (which
internally is essentially a "JS heap pointer -> ID" map).
JS, since ES15, includes a "Symbol" primitive type, which is a unique,
immutable identifier. They are also not iterable by for..in or
GetOwnPropertyName or related.
This means they can be used to store the tag directly on the object
(since it's impossible overwrite a user property).
Thanks to this, we can forgo ObjectIDCache in the serializers, and since
following D2897 it becomes unused, we can delete it, along with the
Finalization code it used.
Part of SM52 migration, stage: SM45-compatible changes.
Patch by: Itms
Tested By: Freagarach
Refs #4893
Differential Revision: https://code.wildfiregames.com/D3085
This was SVN commit r24167.
Range manager has several `std::vector` for fixed-size arrays and 2D
grids. By using proper data structures, the code readability is
improved.
This also moves around the LosVisibility enum.
Comments by: Stan`, nani
Differential Revision: https://code.wildfiregames.com/D2770
This was SVN commit r23769.
This generally makes this code cleaner and easier to maintain, while
also being more modern and fully tested.
Looked at by: leper
Differential Revision: https://code.wildfiregames.com/D8
This was SVN commit r22029.
This fixes an assertion failure in ScriptEngine that can occur when
closing
the game while in the loading screen.
Reviewed By: vladislavbelov, leper
Differential Revision: https://code.wildfiregames.com/D684
This was SVN commit r20035.
Templates should be the same for all players, techs can be used to
provide
different lists of templates if that is needed.
Reviewed By: Itms
Differential Revision: https://code.wildfiregames.com/D25
This was SVN commit r19142.
Summary: Refs #4419.
Test Plan: When running tests, make sure nothing is created.
Reviewers: leper
Reviewed By: leper
Subscribers: leper, Vulcan
Differential Revision: https://code.wildfiregames.com/D1
This was SVN commit r19062.
This avoids vswprintf failures when printing non-ASCII char* strings
from CLogger into the console.
Also convert ScriptInterface::ToString to return UTF-8, to avoid some
utf8_from_wstring calls.
Also remove some unused and redundant CConsole functions.
This was SVN commit r16333.
This fixes the problem where passing a non-ASCII string to
debug_printf(L"%hs", s) caused vswprintf_s to fail on Linux (because it
doesn't know what encoding the char* is meant to have). Now debug
messages will remain as UTF-8 until they reach the OS.
Fixes#3021.
This was SVN commit r16332.
This upgrade also introduces exact stack rooting (see to the wiki:
JSRootingGuide) and fixes problems with moving GC. This allows us to
enable generational garbage collection (GGC).
Measurements a few months ago have shown a performance improvement of a
non-visual replay of around 13.5%. This probably varies quite a bit, but
it should be somewhere between 5-20%. Memory usage has also been
improved. Check the forum thread for details.
Thanks to everyone from the team who helped with this directly or
indirectly (review, finding and fixing issues, the required C++11
upgrade, the new autobuilder etc.)! Also thanks to the SpiderMonkey
developers who helped on the #jsapi channel or elsewhere!
Fixes#2462, #2415, #2428, #2684, #1374
Refs #2973, #2669
This was SVN commit r16214.
sys_vswprintf relies on platform-specific printf implementations, which
vary widely between platforms (in handling of truncation, return values,
use of %s/%S/%hs/%ls for mixing char and wchar_t strings, etc) and are
therefore a pain.
Use cppformat's fmt::sprintf instead, which has very similar syntax to
sprintf but is more C++ish and is portable.
Also, wchar_t is stupid, so use char* strings (which are expected to be
UTF-8) in CLogger. This creates a bit of a pain with changing all
callers to convert to char* strings, but that's their fault for not
using UTF-8 already.
Refs #3011.
This was SVN commit r16182.
