This commit removes a large amount of code partially duplicated
between the text and the graphics windows, and opens the path to
having more than one model window on screen at any given time,
as well as simplifies platform work.
This commit also adds complete support for High-DPI device pixel
ratio. It adds support for font scale factor (a fractional factor
on top of integral device pixel ratio) on the platform side, but not
on the application side.
This commit also adds error checking to all Windows API calls
(within the abstracted code) and fixes a significant number of
misuses and non-future-proof uses of Windows API.
This commit also makes uses of Windows API idiomatic, e.g. using
the built-in vertical scroll bar, native tooltips, control
subclassing instead of hooks in the global dispatch loop, and so on.
It reinstates tooltip support and removes menu-related hacks.
Extrustion top and bottom faces require a normal to be present.
Before this commit, the normal is always taken from the assembled
loop; if the loop could not be assembled (i.e. the loop is broken
or not coplanar) the normal will be (0,0,0), which breaks the sketch.
Also, loops are not generated when generating the sketch
to determine its bounding box.
This may result in spuriously broken sketches when e.g. undoing
a change that has broken a loop.
After this commit, loops are generated when generating for bounding
box, and if the loop could not be assembled, then the workplane
normal is used. This still results in failures when there is
no workplane, but those cases should be quite pathological.
This is a fairly standard CAD feature; it conveys the same
information and has the same recovery path, without erroring out,
so seems like an obvious win.
Before this commit, it was possible to add some redundant constraints
(e.g. vertical, horizontal or midpoint) without failing the sketch,
because SolveBySubstitution() removed the redundant equations.
However, this could result in the solve failing later because
the system didn't converge, without any pointers as to the true
cause of the failure.
This partially reverts commit 3a585ea.
We no longer need this because the VectorsParallel() is gone, and
there is no chance of pivoting wrong when solving.
Before this commit, pt-on-line constraints are buggy. To reproduce,
extrude a circle, then add a datum point and constrain it to the
axis of the circle, then move it. The cylinder will collapse.
To quote Jonathan:
> On investigation, I (a) confirm that the problem is
> the unconstrained extrusion depth going to zero, and (b) retract
> my earlier statement blaming extrude and other similar non-entity
> parameter treatment for this problem; you can easily reproduce it
> with a point in 3d constrained to lie on any line whose length
> is free.
>
> PT_ON_LINE is written using VectorsParallel, for no obvious reason.
> Rewriting that constraint to work on two projected distances (using
> any two basis vectors perpendicular to the line) should fix that
> problem, since replacing the "point on line in 3d" constraint with
> two "point on line in 2d" constraints works. That still has
> the hairy ball problem of choosing the basis vectors, which you
> can't do with a continuous function; you'd need Vector::Normal()
> or equivalent.
>
> You could write three equations and make the constraint itself
> introduce one new parameter for t. I don't know how well that
> would work numerically, but it would avoid the hairy ball problem,
> perhaps elegant at the cost of speed.
Indeed, this commit implements the latter solution: it introduces
an additional free parameter. The point being coincident with
the start of the line corresponds to the parameter being zero, and
point being coincident with the end corresponds to one).
In effect, instead of constraining two of three degrees of freedom
(for which the equations do not exist because of the hairy ball
theorem), it constrains three and adds one more.
To actually achieve improved performance with the OpenGL 2 renderer,
we have to cache geometry that doesn't change when the viewport does
(note that the rendered pixels can change quite dramatically because
we can reconfigure shaders; e.g. stippling can be drawn in screen
coordinates).
This commit adds a BatchCanvas interface that can be implemented
by renderers, and uses it for drawing entities such as lines and
points.
gcc 6 displays these when compiling in release mode; all of these
warnings except the rankOk one were benign because there would have
been an error about the incomplete switch statement.
The rankOk warning highlighted a real problem: bailing early to
didnt_converge would have branched on an uninitialized variable.
GroupsInOrder is an extremely hot function, especially during object
picking: before this commit, it was easy to get second plus latencies
on picking, and after this commit picking is almost real-time.
Sometimes, after a large change in a sketch, constraints that are
geometrically fine may still cause the rank test to fail. One way
this can happen is VectorsParallel() pivoting wrong due to the big
move, converging anyways but ending up singular. It would then
re-pivot correctly on the new solution when you re-solve, making
this a transient error. This is visible when dragging the arm in
the jansen-asm.slvs example.
After this commit, if the rank test fails, equations are regenerated
the Jacobian is rewritten, and the rank test is retried, which
prevents these transient errors from interfering with dragging.
The problem described above was invisible before c011444, as rank
test was only performed before solving.
It's broken. It expects a valid OpenGL context during generation
that immediately pushes changes to the screen. This is never true
on non-Windows as offscreen rendering is used, and also incompatible
with OpenGL core profile.
Further, right now it displays junk on Windows as well due to some
issue with the bitmap font texture loading.
We will restore it later, in a saner form.
This is to ensure that:
* it is clear, when looking at the point of usage, what is
the purpose of "true" or "false";
* when refactoring, a simple search will bring up any places that
need to be changed.
Also, argument names were synchronized between declaration and
implementation.
As an exception, these are not annotated:
* Printf(/*halfLine=*/), to avoid pointless churn.
Specifically, this enables -Wswitch=error on GCC/Clang and its MSVC
equivalent; the exact way it is handled varies slightly, but what
they all have in common is that in a switch statement over an
enumeration, any enumerand that is not explicitly (via case:) or
implicitly (via default:) handled in the switch triggers an error.
Moreover, we also change the switch statements in three ways:
* Switch statements that ought to be extended every time a new
enumerand is added (e.g. Entity::DrawOrGetDistance(), are changed
to explicitly list every single enumerand, and not have a
default: branch.
Note that the assertions are kept because it is legal for
a enumeration to have a value unlike any of its defined
enumerands, and we can e.g. read garbage from a file, or
an uninitialized variable. This requires some rearranging if
a default: branch is undesired.
* Switch statements that ought to only ever see a few select
enumerands, are changed to always assert in the default: branch.
* Switch statements that do something meaningful for a few
enumerands, and ignore everything else, are changed to do nothing
in a default: branch, under the assumption that changing them
every time an enumerand is added or removed would just result
in noise and catch no bugs.
This commit also removes the {Request,Entity,Constraint}::UNKNOWN and
Entity::DATUM_POINT enumerands, as those were just fancy names for
zeroes. They mess up switch exhaustiveness checks and most of the time
were not the best way to implement what they did anyway.
Specifically, take the old code that looks like this:
class Foo {
enum { X = 1, Y = 2 };
int kind;
}
... foo.kind = Foo::X; ...
and convert it to this:
class Foo {
enum class Kind : uint32_t { X = 1, Y = 2 };
Kind kind;
}
... foo.kind = Foo::Kind::X;
(In some cases the enumeration would not be in the class namespace,
such as when it is generally useful.)
The benefits are as follows:
* The type of the field gives a clear indication of intent, both
to humans and tools (such as binding generators).
* The compiler is able to automatically warn when a switch is not
exhaustive; but this is currently suppressed by the
default: ssassert(false, ...)
idiom.
* Integers and plain enums are weakly type checked: they implicitly
convert into each other. This can hide bugs where type conversion
is performed but not intended. Enum classes are strongly type
checked.
* Plain enums pollute parent namespaces; enum classes do not.
Almost every defined enum we have already has a kind of ad-hoc
namespacing via `NAMESPACE_`, which is now explicit.
* Plain enums do not have a well-defined ABI size, which is
important for bindings. Enum classes can have it, if specified.
We specify the base type for all enums as uint32_t, which is
a safe choice and allows us to not change the numeric values
of any variants.
This commit introduces absolutely no functional change to the code,
just renaming and change of types. It handles almost all cases,
except GraphicsWindow::pending.operation, which needs minor
functional change.
This includes explanation and context for non-obvious cases and
shortens debug cycles when just-in-time debugging is not available
(like on Linux) by immediately printing description of the assert
as well as symbolized backtrace.
Most of these were just converting char* into std::string back and
forth; some more used ReadUTF8, which was converted to use nicer
STL-style iterators over UTF-8 text.
The remaining ones are:
* arguments to Expr::From, which we'll change when refactoring
the expression lexer;
* arguments to varargs functions, which we'll change when adding
localization (that requires custom printf-style functions to
allow for changing argument order);
* arguments where only string literals are ever passed, which
are OK;
* in platform-specific code, which is OK.
After commit 2f734d9, inactive groups are no longer regenerated
for trivial changes, e.g. changing parameters, so it's possible to
switch to an earlier group and work on it without incurring
the computational (slowdown) and cognitive (annoyance by red
background) overhead of later groups failing to solve.
However, if a group--any group anywhere--was not solved OK,
the interface reacted accordingly, which diminished usefulness of
the change, especially given that, if we have groups A and B with
B depending on A, if B is broken by a change in A and we activate A
and fix it, B will not be regenerated.
After this commit, only active groups are considered when deciding
if generating the entire sketch would fail.
Commit 89eb208 has improved the overall situation with chord
tolerance, but it changed the display chord tolerance to use
an absolute value in millimeters as a stopgap measure.
This commit changes the display chord tolerance to be specified
in percents of entity bounding box instead of millimeters.
As a result, the linearized curves are both zoom level and sketch
scale independent.
In order to compute the bounding box, all entities are generated
twice. However, this shouldn't result in a noticeable slowdown,
since the bounding box calculation does not need the expensive
triangle mesh generation and the solver will converge immediately
on the second run.
Since the meaning of the preference has changed, a new name is
used (ChordTolerancePct instead of ChordTolerance), so that it
would be reset to the default value after updating SolveSpace.
The default value, 0.5%, was selected using trial and error by
judging whether cylinders of moderate dimensions were looking
aesthetically pleasing enough.
After this change, the only real function of the spacebar
shortcut is to reload imported groups, since manual regeneration
should not change anything anymore unless there is a bug.
Before this change, groups and their meshes were generated even past
the active group, which, in cause the mesh was broken, caused red
marks to appear for no apparent reason. Furthermore, it unnecessarily
slows down regeneration.
The commit 11f29b123 has replaced most of the uses of sprintf,
but there were still many remaining in Screen* functions, and it
was annoyingly inconsistent. Moreover, while most usage of sprintf
there was fine, it is bad hygiene to leave stack overflow prone
code around.
This setting is generally useful, but it especially shines when
assembling, since the "same orientation" and "parallel" constraints
remove three and two rotational degrees of freedom, which makes them
impossible to use with 3d "point on line" constraint that removes
two spatial and two rotational degrees of freedom.
The setting is not enabled for all imported groups by default
because it exhibits some edge case failures. For example:
* draw two line segments sharing a point,
* constrain lengths of line segments,
* constrain line segments perpendicular,
* constrain line segments to a 90° angle.
This is a truly degenerate case and so it is not considered very
important. However, we can fix this later by using Eigen::SparseQR.
Now it is possible to give non-ASCII names to groups
as well as see non-ASCII filenames of imported files.
In the future this makes localization possible.
This works for LTR languages, such as European and CJK,
but not RTL such as Arabic. Does Arabic even exist in
monospaced form? I have no idea.
This will allow us to use non-POD classes inside these objects
in future and is otherwise functionally equivalent, as well
as more concise.
Note that there are some subtleties with handling of
brace-initialization. Specifically:
On aggregates (e.g. simple C-style structures) using an empty
brace-initializer zero-initializes the aggregate, i.e. it makes
all members zero.
On non-aggregates an empty brace-initializer calls the default
constructor. And if the constructor doesn't explicitly initialize
the members (which the auto-generated constructor doesn't) then
the members will be constructed but otherwise uninitialized.
So, what is an aggregate class? To quote the C++ standard
(C++03 8.5.1 §1):
An aggregate is an array or a class (clause 9) with no
user-declared constructors (12.1), no private or protected
non-static data members (clause 11), no base classes (clause 10),
and no virtual functions (10.3).
In SolveSpace, we only have to handle the case of base classes;
Constraint and Entity have those. Thus, they had to gain a default
constructor that does nothing but initializes the members to zero.
This is required to avoid name conflicts with the Cocoa libraries
on OS X.
I renamed the `class SolveSpace` to `class SolveSpaceUI`, because
that's what it does, and because otherwise the namespace would
have to be called something else than `namespace SolveSpace`.
In principle, GTK3 is the way forward, and GTK2 is officially
deprecated, though still maintained. In practice however, GTK3
is often unbearably buggy; e.g. on my system, combo boxes
don't ever roll up in GTK3 windows. So I have added support
for both.
This required a few minor changes to the core, namely:
* GTK wants to know beforehand whether a menu item is a check
menu item or a regular one.
* GTK doesn't give us an easy way to execute something after
any event is processed, so an explicit idle timer is added.
This is a no-op on Win32.
* A few function signatures were const'ed, since GTK expects
immutable strings when converting to Glib::ustring.
However, don't use ssglLineWidth for UI drawing operations.
These only draw horizontal or vertical lines that don't need to
be antialiased, and thus don't require the workaround. In fact
the workaround would make them thicker than needed.
The SolveSpace top-level directory was getting a bit cluttered, so
following the example of numerous other free-software projects, we move the
main application source into a subdirectory and adjust the build systems
accordingly.
Also, got rid of the obj/ directory in favor of creating it on the fly in
Makefile.msvc.
Ryan Pavlik
whitequark
EvilSpirit
Daniel Richard G