solvespace/src/graphicswin.cpp

1311 lines
53 KiB
C++
Raw Normal View History

//-----------------------------------------------------------------------------
// Top-level implementation of the program's main window, in which a graphical
// representation of the model is drawn and edited by the user.
//
// Copyright 2008-2013 Jonathan Westhues.
//-----------------------------------------------------------------------------
#include "solvespace.h"
typedef void MenuHandler(Command id);
using MenuKind = Platform::MenuItem::Indicator;
struct MenuEntry {
int level; // 0 == on menu bar, 1 == one level down
const char *label; // or NULL for a separator
Command cmd; // command ID
int accel; // keyboard accelerator
MenuKind kind;
MenuHandler *fn;
};
#define mView (&GraphicsWindow::MenuView)
#define mEdit (&GraphicsWindow::MenuEdit)
#define mClip (&GraphicsWindow::MenuClipboard)
#define mReq (&GraphicsWindow::MenuRequest)
#define mCon (&Constraint::MenuConstrain)
#define mFile (&SolveSpaceUI::MenuFile)
#define mGrp (&Group::MenuGroup)
#define mAna (&SolveSpaceUI::MenuAnalyze)
#define mHelp (&SolveSpaceUI::MenuHelp)
#define SHIFT_MASK 0x100
#define CTRL_MASK 0x200
#define FN_MASK 0x400
#define S SHIFT_MASK
#define C CTRL_MASK
#define F FN_MASK
#define KN MenuKind::NONE
#define KC MenuKind::CHECK_MARK
#define KR MenuKind::RADIO_MARK
const MenuEntry Menu[] = {
//lv label cmd accel kind
{ 0, N_("&File"), Command::NONE, 0, KN, NULL },
{ 1, N_("&New"), Command::NEW, C|'n', KN, mFile },
{ 1, N_("&Open..."), Command::OPEN, C|'o', KN, mFile },
{ 1, N_("Open &Recent"), Command::OPEN_RECENT, 0, KN, mFile },
{ 1, N_("&Save"), Command::SAVE, C|'s', KN, mFile },
{ 1, N_("Save &As..."), Command::SAVE_AS, 0, KN, mFile },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Export &Image..."), Command::EXPORT_IMAGE, 0, KN, mFile },
{ 1, N_("Export 2d &View..."), Command::EXPORT_VIEW, 0, KN, mFile },
{ 1, N_("Export 2d &Section..."), Command::EXPORT_SECTION, 0, KN, mFile },
{ 1, N_("Export 3d &Wireframe..."), Command::EXPORT_WIREFRAME, 0, KN, mFile },
{ 1, N_("Export Triangle &Mesh..."), Command::EXPORT_MESH, 0, KN, mFile },
{ 1, N_("Export &Surfaces..."), Command::EXPORT_SURFACES, 0, KN, mFile },
{ 1, N_("Im&port..."), Command::IMPORT, 0, KN, mFile },
2015-03-24 06:45:53 +00:00
#ifndef __APPLE__
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("E&xit"), Command::EXIT, C|'q', KN, mFile },
2015-03-24 06:45:53 +00:00
#endif
{ 0, N_("&Edit"), Command::NONE, 0, KN, NULL },
{ 1, N_("&Undo"), Command::UNDO, C|'z', KN, mEdit },
{ 1, N_("&Redo"), Command::REDO, C|'y', KN, mEdit },
{ 1, N_("Re&generate All"), Command::REGEN_ALL, ' ', KN, mEdit },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Snap Selection to &Grid"), Command::SNAP_TO_GRID, '.', KN, mEdit },
{ 1, N_("Rotate Imported &90°"), Command::ROTATE_90, '9', KN, mEdit },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Cu&t"), Command::CUT, C|'x', KN, mClip },
{ 1, N_("&Copy"), Command::COPY, C|'c', KN, mClip },
{ 1, N_("&Paste"), Command::PASTE, C|'v', KN, mClip },
{ 1, N_("Paste &Transformed..."), Command::PASTE_TRANSFORM, C|'t', KN, mClip },
{ 1, N_("&Delete"), Command::DELETE, '\x7f', KN, mClip },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Select &Edge Chain"), Command::SELECT_CHAIN, C|'e', KN, mEdit },
{ 1, N_("Select &All"), Command::SELECT_ALL, C|'a', KN, mEdit },
{ 1, N_("&Unselect All"), Command::UNSELECT_ALL, '\x1b', KN, mEdit },
{ 0, N_("&View"), Command::NONE, 0, KN, mView },
{ 1, N_("Zoom &In"), Command::ZOOM_IN, '+', KN, mView },
{ 1, N_("Zoom &Out"), Command::ZOOM_OUT, '-', KN, mView },
{ 1, N_("Zoom To &Fit"), Command::ZOOM_TO_FIT, 'f', KN, mView },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Align View to &Workplane"), Command::ONTO_WORKPLANE, 'w', KN, mView },
{ 1, N_("Nearest &Ortho View"), Command::NEAREST_ORTHO, F|2, KN, mView },
{ 1, N_("Nearest &Isometric View"), Command::NEAREST_ISO, F|3, KN, mView },
{ 1, N_("&Center View At Point"), Command::CENTER_VIEW, F|4, KN, mView },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Show Snap &Grid"), Command::SHOW_GRID, '>', KC, mView },
{ 1, N_("Use &Perspective Projection"), Command::PERSPECTIVE_PROJ, '`', KC, mView },
{ 1, N_("Dimension &Units"), Command::NONE, 0, KN, NULL },
{ 2, N_("Dimensions in &Millimeters"), Command::UNITS_MM, 0, KR, mView },
{ 2, N_("Dimensions in M&eters"), Command::UNITS_METERS, 0, KR, mView },
{ 2, N_("Dimensions in &Inches"), Command::UNITS_INCHES, 0, KR, mView },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Show &Toolbar"), Command::SHOW_TOOLBAR, 0, KC, mView },
{ 1, N_("Show Property Bro&wser"), Command::SHOW_TEXT_WND, '\t', KC, mView },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("&Full Screen"), Command::FULL_SCREEN, C|F|11, KC, mView },
{ 0, N_("&New Group"), Command::NONE, 0, KN, mGrp },
{ 1, N_("Sketch In &3d"), Command::GROUP_3D, S|'3', KN, mGrp },
{ 1, N_("Sketch In New &Workplane"), Command::GROUP_WRKPL, S|'w', KN, mGrp },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Step &Translating"), Command::GROUP_TRANS, S|'t', KN, mGrp },
{ 1, N_("Step &Rotating"), Command::GROUP_ROT, S|'r', KN, mGrp },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("E&xtrude"), Command::GROUP_EXTRUDE, S|'x', KN, mGrp },
{ 1, N_("&Lathe"), Command::GROUP_LATHE, S|'l', KN, mGrp },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Link / Assemble..."), Command::GROUP_LINK, S|'i', KN, mGrp },
{ 1, N_("Link Recent"), Command::GROUP_RECENT, 0, KN, mGrp },
{ 0, N_("&Sketch"), Command::NONE, 0, KN, mReq },
{ 1, N_("In &Workplane"), Command::SEL_WORKPLANE, '2', KR, mReq },
{ 1, N_("Anywhere In &3d"), Command::FREE_IN_3D, '3', KR, mReq },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Datum &Point"), Command::DATUM_POINT, 'p', KN, mReq },
{ 1, N_("&Workplane"), Command::WORKPLANE, 0, KN, mReq },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Line &Segment"), Command::LINE_SEGMENT, 's', KN, mReq },
{ 1, N_("C&onstruction Line Segment"), Command::CONSTR_SEGMENT, S|'s', KN, mReq },
{ 1, N_("&Rectangle"), Command::RECTANGLE, 'r', KN, mReq },
{ 1, N_("&Circle"), Command::CIRCLE, 'c', KN, mReq },
{ 1, N_("&Arc of a Circle"), Command::ARC, 'a', KN, mReq },
{ 1, N_("&Bezier Cubic Spline"), Command::CUBIC, 'b', KN, mReq },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("&Text in TrueType Font"), Command::TTF_TEXT, 't', KN, mReq },
{ 1, N_("&Image"), Command::IMAGE, 0, KN, mReq },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("To&ggle Construction"), Command::CONSTRUCTION, 'g', KN, mReq },
{ 1, N_("Tangent &Arc at Point"), Command::TANGENT_ARC, S|'a', KN, mReq },
{ 1, N_("Split Curves at &Intersection"), Command::SPLIT_CURVES, 'i', KN, mReq },
{ 0, N_("&Constrain"), Command::NONE, 0, KN, mCon },
{ 1, N_("&Distance / Diameter"), Command::DISTANCE_DIA, 'd', KN, mCon },
{ 1, N_("Re&ference Dimension"), Command::REF_DISTANCE, S|'d', KN, mCon },
{ 1, N_("A&ngle"), Command::ANGLE, 'n', KN, mCon },
{ 1, N_("Reference An&gle"), Command::REF_ANGLE, S|'n', KN, mCon },
{ 1, N_("Other S&upplementary Angle"), Command::OTHER_ANGLE, 'u', KN, mCon },
{ 1, N_("Toggle R&eference Dim"), Command::REFERENCE, 'e', KN, mCon },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("&Horizontal"), Command::HORIZONTAL, 'h', KN, mCon },
{ 1, N_("&Vertical"), Command::VERTICAL, 'v', KN, mCon },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("&On Point / Curve / Plane"), Command::ON_ENTITY, 'o', KN, mCon },
{ 1, N_("E&qual Length / Radius / Angle"), Command::EQUAL, 'q', KN, mCon },
{ 1, N_("Length Ra&tio"), Command::RATIO, 'z', KN, mCon },
{ 1, N_("Length Diff&erence"), Command::DIFFERENCE, 'j', KN, mCon },
{ 1, N_("At &Midpoint"), Command::AT_MIDPOINT, 'm', KN, mCon },
{ 1, N_("S&ymmetric"), Command::SYMMETRIC, 'y', KN, mCon },
{ 1, N_("Para&llel / Tangent"), Command::PARALLEL, 'l', KN, mCon },
{ 1, N_("&Perpendicular"), Command::PERPENDICULAR, '[', KN, mCon },
{ 1, N_("Same Orient&ation"), Command::ORIENTED_SAME, 'x', KN, mCon },
{ 1, N_("Lock Point Where &Dragged"), Command::WHERE_DRAGGED, ']', KN, mCon },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Comment"), Command::COMMENT, ';', KN, mCon },
{ 0, N_("&Analyze"), Command::NONE, 0, KN, mAna },
{ 1, N_("Measure &Volume"), Command::VOLUME, C|S|'v', KN, mAna },
{ 1, N_("Measure A&rea"), Command::AREA, C|S|'a', KN, mAna },
{ 1, N_("Measure &Perimeter"), Command::PERIMETER, C|S|'p', KN, mAna },
{ 1, N_("Show &Interfering Parts"), Command::INTERFERENCE, C|S|'i', KN, mAna },
{ 1, N_("Show &Naked Edges"), Command::NAKED_EDGES, C|S|'n', KN, mAna },
{ 1, N_("Show &Center of Mass"), Command::CENTER_OF_MASS, C|S|'c', KN, mAna },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("Show Degrees of &Freedom"), Command::SHOW_DOF, C|S|'f', KN, mAna },
{ 1, NULL, Command::NONE, 0, KN, NULL },
{ 1, N_("&Trace Point"), Command::TRACE_PT, C|S|'t', KN, mAna },
{ 1, N_("&Stop Tracing..."), Command::STOP_TRACING, C|S|'s', KN, mAna },
{ 1, N_("Step &Dimension..."), Command::STEP_DIM, C|S|'d', KN, mAna },
{ 0, N_("&Help"), Command::NONE, 0, KN, mHelp },
{ 1, N_("&Language"), Command::LOCALE, 0, KN, mHelp },
{ 1, N_("&Website / Manual"), Command::WEBSITE, 0, KN, mHelp },
2015-03-24 06:45:53 +00:00
#ifndef __APPLE__
{ 1, N_("&About"), Command::ABOUT, 0, KN, mHelp },
2015-03-24 06:45:53 +00:00
#endif
{ -1, 0, Command::NONE, 0, KN, NULL }
};
#undef S
#undef C
#undef F
#undef KN
#undef KC
#undef KR
void GraphicsWindow::ActivateCommand(Command cmd) {
for(int i = 0; Menu[i].level >= 0; i++) {
if(cmd == Menu[i].cmd) {
(Menu[i].fn)((Command)Menu[i].cmd);
break;
}
}
}
Platform::KeyboardEvent GraphicsWindow::AcceleratorForCommand(Command cmd) {
int rawAccel = 0;
for(int i = 0; Menu[i].level >= 0; i++) {
if(cmd == Menu[i].cmd) {
rawAccel = Menu[i].accel;
break;
}
}
Platform::KeyboardEvent accel = {};
if(rawAccel & SHIFT_MASK) {
accel.shiftDown = true;
}
if(rawAccel & CTRL_MASK) {
accel.controlDown = true;
}
if(rawAccel & FN_MASK) {
accel.key = Platform::KeyboardEvent::Key::FUNCTION;
accel.num = rawAccel & 0xff;
} else {
accel.key = Platform::KeyboardEvent::Key::CHARACTER;
accel.chr = (char)(rawAccel & 0xff);
}
return accel;
}
bool GraphicsWindow::KeyboardEvent(Platform::KeyboardEvent event) {
using Platform::KeyboardEvent;
if(event.type == KeyboardEvent::Type::RELEASE)
return true;
if(event.key == KeyboardEvent::Key::CHARACTER) {
if(event.chr == '\b') {
// Treat backspace identically to escape.
MenuEdit(Command::UNSELECT_ALL);
return true;
} else if(event.chr == '=') {
// Treat = as +. This is specific to US (and US-compatible) keyboard layouts,
// but makes zooming from keyboard much more usable on these.
// Ideally we'd have a platform-independent way of binding to a particular
// physical key regardless of shift status...
MenuView(Command::ZOOM_IN);
return true;
}
}
// On some platforms, the OS does not handle some or all keyboard accelerators,
// so handle them here.
for(int i = 0; Menu[i].level >= 0; i++) {
if(AcceleratorForCommand(Menu[i].cmd).Equals(event)) {
ActivateCommand(Menu[i].cmd);
return true;
}
}
return false;
}
void GraphicsWindow::PopulateMainMenu() {
bool unique = false;
Platform::MenuBarRef mainMenu = Platform::GetOrCreateMainMenu(&unique);
if(unique) mainMenu->Clear();
Platform::MenuRef currentSubMenu;
std::vector<Platform::MenuRef> subMenuStack;
for(int i = 0; Menu[i].level >= 0; i++) {
while(Menu[i].level > 0 && Menu[i].level <= (int)subMenuStack.size()) {
currentSubMenu = subMenuStack.back();
subMenuStack.pop_back();
}
if(Menu[i].label == NULL) {
currentSubMenu->AddSeparator();
continue;
}
std::string label = Translate(Menu[i].label);
if(Menu[i].level == 0) {
currentSubMenu = mainMenu->AddSubMenu(label);
} else if(Menu[i].cmd == Command::OPEN_RECENT) {
openRecentMenu = currentSubMenu->AddSubMenu(label);
} else if(Menu[i].cmd == Command::GROUP_RECENT) {
linkRecentMenu = currentSubMenu->AddSubMenu(label);
} else if(Menu[i].cmd == Command::LOCALE) {
Platform::MenuRef localeMenu = currentSubMenu->AddSubMenu(label);
for(const Locale &locale : Locales()) {
localeMenu->AddItem(locale.displayName, [&]() {
SetLocale(locale.Culture());
Platform::GetSettings()->FreezeString("Locale", locale.Culture());
SS.UpdateWindowTitles();
PopulateMainMenu();
EnsureValidActives();
});
}
} else if(Menu[i].fn == NULL) {
subMenuStack.push_back(currentSubMenu);
currentSubMenu = currentSubMenu->AddSubMenu(label);
} else {
Platform::MenuItemRef menuItem = currentSubMenu->AddItem(label);
menuItem->SetIndicator(Menu[i].kind);
if(Menu[i].accel != 0) {
menuItem->SetAccelerator(AcceleratorForCommand(Menu[i].cmd));
}
menuItem->onTrigger = std::bind(Menu[i].fn, Menu[i].cmd);
if(Menu[i].cmd == Command::SHOW_GRID) {
showGridMenuItem = menuItem;
} else if(Menu[i].cmd == Command::PERSPECTIVE_PROJ) {
perspectiveProjMenuItem = menuItem;
} else if(Menu[i].cmd == Command::SHOW_TOOLBAR) {
showToolbarMenuItem = menuItem;
} else if(Menu[i].cmd == Command::SHOW_TEXT_WND) {
showTextWndMenuItem = menuItem;
} else if(Menu[i].cmd == Command::FULL_SCREEN) {
fullScreenMenuItem = menuItem;
} else if(Menu[i].cmd == Command::UNITS_MM) {
unitsMmMenuItem = menuItem;
} else if(Menu[i].cmd == Command::UNITS_METERS) {
unitsMetersMenuItem = menuItem;
} else if(Menu[i].cmd == Command::UNITS_INCHES) {
unitsInchesMenuItem = menuItem;
} else if(Menu[i].cmd == Command::SEL_WORKPLANE) {
inWorkplaneMenuItem = menuItem;
} else if(Menu[i].cmd == Command::FREE_IN_3D) {
in3dMenuItem = menuItem;
} else if(Menu[i].cmd == Command::UNDO) {
undoMenuItem = menuItem;
} else if(Menu[i].cmd == Command::REDO) {
redoMenuItem = menuItem;
}
}
}
PopulateRecentFiles();
SS.UndoEnableMenus();
window->SetMenuBar(mainMenu);
}
static void PopulateMenuWithPathnames(Platform::MenuRef menu,
std::vector<Platform::Path> pathnames,
std::function<void(const Platform::Path &)> onTrigger) {
menu->Clear();
if(pathnames.empty()) {
Platform::MenuItemRef menuItem = menu->AddItem(_("(no recent files)"));
menuItem->SetEnabled(false);
} else {
for(Platform::Path pathname : pathnames) {
Platform::MenuItemRef menuItem = menu->AddItem(pathname.raw);
menuItem->onTrigger = [=]() { onTrigger(pathname); };
}
}
}
void GraphicsWindow::PopulateRecentFiles() {
PopulateMenuWithPathnames(openRecentMenu, SS.recentFiles, [](const Platform::Path &path) {
if(!SS.OkayToStartNewFile()) return;
SS.Load(path);
});
PopulateMenuWithPathnames(linkRecentMenu, SS.recentFiles, [](const Platform::Path &path) {
Group::MenuGroup(Command::GROUP_LINK, path);
});
}
void GraphicsWindow::Init() {
scale = 5;
offset = Vector::From(0, 0, 0);
projRight = Vector::From(1, 0, 0);
projUp = Vector::From(0, 1, 0);
// Make sure those are valid; could get a mouse move without a mouse
// down if someone depresses the button, then drags into our window.
orig.projRight = projRight;
orig.projUp = projUp;
// And with the last group active
activeGroup = SK.groupOrder.elem[SK.groupOrder.n - 1];
SK.GetGroup(activeGroup)->Activate();
showWorkplanes = false;
showNormals = true;
showPoints = true;
showConstraints = true;
showShaded = true;
showEdges = true;
showMesh = false;
showOutlines = false;
drawOccludedAs = DrawOccludedAs::INVISIBLE;
showTextWindow = true;
showSnapGrid = false;
context.active = false;
if(!window) {
window = Platform::CreateWindow();
if(window) {
canvas = CreateRenderer();
if(canvas) {
persistentCanvas = canvas->CreateBatch();
persistentDirty = true;
}
using namespace std::placeholders;
window->onClose = std::bind(&SolveSpaceUI::MenuFile, Command::EXIT);
window->onRender = std::bind(&GraphicsWindow::Paint, this);
window->onKeyboardEvent = std::bind(&GraphicsWindow::KeyboardEvent, this, _1);
window->onMouseEvent = std::bind(&GraphicsWindow::MouseEvent, this, _1);
window->onSixDofEvent = std::bind(&GraphicsWindow::SixDofEvent, this, _1);
window->onEditingDone = std::bind(&GraphicsWindow::EditControlDone, this, _1);
window->SetMinContentSize(720, 670);
PopulateMainMenu();
}
}
// Do this last, so that all the menus get updated correctly.
ClearSuper();
}
void GraphicsWindow::AnimateOntoWorkplane() {
if(!LockedInWorkplane()) return;
Entity *w = SK.GetEntity(ActiveWorkplane());
Quaternion quatf = w->Normal()->NormalGetNum();
Vector offsetf = (SK.GetEntity(w->point[0])->PointGetNum()).ScaledBy(-1);
// If the view screen is open, then we need to refresh it.
SS.ScheduleShowTW();
AnimateOnto(quatf, offsetf);
}
void GraphicsWindow::AnimateOnto(Quaternion quatf, Vector offsetf) {
// Get our initial orientation and translation.
Quaternion quat0 = Quaternion::From(projRight, projUp);
Vector offset0 = offset;
// Make sure we take the shorter of the two possible paths.
double mp = (quatf.Minus(quat0)).Magnitude();
double mm = (quatf.Plus(quat0)).Magnitude();
if(mp > mm) {
quatf = quatf.ScaledBy(-1);
mp = mm;
}
double mo = (offset0.Minus(offsetf)).Magnitude()*scale;
// Animate transition, unless it's a tiny move.
int64_t t0 = GetMilliseconds();
int32_t dt = (mp < 0.01 && mo < 10) ? (-20) :
(int32_t)(100 + 1000*mp + 0.4*mo);
// Don't ever animate for longer than 2000 ms; we can get absurdly
// long translations (as measured in pixels) if the user zooms out, moves,
// and then zooms in again.
if(dt > 2000) dt = 2000;
Quaternion dq = quatf.Times(quat0.Inverse());
if(!animateTimer) {
animateTimer = Platform::CreateTimer();
}
animateTimer->onTimeout = [=] {
int64_t tn = GetMilliseconds();
if((tn - t0) < dt) {
animateTimer->RunAfterNextFrame();
double s = (tn - t0)/((double)dt);
offset = (offset0.ScaledBy(1 - s)).Plus(offsetf.ScaledBy(s));
Quaternion quat = (dq.ToThe(s)).Times(quat0).WithMagnitude(1);
projRight = quat.RotationU();
projUp = quat.RotationV();
} else {
projRight = quatf.RotationU();
projUp = quatf.RotationV();
offset = offsetf;
}
window->Invalidate();
};
animateTimer->RunAfterNextFrame();
}
void GraphicsWindow::HandlePointForZoomToFit(Vector p, Point2d *pmax, Point2d *pmin,
double *wmin, bool usePerspective,
const Camera &camera)
{
double w;
Vector pp = camera.ProjectPoint4(p, &w);
// If usePerspective is true, then we calculate a perspective projection of the point.
// If not, then we do a parallel projection regardless of the current
// scale factor.
if(usePerspective) {
pp = pp.ScaledBy(1.0/w);
}
pmax->x = max(pmax->x, pp.x);
pmax->y = max(pmax->y, pp.y);
pmin->x = min(pmin->x, pp.x);
pmin->y = min(pmin->y, pp.y);
*wmin = min(*wmin, w);
}
void GraphicsWindow::LoopOverPoints(const std::vector<Entity *> &entities,
const std::vector<Constraint *> &constraints,
const std::vector<hEntity> &faces,
Point2d *pmax, Point2d *pmin, double *wmin,
bool usePerspective, bool includeMesh,
const Camera &camera) {
for(Entity *e : entities) {
if(e->IsPoint()) {
HandlePointForZoomToFit(e->PointGetNum(), pmax, pmin, wmin, usePerspective, camera);
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
} else if(e->type == Entity::Type::CIRCLE) {
// Lots of entities can extend outside the bbox of their points,
// but circles are particularly bad. We want to get things halfway
// reasonable without the mesh, because a zoom to fit is used to
// set the zoom level to set the chord tol.
double r = e->CircleGetRadiusNum();
Vector c = SK.GetEntity(e->point[0])->PointGetNum();
Quaternion q = SK.GetEntity(e->normal)->NormalGetNum();
for(int j = 0; j < 4; j++) {
Vector p = (j == 0) ? (c.Plus(q.RotationU().ScaledBy( r))) :
(j == 1) ? (c.Plus(q.RotationU().ScaledBy(-r))) :
(j == 2) ? (c.Plus(q.RotationV().ScaledBy( r))) :
(c.Plus(q.RotationV().ScaledBy(-r)));
HandlePointForZoomToFit(p, pmax, pmin, wmin, usePerspective, camera);
}
} else {
// We have to iterate children points, because we can select entities without points
for(int i = 0; i < MAX_POINTS_IN_ENTITY; i++) {
if(e->point[i].v == 0) break;
Vector p = SK.GetEntity(e->point[i])->PointGetNum();
HandlePointForZoomToFit(p, pmax, pmin, wmin, usePerspective, camera);
}
}
}
for(Constraint *c : constraints) {
std::vector<Vector> refs;
c->GetReferencePoints(camera, &refs);
for(Vector p : refs) {
HandlePointForZoomToFit(p, pmax, pmin, wmin, usePerspective, camera);
}
}
if(!includeMesh && faces.empty()) return;
Group *g = SK.GetGroup(activeGroup);
g->GenerateDisplayItems();
for(int i = 0; i < g->displayMesh.l.n; i++) {
STriangle *tr = &(g->displayMesh.l.elem[i]);
if(!includeMesh) {
bool found = false;
for(const hEntity &face : faces) {
if(face.v != tr->meta.face) continue;
found = true;
break;
}
if(!found) continue;
}
HandlePointForZoomToFit(tr->a, pmax, pmin, wmin, usePerspective, camera);
HandlePointForZoomToFit(tr->b, pmax, pmin, wmin, usePerspective, camera);
HandlePointForZoomToFit(tr->c, pmax, pmin, wmin, usePerspective, camera);
}
if(!includeMesh) return;
for(int i = 0; i < g->polyLoops.l.n; i++) {
SContour *sc = &(g->polyLoops.l.elem[i]);
for(int j = 0; j < sc->l.n; j++) {
HandlePointForZoomToFit(sc->l.elem[j].p, pmax, pmin, wmin, usePerspective, camera);
}
}
}
void GraphicsWindow::ZoomToFit(bool includingInvisibles, bool useSelection) {
if(!window) return;
scale = ZoomToFit(GetCamera(), includingInvisibles, useSelection);
}
double GraphicsWindow::ZoomToFit(const Camera &camera,
bool includingInvisibles, bool useSelection) {
std::vector<Entity *> entities;
std::vector<Constraint *> constraints;
std::vector<hEntity> faces;
if(useSelection) {
for(int i = 0; i < selection.n; i++) {
Selection *s = &selection.elem[i];
if(s->entity.v != 0) {
Entity *e = SK.entity.FindById(s->entity);
if(e->IsFace()) {
faces.push_back(e->h);
continue;
}
entities.push_back(e);
}
if(s->constraint.v != 0) {
Constraint *c = SK.constraint.FindById(s->constraint);
constraints.push_back(c);
}
}
}
bool selectionUsed = !entities.empty() || !constraints.empty() || !faces.empty();
if(!selectionUsed) {
for(Entity &e : SK.entity) {
// we don't want to handle separate points, because we will iterate them inside entities.
if(e.IsPoint()) continue;
if(!includingInvisibles && !e.IsVisible()) continue;
entities.push_back(&e);
}
for(Constraint &c : SK.constraint) {
if(!c.IsVisible()) continue;
constraints.push_back(&c);
}
}
// On the first run, ignore perspective.
Point2d pmax = { -1e12, -1e12 }, pmin = { 1e12, 1e12 };
double wmin = 1;
LoopOverPoints(entities, constraints, faces, &pmax, &pmin, &wmin,
/*usePerspective=*/false, /*includeMesh=*/!selectionUsed,
camera);
double xm = (pmax.x + pmin.x)/2, ym = (pmax.y + pmin.y)/2;
double dx = pmax.x - pmin.x, dy = pmax.y - pmin.y;
offset = offset.Plus(projRight.ScaledBy(-xm)).Plus(
projUp. ScaledBy(-ym));
2015-03-29 00:30:52 +00:00
// And based on this, we calculate the scale and offset
double scale;
if(EXACT(dx == 0 && dy == 0)) {
scale = 5;
} else {
double scalex = 1e12, scaley = 1e12;
if(EXACT(dx != 0)) scalex = 0.9*camera.width /dx;
if(EXACT(dy != 0)) scaley = 0.9*camera.height/dy;
scale = min(scalex, scaley);
scale = min(300.0, scale);
scale = max(0.003, scale);
}
// Then do another run, considering the perspective.
pmax.x = -1e12; pmax.y = -1e12;
pmin.x = 1e12; pmin.y = 1e12;
wmin = 1;
LoopOverPoints(entities, constraints, faces, &pmax, &pmin, &wmin,
/*usePerspective=*/true, /*includeMesh=*/!selectionUsed,
camera);
// Adjust the scale so that no points are behind the camera
if(wmin < 0.1) {
double k = camera.tangent;
// w = 1+k*scale*z
double zmin = (wmin - 1)/(k*scale);
// 0.1 = 1 + k*scale*zmin
// (0.1 - 1)/(k*zmin) = scale
scale = min(scale, (0.1 - 1)/(k*zmin));
}
return scale;
}
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
void GraphicsWindow::MenuView(Command id) {
switch(id) {
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::ZOOM_IN:
SS.GW.scale *= 1.2;
SS.ScheduleShowTW();
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::ZOOM_OUT:
SS.GW.scale /= 1.2;
SS.ScheduleShowTW();
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::ZOOM_TO_FIT:
SS.GW.ZoomToFit(/*includingInvisibles=*/false, /*useSelection=*/true);
SS.ScheduleShowTW();
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::SHOW_GRID:
SS.GW.showSnapGrid = !SS.GW.showSnapGrid;
if(SS.GW.showSnapGrid && !SS.GW.LockedInWorkplane()) {
Message(_("No workplane is active, so the grid will not appear."));
}
SS.GW.EnsureValidActives();
SS.GW.Invalidate();
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::PERSPECTIVE_PROJ:
SS.usePerspectiveProj = !SS.usePerspectiveProj;
if(SS.cameraTangent < 1e-6) {
Error(_("The perspective factor is set to zero, so the view will "
"always be a parallel projection.\n\n"
"For a perspective projection, modify the perspective "
"factor in the configuration screen. A value around 0.3 "
"is typical."));
}
SS.GW.EnsureValidActives();
SS.GW.Invalidate();
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::ONTO_WORKPLANE:
if(SS.GW.LockedInWorkplane()) {
SS.GW.AnimateOntoWorkplane();
break;
} // if not in 2d mode fall through and use ORTHO logic
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::NEAREST_ORTHO:
case Command::NEAREST_ISO: {
static const Vector ortho[3] = {
Vector::From(1, 0, 0),
Vector::From(0, 1, 0),
Vector::From(0, 0, 1)
};
double sqrt2 = sqrt(2.0), sqrt6 = sqrt(6.0);
Quaternion quat0 = Quaternion::From(SS.GW.projRight, SS.GW.projUp);
Quaternion quatf = quat0;
double dmin = 1e10;
// There are 24 possible views; 3*2*2*2
int i, j, negi, negj;
for(i = 0; i < 3; i++) {
for(j = 0; j < 3; j++) {
if(i == j) continue;
for(negi = 0; negi < 2; negi++) {
for(negj = 0; negj < 2; negj++) {
Vector ou = ortho[i], ov = ortho[j];
if(negi) ou = ou.ScaledBy(-1);
if(negj) ov = ov.ScaledBy(-1);
Vector on = ou.Cross(ov);
Vector u, v;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
if(id == Command::NEAREST_ORTHO || id == Command::ONTO_WORKPLANE) {
u = ou;
v = ov;
} else {
u =
ou.ScaledBy(1/sqrt2).Plus(
on.ScaledBy(-1/sqrt2));
v =
ou.ScaledBy(-1/sqrt6).Plus(
ov.ScaledBy(2/sqrt6).Plus(
on.ScaledBy(-1/sqrt6)));
}
Quaternion quatt = Quaternion::From(u, v);
double d = min(
(quatt.Minus(quat0)).Magnitude(),
(quatt.Plus(quat0)).Magnitude());
if(d < dmin) {
dmin = d;
quatf = quatt;
}
}
}
}
}
SS.GW.AnimateOnto(quatf, SS.GW.offset);
break;
}
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::CENTER_VIEW:
SS.GW.GroupSelection();
if(SS.GW.gs.n == 1 && SS.GW.gs.points == 1) {
Quaternion quat0;
// Offset is the selected point, quaternion is same as before
Vector pt = SK.GetEntity(SS.GW.gs.point[0])->PointGetNum();
quat0 = Quaternion::From(SS.GW.projRight, SS.GW.projUp);
SS.GW.ClearSelection();
SS.GW.AnimateOnto(quat0, pt.ScaledBy(-1));
} else {
Error(_("Select a point; this point will become the center "
"of the view on screen."));
}
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::SHOW_TOOLBAR:
SS.showToolbar = !SS.showToolbar;
SS.GW.EnsureValidActives();
SS.GW.Invalidate();
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::SHOW_TEXT_WND:
SS.GW.showTextWindow = !SS.GW.showTextWindow;
SS.GW.EnsureValidActives();
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::UNITS_INCHES:
SS.viewUnits = Unit::INCHES;
SS.ScheduleShowTW();
SS.GW.EnsureValidActives();
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::UNITS_MM:
SS.viewUnits = Unit::MM;
SS.ScheduleShowTW();
SS.GW.EnsureValidActives();
break;
case Command::UNITS_METERS:
SS.viewUnits = Unit::METERS;
SS.ScheduleShowTW();
SS.GW.EnsureValidActives();
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::FULL_SCREEN:
SS.GW.window->SetFullScreen(!SS.GW.window->IsFullScreen());
SS.GW.EnsureValidActives();
break;
default: ssassert(false, "Unexpected menu ID");
}
SS.GW.Invalidate();
}
void GraphicsWindow::EnsureValidActives() {
bool change = false;
// The active group must exist, and not be the references.
Group *g = SK.group.FindByIdNoOops(activeGroup);
if((!g) || (g->h.v == Group::HGROUP_REFERENCES.v)) {
int i;
for(i = 0; i < SK.groupOrder.n; i++) {
if(SK.groupOrder.elem[i].v != Group::HGROUP_REFERENCES.v) {
break;
}
}
if(i >= SK.groupOrder.n) {
// This can happen if the user deletes all the groups in the
// sketch. It's difficult to prevent that, because the last
// group might have been deleted automatically, because it failed
// a dependency. There needs to be something, so create a plane
// drawing group and activate that. They should never be able
// to delete the references, though.
activeGroup = SS.CreateDefaultDrawingGroup();
// We've created the default group, but not the workplane entity;
// do it now so that drawing mode isn't switched to "Free in 3d".
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
SS.GenerateAll(SolveSpaceUI::Generate::ALL);
} else {
activeGroup = SK.groupOrder.elem[i];
}
SK.GetGroup(activeGroup)->Activate();
change = true;
}
// The active coordinate system must also exist.
if(LockedInWorkplane()) {
Entity *e = SK.entity.FindByIdNoOops(ActiveWorkplane());
if(e) {
hGroup hgw = e->group;
if(hgw.v != activeGroup.v && SS.GroupsInOrder(activeGroup, hgw)) {
// The active workplane is in a group that comes after the
// active group; so any request or constraint will fail.
SetWorkplaneFreeIn3d();
change = true;
}
} else {
SetWorkplaneFreeIn3d();
change = true;
}
}
if(!window) return;
// And update the checked state for various menus
bool locked = LockedInWorkplane();
in3dMenuItem->SetActive(!locked);
inWorkplaneMenuItem->SetActive(locked);
SS.UndoEnableMenus();
switch(SS.viewUnits) {
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Unit::MM:
case Unit::METERS:
case Unit::INCHES:
break;
default:
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
SS.viewUnits = Unit::MM;
break;
}
unitsMmMenuItem->SetActive(SS.viewUnits == Unit::MM);
unitsMetersMenuItem->SetActive(SS.viewUnits == Unit::METERS);
unitsInchesMenuItem->SetActive(SS.viewUnits == Unit::INCHES);
if(SS.TW.window) SS.TW.window->SetVisible(SS.GW.showTextWindow);
showTextWndMenuItem->SetActive(SS.GW.showTextWindow);
showGridMenuItem->SetActive(SS.GW.showSnapGrid);
perspectiveProjMenuItem->SetActive(SS.usePerspectiveProj);
showToolbarMenuItem->SetActive(SS.showToolbar);
fullScreenMenuItem->SetActive(SS.GW.window->IsFullScreen());
if(change) SS.ScheduleShowTW();
}
void GraphicsWindow::SetWorkplaneFreeIn3d() {
SK.GetGroup(activeGroup)->activeWorkplane = Entity::FREE_IN_3D;
}
hEntity GraphicsWindow::ActiveWorkplane() {
Group *g = SK.group.FindByIdNoOops(activeGroup);
if(g) {
return g->activeWorkplane;
} else {
return Entity::FREE_IN_3D;
}
}
bool GraphicsWindow::LockedInWorkplane() {
return (SS.GW.ActiveWorkplane().v != Entity::FREE_IN_3D.v);
}
void GraphicsWindow::ForceTextWindowShown() {
if(!showTextWindow) {
showTextWindow = true;
showTextWndMenuItem->SetActive(true);
SS.TW.window->SetVisible(true);
}
}
void GraphicsWindow::DeleteTaggedRequests() {
// Rewrite any point-coincident constraints that were affected by this
// deletion.
Request *r;
for(r = SK.request.First(); r; r = SK.request.NextAfter(r)) {
if(!r->tag) continue;
FixConstraintsForRequestBeingDeleted(r->h);
}
// and then delete the tagged requests.
SK.request.RemoveTagged();
// An edit might be in progress for the just-deleted item. So
// now it's not.
window->HideEditor();
SS.TW.HideEditControl();
// And clear out the selection, which could contain that item.
ClearSuper();
// And regenerate to get rid of what it generates, plus anything
// that references it (since the regen code checks for that).
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
SS.GenerateAll(SolveSpaceUI::Generate::ALL);
EnsureValidActives();
SS.ScheduleShowTW();
}
Vector GraphicsWindow::SnapToGrid(Vector p) {
if(!LockedInWorkplane()) return p;
EntityBase *wrkpl = SK.GetEntity(ActiveWorkplane()),
*norm = wrkpl->Normal();
Vector wo = SK.GetEntity(wrkpl->point[0])->PointGetNum(),
wu = norm->NormalU(),
wv = norm->NormalV(),
wn = norm->NormalN();
Vector pp = (p.Minus(wo)).DotInToCsys(wu, wv, wn);
pp.x = floor((pp.x / SS.gridSpacing) + 0.5)*SS.gridSpacing;
pp.y = floor((pp.y / SS.gridSpacing) + 0.5)*SS.gridSpacing;
pp.z = 0;
2015-03-29 00:30:52 +00:00
return pp.ScaleOutOfCsys(wu, wv, wn).Plus(wo);
}
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
void GraphicsWindow::MenuEdit(Command id) {
switch(id) {
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::UNSELECT_ALL:
SS.GW.GroupSelection();
// If there's nothing selected to de-select, and no operation
// to cancel, then perhaps they want to return to the home
// screen in the text window.
if(SS.GW.gs.n == 0 &&
SS.GW.gs.constraints == 0 &&
SS.GW.pending.operation == Pending::NONE)
{
if(!(SS.TW.window->IsEditorVisible() ||
SS.GW.window->IsEditorVisible()))
{
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
if(SS.TW.shown.screen == TextWindow::Screen::STYLE_INFO) {
SS.TW.GoToScreen(TextWindow::Screen::LIST_OF_STYLES);
} else {
SS.TW.ClearSuper();
}
}
}
SS.GW.ClearSuper();
SS.TW.HideEditControl();
SS.nakedEdges.Clear();
SS.justExportedInfo.draw = false;
SS.centerOfMass.draw = false;
// This clears the marks drawn to indicate which points are
// still free to drag.
Param *p;
for(p = SK.param.First(); p; p = SK.param.NextAfter(p)) {
p->free = false;
}
if(SS.exportMode) {
SS.exportMode = false;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
SS.GenerateAll(SolveSpaceUI::Generate::ALL);
}
SS.GW.persistentDirty = true;
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::SELECT_ALL: {
Entity *e;
for(e = SK.entity.First(); e; e = SK.entity.NextAfter(e)) {
if(e->group.v != SS.GW.activeGroup.v) continue;
if(e->IsFace() || e->IsDistance()) continue;
if(!e->IsVisible()) continue;
SS.GW.MakeSelected(e->h);
}
SS.GW.Invalidate();
SS.ScheduleShowTW();
break;
}
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::SELECT_CHAIN: {
Entity *e;
int newlySelected = 0;
bool didSomething;
do {
didSomething = false;
for(e = SK.entity.First(); e; e = SK.entity.NextAfter(e)) {
if(e->group.v != SS.GW.activeGroup.v) continue;
if(!e->HasEndpoints()) continue;
if(!e->IsVisible()) continue;
Vector st = e->EndpointStart(),
fi = e->EndpointFinish();
2015-03-29 00:30:52 +00:00
bool onChain = false, alreadySelected = false;
List<Selection> *ls = &(SS.GW.selection);
for(Selection *s = ls->First(); s; s = ls->NextAfter(s)) {
if(!s->entity.v) continue;
if(s->entity.v == e->h.v) {
alreadySelected = true;
continue;
}
Entity *se = SK.GetEntity(s->entity);
if(!se->HasEndpoints()) continue;
Vector sst = se->EndpointStart(),
sfi = se->EndpointFinish();
if(sst.Equals(st) || sst.Equals(fi) ||
sfi.Equals(st) || sfi.Equals(fi))
{
onChain = true;
}
}
if(onChain && !alreadySelected) {
SS.GW.MakeSelected(e->h);
newlySelected++;
didSomething = true;
}
}
} while(didSomething);
if(newlySelected == 0) {
Error(_("No additional entities share endpoints with the selected entities."));
}
SS.GW.Invalidate();
SS.ScheduleShowTW();
break;
}
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::ROTATE_90: {
SS.GW.GroupSelection();
Entity *e = NULL;
if(SS.GW.gs.n == 1 && SS.GW.gs.points == 1) {
e = SK.GetEntity(SS.GW.gs.point[0]);
} else if(SS.GW.gs.n == 1 && SS.GW.gs.entities == 1) {
e = SK.GetEntity(SS.GW.gs.entity[0]);
}
SS.GW.ClearSelection();
hGroup hg = e ? e->group : SS.GW.activeGroup;
Group *g = SK.GetGroup(hg);
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
if(g->type != Group::Type::LINKED) {
Error(_("To use this command, select a point or other "
"entity from an linked part, or make a link "
"group the active group."));
break;
}
SS.UndoRemember();
// Rotate by ninety degrees about the coordinate axis closest
// to the screen normal.
Vector norm = SS.GW.projRight.Cross(SS.GW.projUp);
norm = norm.ClosestOrtho();
norm = norm.WithMagnitude(1);
Quaternion qaa = Quaternion::From(norm, PI/2);
g->TransformImportedBy(Vector::From(0, 0, 0), qaa);
// and regenerate as necessary.
SS.MarkGroupDirty(hg);
break;
}
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::SNAP_TO_GRID: {
if(!SS.GW.LockedInWorkplane()) {
Error(_("No workplane is active. Activate a workplane "
"(with Sketch -> In Workplane) to define the plane "
"for the snap grid."));
break;
}
SS.GW.GroupSelection();
if(SS.GW.gs.points == 0 && SS.GW.gs.constraintLabels == 0) {
Error(_("Can't snap these items to grid; select points, "
"text comments, or constraints with a label. "
"To snap a line, select its endpoints."));
break;
}
SS.UndoRemember();
List<Selection> *ls = &(SS.GW.selection);
for(Selection *s = ls->First(); s; s = ls->NextAfter(s)) {
if(s->entity.v) {
hEntity hp = s->entity;
Entity *ep = SK.GetEntity(hp);
if(!ep->IsPoint()) continue;
Vector p = ep->PointGetNum();
ep->PointForceTo(SS.GW.SnapToGrid(p));
SS.GW.pending.points.Add(&hp);
SS.MarkGroupDirty(ep->group);
} else if(s->constraint.v) {
Constraint *c = SK.GetConstraint(s->constraint);
Abstract all (ex-OpenGL) drawing operations into a Canvas interface. This has several desirable consequences: * It is now possible to port SolveSpace to a later version of OpenGL, such as OpenGLES 2, so that it runs on platforms that only have that OpenGL version; * The majority of geometry is now rendered without references to the camera in C++ code, so a renderer can now submit it to the video card once and re-rasterize with a different projection matrix every time the projection is changed, avoiding expensive reuploads; * The DOGD (draw or get distance) interface is now a straightforward Canvas implementation; * There are no more direct references to SS.GW.(projection) in sketch rendering code, which allows rendering to multiple viewports; * There are no more unnecessary framebuffer flips on CPU on Cocoa and GTK; * The platform-dependent GL code is now confined to rendergl1.cpp. * The Microsoft and Apple headers required by it that are prone to identifier conflicts are no longer included globally; * The rendergl1.cpp implementation can now be omitted from compilation to run SolveSpace headless or with a different OpenGL version. Note these implementation details of Canvas: * GetCamera currently always returns a reference to the field `Camera camera;`. This is so that a future renderer that caches geometry in the video memory can define it as asserting, which would provide assurance against code that could accidentally put something projection-dependent in the cache; * Line and triangle rendering is specified through a level of indirection, hStroke and hFill. This is so that a future renderer that batches geometry could cheaply group identical styles. * DrawPixmap and DrawVectorText accept a (o,u,v) and not a matrix. This is so that a future renderer into an output format that uses 2d transforms (e.g. SVG) could easily derive those. Some additional internal changes were required to enable this: * Pixmap is now always passed as std::shared_ptr<{const ,}Pixmap>. This is so that the renderer could cache uploaded textures between API calls, which requires it to capture a (weak) reference. * The PlatformPathEqual function was properly extracted into platform-specific code. This is so that the <windows.h> header could be included only where needed (in platform/w32* as well as rendergl1.cpp). * The SBsp{2,3}::DebugDraw functions were removed. They can be rewritten using the Canvas API if they are ever needed. While no visual changes were originally intended, some minor fixes happened anyway: * The "emphasis" yellow line from top-left corner is now correctly rendered much wider. * The marquee rectangle is now pixel grid aligned. * The hidden entities now do not clobber the depth buffer, removing some minor artifacts. * The workplane "tab" now scales with the font used to render the workplane name. * The workplane name font is now taken from the normals style. * Workplane and constraint line stipple is insignificantly different. This is so that it can reuse the existing stipple codepaths; rendering of workplanes and constraints predates those. Some debug functionality was added: * In graphics window, an fps counter that becomes red when rendering under 60fps is drawn.
2016-05-31 00:55:13 +00:00
std::vector<Vector> refs;
c->GetReferencePoints(SS.GW.GetCamera(), &refs);
c->disp.offset = c->disp.offset.Plus(SS.GW.SnapToGrid(refs[0]).Minus(refs[0]));
}
}
// Regenerate, with these points marked as dragged so that they
// get placed as close as possible to our snap grid.
SS.GW.ClearPending();
SS.GW.ClearSelection();
SS.GW.Invalidate();
break;
}
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::UNDO:
SS.UndoUndo();
break;
2015-03-29 00:30:52 +00:00
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::REDO:
SS.UndoRedo();
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::REGEN_ALL:
2016-11-29 16:49:20 +00:00
SS.images.clear();
SS.ReloadAllLinked(SS.saveFile);
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
SS.GenerateAll(SolveSpaceUI::Generate::UNTIL_ACTIVE);
SS.ScheduleShowTW();
break;
default: ssassert(false, "Unexpected menu ID");
}
}
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
void GraphicsWindow::MenuRequest(Command id) {
const char *s;
switch(id) {
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::SEL_WORKPLANE: {
SS.GW.GroupSelection();
Group *g = SK.GetGroup(SS.GW.activeGroup);
if(SS.GW.gs.n == 1 && SS.GW.gs.workplanes == 1) {
// A user-selected workplane
g->activeWorkplane = SS.GW.gs.entity[0];
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
} else if(g->type == Group::Type::DRAWING_WORKPLANE) {
// The group's default workplane
g->activeWorkplane = g->h.entity(0);
Message(_("No workplane selected. Activating default workplane "
"for this group."));
}
if(!SS.GW.LockedInWorkplane()) {
Error(_("No workplane is selected, and the active group does "
"not have a default workplane. Try selecting a "
"workplane, or activating a sketch-in-new-workplane "
"group."));
break;
}
// Align the view with the selected workplane
SS.GW.ClearSuper();
SS.GW.AnimateOntoWorkplane();
break;
}
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::FREE_IN_3D:
SS.GW.SetWorkplaneFreeIn3d();
SS.GW.EnsureValidActives();
SS.ScheduleShowTW();
SS.GW.Invalidate();
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::TANGENT_ARC:
SS.GW.GroupSelection();
if(SS.GW.gs.n == 1 && SS.GW.gs.points == 1) {
SS.GW.MakeTangentArc();
} else if(SS.GW.gs.n != 0) {
Error(_("Bad selection for tangent arc at point. Select a "
"single point, or select nothing to set up arc "
"parameters."));
} else {
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
SS.TW.GoToScreen(TextWindow::Screen::TANGENT_ARC);
SS.GW.ForceTextWindowShown();
SS.ScheduleShowTW();
SS.GW.Invalidate(); // repaint toolbar
}
break;
case Command::ARC: s = _("click point on arc (draws anti-clockwise)"); goto c;
case Command::DATUM_POINT: s = _("click to place datum point"); goto c;
case Command::LINE_SEGMENT: s = _("click first point of line segment"); goto c;
case Command::CONSTR_SEGMENT:
s = _("click first point of construction line segment"); goto c;
case Command::CUBIC: s = _("click first point of cubic segment"); goto c;
case Command::CIRCLE: s = _("click center of circle"); goto c;
case Command::WORKPLANE: s = _("click origin of workplane"); goto c;
case Command::RECTANGLE: s = _("click one corner of rectangle"); goto c;
case Command::TTF_TEXT: s = _("click top left of text"); goto c;
2016-11-29 16:49:20 +00:00
case Command::IMAGE:
if(!SS.ReloadLinkedImage(SS.saveFile, &SS.GW.pending.filename,
/*canCancel=*/true)) {
return;
}
s = _("click top left of image"); goto c;
c:
SS.GW.pending.operation = GraphicsWindow::Pending::COMMAND;
SS.GW.pending.command = id;
SS.GW.pending.description = s;
SS.ScheduleShowTW();
SS.GW.Invalidate(); // repaint toolbar
break;
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::CONSTRUCTION: {
SS.UndoRemember();
SS.GW.GroupSelection();
if(SS.GW.gs.entities == 0) {
Error(_("No entities are selected. Select entities before "
"trying to toggle their construction state."));
}
int i;
for(i = 0; i < SS.GW.gs.entities; i++) {
hEntity he = SS.GW.gs.entity[i];
if(!he.isFromRequest()) continue;
Request *r = SK.GetRequest(he.request());
r->construction = !(r->construction);
SS.MarkGroupDirty(r->group);
}
SS.GW.ClearSelection();
break;
}
Convert all enumerations to use `enum class`. 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.
2016-05-20 08:31:20 +00:00
case Command::SPLIT_CURVES:
SS.GW.SplitLinesOrCurves();
break;
default: ssassert(false, "Unexpected menu ID");
}
}
void GraphicsWindow::ClearSuper() {
if(window) window->HideEditor();
ClearPending();
ClearSelection();
hover.Clear();
EnsureValidActives();
}
void GraphicsWindow::ToggleBool(bool *v) {
*v = !*v;
// The faces are shown as special stippling on the shaded triangle mesh,
// so not meaningful to show them and hide the shaded.
if(!showShaded) showFaces = false;
// If the mesh or edges were previously hidden, they haven't been generated,
// and if we are going to show them, we need to generate them first.
Group *g = SK.GetGroup(SS.GW.activeGroup);
if(*v && (g->displayOutlines.l.n == 0 && (v == &showEdges || v == &showOutlines))) {
SS.GenerateAll(SolveSpaceUI::Generate::UNTIL_ACTIVE);
}
Invalidate(/*clearPersistent=*/true);
SS.ScheduleShowTW();
}
Enable exhaustive switch coverage warnings as an error, and use them. 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.
2016-05-25 06:55:50 +00:00
bool GraphicsWindow::SuggestLineConstraint(hRequest request, Constraint::Type *type) {
if(LockedInWorkplane()) {
Entity *ptA = SK.GetEntity(request.entity(1)),
*ptB = SK.GetEntity(request.entity(2));
Expr *au, *av, *bu, *bv;
ptA->PointGetExprsInWorkplane(ActiveWorkplane(), &au, &av);
ptB->PointGetExprsInWorkplane(ActiveWorkplane(), &bu, &bv);
double du = au->Minus(bu)->Eval();
double dv = av->Minus(bv)->Eval();
const double TOLERANCE_RATIO = 0.02;
Enable exhaustive switch coverage warnings as an error, and use them. 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.
2016-05-25 06:55:50 +00:00
if(fabs(dv) > LENGTH_EPS && fabs(du / dv) < TOLERANCE_RATIO) {
*type = Constraint::Type::VERTICAL;
return true;
} else if(fabs(du) > LENGTH_EPS && fabs(dv / du) < TOLERANCE_RATIO) {
*type = Constraint::Type::HORIZONTAL;
return true;
}
}
Enable exhaustive switch coverage warnings as an error, and use them. 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.
2016-05-25 06:55:50 +00:00
return false;
}