solvespace/src/solvespace.cpp

1036 lines
38 KiB
C++
Raw Normal View History

//-----------------------------------------------------------------------------
// Entry point in to the program, our registry-stored settings and top-level
// housekeeping when we open, save, and create new files.
//
// Copyright 2008-2013 Jonathan Westhues.
//-----------------------------------------------------------------------------
#include "solvespace.h"
#include "config.h"
SolveSpaceUI SolveSpace::SS = {};
Sketch SolveSpace::SK = {};
2015-03-24 06:45:53 +00:00
void SolveSpaceUI::Init() {
#if !defined(HEADLESS)
Implement a resource system. Currently, icons, fonts, etc are converted to C structures at compile time and are hardcoded to the binary. This presents several problems: * Cross-compilation is complicated. Right now, it is necessary to be able to run executables for the target platform; this happens to work with wine-binfmt installed, but is rather ugly. * Icons can only have one resolution. On OS X, modern software is expected to take advantage of high-DPI ("Retina") screens and use so-called @2x assets when ran in high-DPI mode. * Localization is complicated. Win32 and OS X provide built-in support for loading the resource appropriate for the user's locale. * Embedding strings can only be done as raw strings, using C++'s R"(...)" literals. This precludes embedding sizable strings, e.g. JavaScript libraries as used in Three.js export, and makes git history less useful. Not embedding the libraries means we have to rely on external CDNs, which requires an Internet connection and adds a glaring point of failure. * Linux distribution guidelines are violated. All architecture- independent data, especially large data such as fonts, is expected to be in /usr/share, not in the binary. * Customization is impossible without recompilation. Minor modifications like adding a few missing vector font characters or adjusting localization require a complete development environment, which is unreasonable to expect from users of a mechanical CAD. As such, this commit adds a resource system that bundles (and sometimes builds) resources with the executable. Where they go is platform-dependent: * on Win32: into resources of the executable, which allows us to keep distributing one file; * on OS X: into the app bundle; * on other *nix: into /usr/share/solvespace/ or ../res/ (relative to the executable path), the latter allowing us to run freshly built executables without installation. It also subsides the platform-specific resources that are in src/. The resource system is not yet used for anything; this will be added in later commits.
2016-04-21 15:54:18 +00:00
// Check that the resource system works.
dbp("%s", LoadString("banner.txt").data());
#endif
Implement a resource system. Currently, icons, fonts, etc are converted to C structures at compile time and are hardcoded to the binary. This presents several problems: * Cross-compilation is complicated. Right now, it is necessary to be able to run executables for the target platform; this happens to work with wine-binfmt installed, but is rather ugly. * Icons can only have one resolution. On OS X, modern software is expected to take advantage of high-DPI ("Retina") screens and use so-called @2x assets when ran in high-DPI mode. * Localization is complicated. Win32 and OS X provide built-in support for loading the resource appropriate for the user's locale. * Embedding strings can only be done as raw strings, using C++'s R"(...)" literals. This precludes embedding sizable strings, e.g. JavaScript libraries as used in Three.js export, and makes git history less useful. Not embedding the libraries means we have to rely on external CDNs, which requires an Internet connection and adds a glaring point of failure. * Linux distribution guidelines are violated. All architecture- independent data, especially large data such as fonts, is expected to be in /usr/share, not in the binary. * Customization is impossible without recompilation. Minor modifications like adding a few missing vector font characters or adjusting localization require a complete development environment, which is unreasonable to expect from users of a mechanical CAD. As such, this commit adds a resource system that bundles (and sometimes builds) resources with the executable. Where they go is platform-dependent: * on Win32: into resources of the executable, which allows us to keep distributing one file; * on OS X: into the app bundle; * on other *nix: into /usr/share/solvespace/ or ../res/ (relative to the executable path), the latter allowing us to run freshly built executables without installation. It also subsides the platform-specific resources that are in src/. The resource system is not yet used for anything; this will be added in later commits.
2016-04-21 15:54:18 +00:00
Platform::SettingsRef settings = Platform::GetSettings();
SS.tangentArcRadius = 10.0;
// Then, load the registry settings.
// Default list of colors for the model material
modelColor[0] = settings->ThawColor("ModelColor_0", RGBi(150, 150, 150));
modelColor[1] = settings->ThawColor("ModelColor_1", RGBi(100, 100, 100));
modelColor[2] = settings->ThawColor("ModelColor_2", RGBi( 30, 30, 30));
modelColor[3] = settings->ThawColor("ModelColor_3", RGBi(150, 0, 0));
modelColor[4] = settings->ThawColor("ModelColor_4", RGBi( 0, 100, 0));
modelColor[5] = settings->ThawColor("ModelColor_5", RGBi( 0, 80, 80));
modelColor[6] = settings->ThawColor("ModelColor_6", RGBi( 0, 0, 130));
modelColor[7] = settings->ThawColor("ModelColor_7", RGBi( 80, 0, 80));
// Light intensities
lightIntensity[0] = settings->ThawFloat("LightIntensity_0", 1.0);
lightIntensity[1] = settings->ThawFloat("LightIntensity_1", 0.5);
ambientIntensity = 0.3; // no setting for that yet
// Light positions
lightDir[0].x = settings->ThawFloat("LightDir_0_Right", -1.0);
lightDir[0].y = settings->ThawFloat("LightDir_0_Up", 1.0);
lightDir[0].z = settings->ThawFloat("LightDir_0_Forward", 0.0);
lightDir[1].x = settings->ThawFloat("LightDir_1_Right", 1.0);
lightDir[1].y = settings->ThawFloat("LightDir_1_Up", 0.0);
lightDir[1].z = settings->ThawFloat("LightDir_1_Forward", 0.0);
exportMode = false;
// Chord tolerance
chordTol = settings->ThawFloat("ChordTolerancePct", 0.5);
// Max pwl segments to generate
maxSegments = settings->ThawInt("MaxSegments", 10);
// Chord tolerance
exportChordTol = settings->ThawFloat("ExportChordTolerance", 0.1);
// Max pwl segments to generate
exportMaxSegments = settings->ThawInt("ExportMaxSegments", 64);
// View units
viewUnits = (Unit)settings->ThawInt("ViewUnits", (uint32_t)Unit::MM);
// Number of digits after the decimal point
afterDecimalMm = settings->ThawInt("AfterDecimalMm", 2);
afterDecimalInch = settings->ThawInt("AfterDecimalInch", 3);
// Camera tangent (determines perspective)
cameraTangent = settings->ThawFloat("CameraTangent", 0.3f/1e3);
// Grid spacing
gridSpacing = settings->ThawFloat("GridSpacing", 5.0);
// Export scale factor
exportScale = settings->ThawFloat("ExportScale", 1.0);
// Export offset (cutter radius comp)
exportOffset = settings->ThawFloat("ExportOffset", 0.0);
// Rewrite exported colors close to white into black (assuming white bg)
fixExportColors = settings->ThawBool("FixExportColors", true);
// Draw back faces of triangles (when mesh is leaky/self-intersecting)
drawBackFaces = settings->ThawBool("DrawBackFaces", true);
// Check that contours are closed and not self-intersecting
checkClosedContour = settings->ThawBool("CheckClosedContour", true);
// Enable automatic constrains for lines
automaticLineConstraints = settings->ThawBool("AutomaticLineConstraints", true);
// Draw closed polygons areas
showContourAreas = settings->ThawBool("ShowContourAreas", false);
// Export shaded triangles in a 2d view
exportShadedTriangles = settings->ThawBool("ExportShadedTriangles", true);
// Export pwl curves (instead of exact) always
exportPwlCurves = settings->ThawBool("ExportPwlCurves", false);
// Background color on-screen
backgroundColor = settings->ThawColor("BackgroundColor", RGBi(0, 0, 0));
// Whether export canvas size is fixed or derived from bbox
exportCanvasSizeAuto = settings->ThawBool("ExportCanvasSizeAuto", true);
// Margins for automatic canvas size
exportMargin.left = settings->ThawFloat("ExportMargin_Left", 5.0);
exportMargin.right = settings->ThawFloat("ExportMargin_Right", 5.0);
exportMargin.bottom = settings->ThawFloat("ExportMargin_Bottom", 5.0);
exportMargin.top = settings->ThawFloat("ExportMargin_Top", 5.0);
// Dimensions for fixed canvas size
exportCanvas.width = settings->ThawFloat("ExportCanvas_Width", 100.0);
exportCanvas.height = settings->ThawFloat("ExportCanvas_Height", 100.0);
exportCanvas.dx = settings->ThawFloat("ExportCanvas_Dx", 5.0);
exportCanvas.dy = settings->ThawFloat("ExportCanvas_Dy", 5.0);
// Extra parameters when exporting G code
gCode.depth = settings->ThawFloat("GCode_Depth", 10.0);
gCode.passes = settings->ThawInt("GCode_Passes", 1);
gCode.feed = settings->ThawFloat("GCode_Feed", 10.0);
gCode.plungeFeed = settings->ThawFloat("GCode_PlungeFeed", 10.0);
// Show toolbar in the graphics window
showToolbar = settings->ThawBool("ShowToolbar", true);
// Recent files menus
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
for(size_t i = 0; i < MAX_RECENT; i++) {
std::string rawPath = settings->ThawString("RecentFile_" + std::to_string(i), "");
if(rawPath.empty()) continue;
recentFiles.push_back(Platform::Path::From(rawPath));
}
// Autosave timer
autosaveInterval = settings->ThawInt("AutosaveInterval", 5);
2017-01-05 10:39:08 +00:00
// Locale
std::string locale = settings->ThawString("Locale", "");
2017-01-05 10:39:08 +00:00
if(!locale.empty()) {
SetLocale(locale);
}
generateAllTimer = Platform::CreateTimer();
generateAllTimer->onTimeout = std::bind(&SolveSpaceUI::GenerateAll, &SS, Generate::DIRTY,
/*andFindFree=*/false, /*genForBBox=*/false);
showTWTimer = Platform::CreateTimer();
showTWTimer->onTimeout = std::bind(&TextWindow::Show, &TW);
autosaveTimer = Platform::CreateTimer();
autosaveTimer->onTimeout = std::bind(&SolveSpaceUI::Autosave, &SS);
// The default styles (colors, line widths, etc.) are also stored in the
// configuration file, but we will automatically load those as we need
// them.
ScheduleAutosave();
NewFile();
AfterNewFile();
if(TW.window && GW.window) {
TW.window->ThawPosition(settings, "TextWindow");
GW.window->ThawPosition(settings, "GraphicsWindow");
TW.window->SetVisible(true);
GW.window->SetVisible(true);
GW.window->Focus();
// Do this once the window is created.
Request3DConnexionEventsForWindow(GW.window);
}
2015-03-24 06:45:53 +00:00
}
bool SolveSpaceUI::LoadAutosaveFor(const Platform::Path &filename) {
Platform::Path autosaveFile = filename.WithExtension(AUTOSAVE_EXT);
FILE *f = OpenFile(autosaveFile, "rb");
if(!f)
return false;
fclose(f);
Platform::MessageDialogRef dialog = CreateMessageDialog(GW.window);
using Platform::MessageDialog;
dialog->SetType(MessageDialog::Type::QUESTION);
dialog->SetTitle(C_("title", "Autosave Available"));
dialog->SetMessage(C_("dialog", "An autosave file is available for this sketch."));
dialog->SetDescription(C_("dialog", "Do you want to load the autosave file instead?"));
dialog->AddButton(C_("button", "&Load autosave"), MessageDialog::Response::YES,
/*isDefault=*/true);
dialog->AddButton(C_("button", "Do&n't Load"), MessageDialog::Response::NO);
// FIXME(async): asyncify this call
if(dialog->RunModal() == MessageDialog::Response::YES) {
unsaved = true;
return LoadFromFile(autosaveFile, /*canCancel=*/true);
}
return false;
}
bool SolveSpaceUI::Load(const Platform::Path &filename) {
bool autosaveLoaded = LoadAutosaveFor(filename);
bool fileLoaded = autosaveLoaded || LoadFromFile(filename, /*canCancel=*/true);
if(fileLoaded) {
saveFile = filename;
2015-03-24 06:45:53 +00:00
AddToRecentList(filename);
} else {
saveFile.Clear();
2015-03-24 06:45:53 +00:00
NewFile();
}
AfterNewFile();
unsaved = autosaveLoaded;
return fileLoaded;
}
void SolveSpaceUI::Exit() {
Platform::SettingsRef settings = Platform::GetSettings();
GW.window->FreezePosition(settings, "GraphicsWindow");
TW.window->FreezePosition(settings, "TextWindow");
// Recent files
for(size_t i = 0; i < MAX_RECENT; i++) {
std::string rawPath;
if(recentFiles.size() > i) {
rawPath = recentFiles[i].raw;
}
settings->FreezeString("RecentFile_" + std::to_string(i), rawPath);
}
// Model colors
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
for(size_t i = 0; i < MODEL_COLORS; i++)
settings->FreezeColor("ModelColor_" + std::to_string(i), modelColor[i]);
// Light intensities
settings->FreezeFloat("LightIntensity_0", (float)lightIntensity[0]);
settings->FreezeFloat("LightIntensity_1", (float)lightIntensity[1]);
// Light directions
settings->FreezeFloat("LightDir_0_Right", (float)lightDir[0].x);
settings->FreezeFloat("LightDir_0_Up", (float)lightDir[0].y);
settings->FreezeFloat("LightDir_0_Forward", (float)lightDir[0].z);
settings->FreezeFloat("LightDir_1_Right", (float)lightDir[1].x);
settings->FreezeFloat("LightDir_1_Up", (float)lightDir[1].y);
settings->FreezeFloat("LightDir_1_Forward", (float)lightDir[1].z);
// Chord tolerance
settings->FreezeFloat("ChordTolerancePct", (float)chordTol);
// Max pwl segments to generate
settings->FreezeInt("MaxSegments", (uint32_t)maxSegments);
// Export Chord tolerance
settings->FreezeFloat("ExportChordTolerance", (float)exportChordTol);
// Export Max pwl segments to generate
settings->FreezeInt("ExportMaxSegments", (uint32_t)exportMaxSegments);
// View units
settings->FreezeInt("ViewUnits", (uint32_t)viewUnits);
// Number of digits after the decimal point
settings->FreezeInt("AfterDecimalMm", (uint32_t)afterDecimalMm);
settings->FreezeInt("AfterDecimalInch", (uint32_t)afterDecimalInch);
// Camera tangent (determines perspective)
settings->FreezeFloat("CameraTangent", (float)cameraTangent);
// Grid spacing
settings->FreezeFloat("GridSpacing", gridSpacing);
// Export scale
settings->FreezeFloat("ExportScale", exportScale);
// Export offset (cutter radius comp)
settings->FreezeFloat("ExportOffset", exportOffset);
// Rewrite exported colors close to white into black (assuming white bg)
settings->FreezeBool("FixExportColors", fixExportColors);
// Draw back faces of triangles (when mesh is leaky/self-intersecting)
settings->FreezeBool("DrawBackFaces", drawBackFaces);
// Draw closed polygons areas
settings->FreezeBool("ShowContourAreas", showContourAreas);
// Check that contours are closed and not self-intersecting
settings->FreezeBool("CheckClosedContour", checkClosedContour);
// Enable automatic constrains for lines
settings->FreezeBool("AutomaticLineConstraints", automaticLineConstraints);
// Export shaded triangles in a 2d view
settings->FreezeBool("ExportShadedTriangles", exportShadedTriangles);
// Export pwl curves (instead of exact) always
settings->FreezeBool("ExportPwlCurves", exportPwlCurves);
// Background color on-screen
settings->FreezeColor("BackgroundColor", backgroundColor);
// Whether export canvas size is fixed or derived from bbox
settings->FreezeBool("ExportCanvasSizeAuto", exportCanvasSizeAuto);
// Margins for automatic canvas size
settings->FreezeFloat("ExportMargin_Left", exportMargin.left);
settings->FreezeFloat("ExportMargin_Right", exportMargin.right);
settings->FreezeFloat("ExportMargin_Bottom", exportMargin.bottom);
settings->FreezeFloat("ExportMargin_Top", exportMargin.top);
// Dimensions for fixed canvas size
settings->FreezeFloat("ExportCanvas_Width", exportCanvas.width);
settings->FreezeFloat("ExportCanvas_Height", exportCanvas.height);
settings->FreezeFloat("ExportCanvas_Dx", exportCanvas.dx);
settings->FreezeFloat("ExportCanvas_Dy", exportCanvas.dy);
// Extra parameters when exporting G code
settings->FreezeFloat("GCode_Depth", gCode.depth);
settings->FreezeInt("GCode_Passes", gCode.passes);
settings->FreezeFloat("GCode_Feed", gCode.feed);
settings->FreezeFloat("GCode_PlungeFeed", gCode.plungeFeed);
// Show toolbar in the graphics window
settings->FreezeBool("ShowToolbar", showToolbar);
// Autosave timer
settings->FreezeInt("AutosaveInterval", autosaveInterval);
// And the default styles, colors and line widths and such.
Style::FreezeDefaultStyles(settings);
Platform::ExitGui();
}
void SolveSpaceUI::ScheduleGenerateAll() {
generateAllTimer->RunAfterProcessingEvents();
}
void SolveSpaceUI::ScheduleShowTW() {
showTWTimer->RunAfterProcessingEvents();
}
void SolveSpaceUI::ScheduleAutosave() {
autosaveTimer->RunAfter(autosaveInterval * 60 * 1000);
}
double SolveSpaceUI::MmPerUnit() {
switch(viewUnits) {
case Unit::INCHES: return 25.4;
case Unit::METERS: return 1000.0;
case Unit::MM: return 1.0;
}
return 1.0;
}
const char *SolveSpaceUI::UnitName() {
switch(viewUnits) {
case Unit::INCHES: return "inch";
case Unit::METERS: return "m";
case Unit::MM: return "mm";
}
return "";
}
std::string SolveSpaceUI::MmToString(double v) {
switch(viewUnits) {
case Unit::INCHES: return ssprintf("%.*f", afterDecimalInch, v / 25.4);
case Unit::METERS: return ssprintf("%.*f", afterDecimalMm, v / 1000.0);
case Unit::MM: return ssprintf("%.*f", afterDecimalMm, v);
}
return "";
}
double SolveSpaceUI::ExprToMm(Expr *e) {
return (e->Eval()) * MmPerUnit();
}
double SolveSpaceUI::StringToMm(const std::string &str) {
return std::stod(str) * MmPerUnit();
}
double SolveSpaceUI::ChordTolMm() {
if(exportMode) return ExportChordTolMm();
return chordTolCalculated;
}
double SolveSpaceUI::ExportChordTolMm() {
return exportChordTol / exportScale;
}
int SolveSpaceUI::GetMaxSegments() {
if(exportMode) return exportMaxSegments;
return maxSegments;
}
int SolveSpaceUI::UnitDigitsAfterDecimal() {
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
return (viewUnits == Unit::INCHES) ? afterDecimalInch : afterDecimalMm;
}
void SolveSpaceUI::SetUnitDigitsAfterDecimal(int 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(viewUnits == Unit::INCHES) {
afterDecimalInch = v;
} else {
afterDecimalMm = v;
}
}
double SolveSpaceUI::CameraTangent() {
if(!usePerspectiveProj) {
return 0;
} else {
return cameraTangent;
}
}
void SolveSpaceUI::AfterNewFile() {
// Clear out the traced point, which is no longer valid
traced.point = Entity::NO_ENTITY;
traced.path.l.Clear();
// and the naked edges
nakedEdges.Clear();
2016-04-16 09:05:12 +00:00
// Quit export mode
justExportedInfo.draw = false;
centerOfMass.draw = false;
2016-04-16 09:05:12 +00:00
exportMode = false;
// GenerateAll() expects the view to be valid, because it uses that to
// fill in default values for extrusion depths etc. (which won't matter
// here, but just don't let it work on garbage)
SS.GW.offset = Vector::From(0, 0, 0);
SS.GW.projRight = Vector::From(1, 0, 0);
SS.GW.projUp = Vector::From(0, 1, 0);
GenerateAll(Generate::ALL);
GW.Init();
TW.Init();
unsaved = false;
GW.ZoomToFit();
// Create all the default styles; they'll get created on the fly anyways,
// but can't hurt to do it now.
Style::CreateAllDefaultStyles();
UpdateWindowTitles();
}
void SolveSpaceUI::AddToRecentList(const Platform::Path &filename) {
auto it = std::find_if(recentFiles.begin(), recentFiles.end(),
[&](const Platform::Path &p) { return p.Equals(filename); });
if(it != recentFiles.end()) {
recentFiles.erase(it);
}
if(recentFiles.size() > MAX_RECENT) {
recentFiles.erase(recentFiles.begin() + MAX_RECENT);
}
recentFiles.insert(recentFiles.begin(), filename);
GW.PopulateRecentFiles();
}
bool SolveSpaceUI::GetFilenameAndSave(bool saveAs) {
Platform::SettingsRef settings = Platform::GetSettings();
Platform::Path newSaveFile = saveFile;
if(saveAs || saveFile.IsEmpty()) {
Platform::FileDialogRef dialog = Platform::CreateSaveFileDialog(GW.window);
dialog->AddFilter(C_("file-type", "SolveSpace models"), { "slvs" });
dialog->ThawChoices(settings, "Sketch");
if(!newSaveFile.IsEmpty()) {
dialog->SetFilename(newSaveFile);
}
if(dialog->RunModal()) {
dialog->FreezeChoices(settings, "Sketch");
newSaveFile = dialog->GetFilename();
} else {
return false;
}
}
if(SaveToFile(newSaveFile)) {
AddToRecentList(newSaveFile);
RemoveAutosave();
saveFile = newSaveFile;
unsaved = false;
return true;
} else {
return false;
}
}
void SolveSpaceUI::Autosave()
{
ScheduleAutosave();
if(!saveFile.IsEmpty() && unsaved) {
SaveToFile(saveFile.WithExtension(AUTOSAVE_EXT));
}
}
void SolveSpaceUI::RemoveAutosave()
{
Platform::Path autosaveFile = saveFile.WithExtension(AUTOSAVE_EXT);
RemoveFile(autosaveFile);
}
bool SolveSpaceUI::OkayToStartNewFile() {
if(!unsaved) return true;
Platform::MessageDialogRef dialog = CreateMessageDialog(GW.window);
using Platform::MessageDialog;
dialog->SetType(MessageDialog::Type::QUESTION);
dialog->SetTitle(C_("title", "Modified File"));
if(!SolveSpace::SS.saveFile.IsEmpty()) {
dialog->SetMessage(ssprintf(C_("dialog", "Do you want to save the changes you made to "
"the sketch “%s”?"), saveFile.raw.c_str()));
} else {
dialog->SetMessage(C_("dialog", "Do you want to save the changes you made to "
"the new sketch?"));
}
dialog->SetDescription(C_("dialog", "Your changes will be lost if you don't save them."));
dialog->AddButton(C_("button", "&Save"), MessageDialog::Response::YES,
/*isDefault=*/true);
dialog->AddButton(C_("button", "Do&n't Save"), MessageDialog::Response::NO);
dialog->AddButton(C_("button", "&Cancel"), MessageDialog::Response::CANCEL);
// FIXME(async): asyncify this call
switch(dialog->RunModal()) {
case MessageDialog::Response::YES:
return GetFilenameAndSave(/*saveAs=*/false);
case MessageDialog::Response::NO:
RemoveAutosave();
return true;
default:
return false;
}
}
void SolveSpaceUI::UpdateWindowTitles() {
if(!GW.window || !TW.window) return;
if(saveFile.IsEmpty()) {
GW.window->SetTitle(C_("title", "(new sketch)"));
} else {
if(!GW.window->SetTitleForFilename(saveFile)) {
GW.window->SetTitle(saveFile.raw);
}
}
TW.window->SetTitle(C_("title", "Property Browser"));
}
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 SolveSpaceUI::MenuFile(Command id) {
Platform::SettingsRef settings = Platform::GetSettings();
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::NEW:
if(!SS.OkayToStartNewFile()) break;
SS.saveFile.Clear();
SS.NewFile();
SS.AfterNewFile();
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::OPEN: {
if(!SS.OkayToStartNewFile()) break;
Platform::FileDialogRef dialog = Platform::CreateOpenFileDialog(SS.GW.window);
dialog->AddFilters(Platform::SolveSpaceModelFileFilters);
dialog->ThawChoices(settings, "Sketch");
if(dialog->RunModal()) {
dialog->FreezeChoices(settings, "Sketch");
SS.Load(dialog->GetFilename());
}
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::SAVE:
SS.GetFilenameAndSave(/*saveAs=*/false);
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::SAVE_AS:
SS.GetFilenameAndSave(/*saveAs=*/true);
break;
case Command::EXPORT_IMAGE: {
Platform::FileDialogRef dialog = Platform::CreateSaveFileDialog(SS.GW.window);
dialog->AddFilters(Platform::RasterFileFilters);
dialog->ThawChoices(settings, "ExportImage");
if(dialog->RunModal()) {
dialog->FreezeChoices(settings, "ExportImage");
SS.ExportAsPngTo(dialog->GetFilename());
}
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::EXPORT_VIEW: {
Platform::FileDialogRef dialog = Platform::CreateSaveFileDialog(SS.GW.window);
dialog->AddFilters(Platform::VectorFileFilters);
dialog->ThawChoices(settings, "ExportView");
if(!dialog->RunModal()) break;
dialog->FreezeChoices(settings, "ExportView");
// If the user is exporting something where it would be
// inappropriate to include the constraints, then warn.
if(SS.GW.showConstraints &&
(dialog->GetFilename().HasExtension("txt") ||
fabs(SS.exportOffset) > LENGTH_EPS))
{
Message(_("Constraints are currently shown, and will be exported "
"in the toolpath. This is probably not what you want; "
"hide them by clicking the link at the top of the "
"text window."));
}
SS.ExportViewOrWireframeTo(dialog->GetFilename(), /*exportWireframe=*/false);
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::EXPORT_WIREFRAME: {
Platform::FileDialogRef dialog = Platform::CreateSaveFileDialog(SS.GW.window);
dialog->AddFilters(Platform::Vector3dFileFilters);
dialog->ThawChoices(settings, "ExportWireframe");
if(!dialog->RunModal()) break;
dialog->FreezeChoices(settings, "ExportWireframe");
SS.ExportViewOrWireframeTo(dialog->GetFilename(), /*exportWireframe*/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::EXPORT_SECTION: {
Platform::FileDialogRef dialog = Platform::CreateSaveFileDialog(SS.GW.window);
dialog->AddFilters(Platform::VectorFileFilters);
dialog->ThawChoices(settings, "ExportSection");
if(!dialog->RunModal()) break;
dialog->FreezeChoices(settings, "ExportSection");
SS.ExportSectionTo(dialog->GetFilename());
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::EXPORT_MESH: {
Platform::FileDialogRef dialog = Platform::CreateSaveFileDialog(SS.GW.window);
dialog->AddFilters(Platform::MeshFileFilters);
dialog->ThawChoices(settings, "ExportMesh");
if(!dialog->RunModal()) break;
dialog->FreezeChoices(settings, "ExportMesh");
SS.ExportMeshTo(dialog->GetFilename());
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::EXPORT_SURFACES: {
Platform::FileDialogRef dialog = Platform::CreateSaveFileDialog(SS.GW.window);
dialog->AddFilters(Platform::SurfaceFileFilters);
dialog->ThawChoices(settings, "ExportSurfaces");
if(!dialog->RunModal()) break;
dialog->FreezeChoices(settings, "ExportSurfaces");
StepFileWriter sfw = {};
sfw.ExportSurfacesTo(dialog->GetFilename());
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::IMPORT: {
Platform::FileDialogRef dialog = Platform::CreateOpenFileDialog(SS.GW.window);
dialog->AddFilters(Platform::ImportFileFilters);
dialog->ThawChoices(settings, "Import");
if(!dialog->RunModal()) break;
dialog->FreezeChoices(settings, "Import");
2016-04-13 08:43:06 +00:00
Platform::Path importFile = dialog->GetFilename();
if(importFile.HasExtension("dxf")) {
ImportDxf(importFile);
} else if(importFile.HasExtension("dwg")) {
ImportDwg(importFile);
} else {
Error(_("Can't identify file type from file extension of "
"filename '%s'; try .dxf or .dwg."), importFile.raw.c_str());
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
SS.GenerateAll(SolveSpaceUI::Generate::UNTIL_ACTIVE);
SS.ScheduleShowTW();
2016-04-13 08:43:06 +00:00
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::EXIT:
if(!SS.OkayToStartNewFile()) break;
SS.Exit();
break;
default: ssassert(false, "Unexpected menu ID");
}
SS.UpdateWindowTitles();
}
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 SolveSpaceUI::MenuAnalyze(Command id) {
Platform::SettingsRef settings = Platform::GetSettings();
SS.GW.GroupSelection();
auto const &gs = SS.GW.gs;
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::STEP_DIM:
if(gs.constraints == 1 && gs.n == 0) {
Constraint *c = SK.GetConstraint(gs.constraint[0]);
if(c->HasLabel() && !c->reference) {
SS.TW.stepDim.finish = c->valA;
SS.TW.stepDim.steps = 10;
SS.TW.stepDim.isDistance =
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
(c->type != Constraint::Type::ANGLE) &&
(c->type != Constraint::Type::LENGTH_RATIO) &&
(c->type != Constraint::Type::LENGTH_DIFFERENCE);
SS.TW.shown.constraint = c->h;
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.shown.screen = TextWindow::Screen::STEP_DIMENSION;
// The step params are specified in the text window,
// so force that to be shown.
SS.GW.ForceTextWindowShown();
SS.ScheduleShowTW();
SS.GW.ClearSelection();
} else {
Error(_("Constraint must have a label, and must not be "
"a reference dimension."));
}
} else {
Error(_("Bad selection for step dimension; select a constraint."));
}
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::NAKED_EDGES: {
ShowNakedEdges(/*reportOnlyWhenNotOkay=*/false);
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::INTERFERENCE: {
SS.nakedEdges.Clear();
SMesh *m = &(SK.GetGroup(SS.GW.activeGroup)->displayMesh);
SKdNode *root = SKdNode::From(m);
bool inters, leaks;
root->MakeCertainEdgesInto(&(SS.nakedEdges),
EdgeKind::SELF_INTER, /*coplanarIsInter=*/false, &inters, &leaks);
SS.GW.Invalidate();
if(inters) {
Error("%d edges interfere with other triangles, bad.",
SS.nakedEdges.l.n);
} else {
Message(_("The assembly does not interfere, good."));
}
break;
}
case Command::CENTER_OF_MASS: {
SS.UpdateCenterOfMass();
SS.centerOfMass.draw = true;
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::VOLUME: {
SMesh *m = &(SK.GetGroup(SS.GW.activeGroup)->displayMesh);
2015-03-29 00:30:52 +00:00
double vol = 0;
int i;
for(i = 0; i < m->l.n; i++) {
STriangle tr = m->l.elem[i];
// Translate to place vertex A at (x, y, 0)
Vector trans = Vector::From(tr.a.x, tr.a.y, 0);
tr.a = (tr.a).Minus(trans);
tr.b = (tr.b).Minus(trans);
tr.c = (tr.c).Minus(trans);
// Rotate to place vertex B on the y-axis. Depending on
// whether the triangle is CW or CCW, C is either to the
// right or to the left of the y-axis. This handles the
// sign of our normal.
Vector u = Vector::From(-tr.b.y, tr.b.x, 0);
u = u.WithMagnitude(1);
Vector v = Vector::From(tr.b.x, tr.b.y, 0);
v = v.WithMagnitude(1);
Vector n = Vector::From(0, 0, 1);
tr.a = (tr.a).DotInToCsys(u, v, n);
tr.b = (tr.b).DotInToCsys(u, v, n);
tr.c = (tr.c).DotInToCsys(u, v, n);
n = tr.Normal().WithMagnitude(1);
// Triangles on edge don't contribute
if(fabs(n.z) < LENGTH_EPS) continue;
2015-03-29 00:30:52 +00:00
// The plane has equation p dot n = a dot n
double d = (tr.a).Dot(n);
// nx*x + ny*y + nz*z = d
// nz*z = d - nx*x - ny*y
double A = -n.x/n.z, B = -n.y/n.z, C = d/n.z;
double mac = tr.c.y/tr.c.x, mbc = (tr.c.y - tr.b.y)/tr.c.x;
double xc = tr.c.x, yb = tr.b.y;
2015-03-29 00:30:52 +00:00
// I asked Maple for
// int(int(A*x + B*y +C, y=mac*x..(mbc*x + yb)), x=0..xc);
2015-03-29 00:30:52 +00:00
double integral =
(1.0/3)*(
A*(mbc-mac)+
(1.0/2)*B*(mbc*mbc-mac*mac)
)*(xc*xc*xc)+
(1.0/2)*(A*yb+B*yb*mbc+C*(mbc-mac))*xc*xc+
C*yb*xc+
(1.0/2)*B*yb*yb*xc;
vol += integral;
}
std::string msg = ssprintf(_("The volume of the solid model is:\n\n"
" %.3f %s^3"),
vol / pow(SS.MmPerUnit(), 3),
SS.UnitName());
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.viewUnits == Unit::MM) {
msg += ssprintf("\n %.2f mL", vol/(10*10*10));
}
msg += _("\n\nCurved surfaces have been approximated as triangles.\n"
"This introduces error, typically of around 1%.");
Message("%s", msg.c_str());
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::AREA: {
Group *g = SK.GetGroup(SS.GW.activeGroup);
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->polyError.how != PolyError::GOOD) {
Error(_("This group does not contain a correctly-formed "
"2d closed area. It is open, not coplanar, or self-"
"intersecting."));
break;
}
SEdgeList sel = {};
g->polyLoops.MakeEdgesInto(&sel);
SPolygon sp = {};
sel.AssemblePolygon(&sp, NULL, /*keepDir=*/true);
sp.normal = sp.ComputeNormal();
sp.FixContourDirections();
double area = sp.SignedArea();
double scale = SS.MmPerUnit();
Message(_("The area of the region sketched in this group is:\n\n"
" %.3f %s^2\n\n"
"Curves have been approximated as piecewise linear.\n"
"This introduces error, typically of around 1%%."),
area / (scale*scale),
SS.UnitName());
sel.Clear();
sp.Clear();
break;
}
case Command::PERIMETER: {
if(gs.n > 0 && gs.n == gs.entities) {
double perimeter = 0.0;
for(int i = 0; i < gs.entities; i++) {
Entity *e = SK.entity.FindById(gs.entity[i]);
SEdgeList *el = e->GetOrGenerateEdges();
for(const SEdge &e : el->l) {
perimeter += e.b.Minus(e.a).Magnitude();
}
}
double scale = SS.MmPerUnit();
Message(_("The total length of the selected entities is:\n\n"
" %.3f %s\n\n"
"Curves have been approximated as piecewise linear.\n"
"This introduces error, typically of around 1%%."),
perimeter / scale,
SS.UnitName());
} else {
Error(_("Bad selection for perimeter; select line segments, arcs, and curves."));
}
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_DOF:
// This works like a normal solve, except that it calculates
// which variables are free/bound at the same time.
SS.GenerateAll(SolveSpaceUI::Generate::ALL, /*andFindFree=*/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::TRACE_PT:
if(gs.points == 1 && gs.n == 1) {
SS.traced.point = gs.point[0];
SS.GW.ClearSelection();
} else {
Error(_("Bad selection for trace; select a single point."));
}
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::STOP_TRACING: {
Platform::FileDialogRef dialog = Platform::CreateSaveFileDialog(SS.GW.window);
dialog->AddFilters(Platform::CsvFileFilters);
dialog->ThawChoices(settings, "Trace");
if(dialog->RunModal()) {
dialog->FreezeChoices(settings, "Trace");
FILE *f = OpenFile(dialog->GetFilename(), "wb");
if(f) {
int i;
SContour *sc = &(SS.traced.path);
for(i = 0; i < sc->l.n; i++) {
Vector p = sc->l.elem[i].p;
double s = SS.exportScale;
fprintf(f, "%.10f, %.10f, %.10f\r\n",
p.x/s, p.y/s, p.z/s);
}
fclose(f);
} else {
Error(_("Couldn't write to '%s'"), dialog->GetFilename().raw.c_str());
}
}
// Clear the trace, and stop tracing
SS.traced.point = Entity::NO_ENTITY;
SS.traced.path.l.Clear();
SS.GW.Invalidate();
break;
}
default: ssassert(false, "Unexpected menu ID");
}
}
void SolveSpaceUI::ShowNakedEdges(bool reportOnlyWhenNotOkay) {
SS.nakedEdges.Clear();
Group *g = SK.GetGroup(SS.GW.activeGroup);
SMesh *m = &(g->displayMesh);
SKdNode *root = SKdNode::From(m);
bool inters, leaks;
root->MakeCertainEdgesInto(&(SS.nakedEdges),
EdgeKind::NAKED_OR_SELF_INTER, /*coplanarIsInter=*/true, &inters, &leaks);
if(reportOnlyWhenNotOkay && !inters && !leaks && SS.nakedEdges.l.n == 0) {
return;
}
SS.GW.Invalidate();
const char *intersMsg = inters ?
_("The mesh is self-intersecting (NOT okay, invalid).") :
_("The mesh is not self-intersecting (okay, valid).");
const char *leaksMsg = leaks ?
_("The mesh has naked edges (NOT okay, invalid).") :
_("The mesh is watertight (okay, valid).");
std::string cntMsg = ssprintf(
_("\n\nThe model contains %d triangles, from %d surfaces."),
g->displayMesh.l.n, g->runningShell.surface.n);
if(SS.nakedEdges.l.n == 0) {
Message(_("%s\n\n%s\n\nZero problematic edges, good.%s"),
intersMsg, leaksMsg, cntMsg.c_str());
} else {
Error(_("%s\n\n%s\n\n%d problematic edges, bad.%s"),
intersMsg, leaksMsg, SS.nakedEdges.l.n, cntMsg.c_str());
}
}
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 SolveSpaceUI::MenuHelp(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::WEBSITE:
Platform::OpenInBrowser("http://solvespace.com/helpmenu");
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::ABOUT:
Message(_(
Initial Autotools and FLTK support With this commit, SolveSpace gains an Autotools build system and a new platform-dependent backend implemented using the FLTK GUI toolkit. These will allow the application to be built and run on Linux and other Unix-like operating systems, and prospectively, MacOS X. A number of new files have been added: * Makefile.am: Automake makefile template; this contains some experimental support for MinGW and MSVC++ builds that needs further development * ac-aux/ax_fltk.m4: Autoconf M4 macro to locate and query the system's installation of FLTK; this will eventually be contributed to the GNU Autoconf Archive * autogen.sh: Script to bootstrap the Autotools build system, usually for a tree just checked out from source control * configure.ac: Source for the Autoconf configure script; note that this file specifies a version of 2.1, near the top * fltk/fltkmain.cpp: Main FLTK backend implementation * fltk/fltkutil.cpp: Utility functions for the FLTK backend * fltk/xFl_Gl_Window_Group.{H,cxx}: Implementation of a new Fl_Gl_Window_Group widget for FLTK, needed to facilitate drawing FLTK widgets on top of OpenGL graphics as SolveSpace does. This has been submitted to the FLTK project for (hopefully) eventual upstream inclusion: http://www.fltk.org/str.php?L2992 The following minor changes are also a part of this commit: * Makefile.msvc: Define PACKAGE_VERSION=2.1 for the benefit of solvespace.cpp in MSVC++ builds * solvespace.cpp: In the About dialog text, use PACKAGE_VERSION rather than hard-coding the version of the program * solvespace.h: Don't define the C99 integer types if HAVE_C99_INTEGER_TYPES is defined, to facilitate MinGW builds
2013-10-28 05:28:42 +00:00
"This is SolveSpace version " PACKAGE_VERSION ".\n"
"\n"
"For more information, see http://solvespace.com/\n"
"\n"
"SolveSpace is free software: you are free to modify\n"
"and/or redistribute it under the terms of the GNU\n"
"General Public License (GPL) version 3 or later.\n"
"\n"
"There is NO WARRANTY, to the extent permitted by\n"
"law. For details, visit http://gnu.org/licenses/\n"
"\n"
"© 2008-2016 Jonathan Westhues and other authors.\n"
));
break;
default: ssassert(false, "Unexpected menu ID");
}
}
void SolveSpaceUI::Clear() {
sys.Clear();
for(int i = 0; i < MAX_UNDO; i++) {
if(i < undo.cnt) undo.d[i].Clear();
if(i < redo.cnt) redo.d[i].Clear();
}
TW.window = NULL;
GW.openRecentMenu = NULL;
GW.linkRecentMenu = NULL;
GW.showGridMenuItem = NULL;
GW.perspectiveProjMenuItem = NULL;
GW.showToolbarMenuItem = NULL;
GW.showTextWndMenuItem = NULL;
GW.fullScreenMenuItem = NULL;
GW.unitsMmMenuItem = NULL;
GW.unitsMetersMenuItem = NULL;
GW.unitsInchesMenuItem = NULL;
GW.inWorkplaneMenuItem = NULL;
GW.in3dMenuItem = NULL;
GW.undoMenuItem = NULL;
GW.redoMenuItem = NULL;
GW.window = NULL;
}
void Sketch::Clear() {
group.Clear();
groupOrder.Clear();
constraint.Clear();
request.Clear();
style.Clear();
entity.Clear();
param.Clear();
}
BBox Sketch::CalculateEntityBBox(bool includingInvisible) {
2016-02-18 09:53:31 +00:00
BBox box = {};
bool first = true;
auto includePoint = [&](const Vector &point) {
if(first) {
box.minp = point;
box.maxp = point;
first = false;
} else {
box.Include(point);
}
};
for(const Entity &e : entity) {
if(e.construction) continue;
if(!(includingInvisible || e.IsVisible())) continue;
if(e.IsPoint()) {
includePoint(e.PointGetNum());
continue;
}
switch(e.type) {
// Circles and arcs are special cases. We calculate their bounds
// based on Bezier curve bounds. This is not exact for arcs,
// but the implementation is rather simple.
case Entity::Type::CIRCLE:
case Entity::Type::ARC_OF_CIRCLE: {
SBezierList sbl = {};
e.GenerateBezierCurves(&sbl);
for(const SBezier &sb : sbl.l) {
for(int j = 0; j <= sb.deg; j++) {
includePoint(sb.ctrl[j]);
}
}
sbl.Clear();
continue;
}
default:
continue;
}
}
return box;
}
Group *Sketch::GetRunningMeshGroupFor(hGroup h) {
Group *g = GetGroup(h);
while(g != NULL) {
if(g->IsMeshGroup()) {
return g;
}
g = g->PreviousGroup();
}
return NULL;
}