Make it easier to link to READMEs by removing the level name.

After make install they will still be installed to a level0, 1, 2 sub-folder,
as the cmake instructions are generated by generate.py

BUG: 394237

15 files changed, 0 insertions, 199 deletions

diff --git a/src/checks/level2/README-base-class-event.md b/src/checks/level2/README-base-class-event.md
deleted file mode 100644
index fc61fca8..00000000
--- a/src/checks/level2/README-base-class-event.md
+++ /dev/null
@@ -1,4 +0,0 @@
-# base-class-event
-
-Warns when you `return false` inside your `QObject::event()` or `QObject::eventFilter()` reimplementation.
-Instead you should probably call the base class method, so the event is correctly handled.
diff --git a/src/checks/level2/README-copyable-polymorphic.md b/src/checks/level2/README-copyable-polymorphic.md
deleted file mode 100644
index a41ead02..00000000
--- a/src/checks/level2/README-copyable-polymorphic.md
+++ /dev/null
@@ -1,8 +0,0 @@
-# copyable-polymorphic
-
-Finds polymorphic classes that are copyable.
-These classes are usually vulnerable to slicing [1].
-
-To fix these warnings use Q_DISABLE_COPY or delete the copy-ctor yourself.
-
-[1] <https://en.wikipedia.org/wiki/Object_slicing>
diff --git a/src/checks/level2/README-ctor-missing-parent-argument.md b/src/checks/level2/README-ctor-missing-parent-argument.md
deleted file mode 100644
index 90b9b3ab..00000000
--- a/src/checks/level2/README-ctor-missing-parent-argument.md
+++ /dev/null
@@ -1,8 +0,0 @@
-# ctor-missing-parent-argument
-
-Warns when `QObject` derived classes don't have at least one CTOR receiving a QObject.
-
-This is an attempt to catch classes which are missing the parent argument.
-It doesn't have false-positives, but might miss true-positives. For example,
-if you have a CTOR which receives a `QObject* but then forget to pass it to the
-base CTOR.
diff --git a/src/checks/level2/README-function-args-by-ref.md b/src/checks/level2/README-function-args-by-ref.md
deleted file mode 100644
index d755bfbc..00000000
--- a/src/checks/level2/README-function-args-by-ref.md
+++ /dev/null
@@ -1,11 +0,0 @@
-# function-args-by-ref
-
-Warns when you should be passing by const-ref.
-Types with sizeof > 16 bytes [1] or types which are not trivially-copyable [2] or not trivially-destructible [3] should be passed by ref. A rule of thumb is that if passing by value would trigger copy-ctor and/or dtor then pass by ref instead.
-
-This check will ignore shared pointers, you're on your own. Most of the times passing shared pointers by const-ref is the best thing to do, but occasionally that will lead to crashes if you're in a method that calls something else that makes the shared pointer ref count go down to zero.
-
-
-- [1] <http://www.macieira.org/blog/2012/02/the-value-of-passing-by-value/>
-- [2] <http://en.cppreference.com/w/cpp/concept/TriviallyCopyable>
-- [3] <http://www.cplusplus.com/reference/type_traits/is_trivially_destructible/>
diff --git a/src/checks/level2/README-function-args-by-value.md b/src/checks/level2/README-function-args-by-value.md
deleted file mode 100644
index 1e96e4ae..00000000
--- a/src/checks/level2/README-function-args-by-value.md
+++ /dev/null
@@ -1,10 +0,0 @@
-# function-args-by-value
-
-Warns when you should be passing by value instead of by-ref.
-Types with sizeof <= 16 bytes [1] which are trivially-copyable [2] and trivially-destructible [3] should be passed by value.
-
-Only fix these warnings if you're sure that the value would be passed in a CPU register instead on the stack.
-
-- [1] <http://www.macieira.org/blog/2012/02/the-value-of-passing-by-value/>
-- [2] <http://en.cppreference.com/w/cpp/concept/TriviallyCopyable>
-- [3] <http://www.cplusplus.com/reference/type_traits/is_trivially_destructible/>
diff --git a/src/checks/level2/README-global-const-char-pointer.md b/src/checks/level2/README-global-const-char-pointer.md
deleted file mode 100644
index 42cfa996..00000000
--- a/src/checks/level2/README-global-const-char-pointer.md
+++ /dev/null
@@ -1,4 +0,0 @@
-# global-const-char-pointer
-
-Finds where you're using `const char *foo` instead of `const char *const foo` or `const char []foo`.
-The former case adds a pointer in .data, pointing to .rodata. The later cases only use .rodata.
diff --git a/src/checks/level2/README-implicit-casts.md b/src/checks/level2/README-implicit-casts.md
deleted file mode 100644
index 8a5f3b0f..00000000
--- a/src/checks/level2/README-implicit-casts.md
+++ /dev/null
@@ -1,19 +0,0 @@
-# implicit-casts
-
-Finds places with unwanted implicit casts in function calls.
-
-#### Supported cases
-
-* pointer->bool cast in functions accepting bool and pointers, example:
-
-        MyWidget(bool b, QObject *parent = nullptr) {}
-        MyWidget(parent);
-
-* bool->int
-
-        void func(int duration);
-        func(someBool);
-
-This last case is disabled due to false positives when calling C code.
-You can enable it by with:
-`export CLAZY_EXTRA_OPTIONS=implicit-casts-bool-to-int`
diff --git a/src/checks/level2/README-missing-qobject-macro.md b/src/checks/level2/README-missing-qobject-macro.md
deleted file mode 100644
index 4b56013a..00000000
--- a/src/checks/level2/README-missing-qobject-macro.md
+++ /dev/null
@@ -1,17 +0,0 @@
-# missing-qobject-macro
-
-Finds `QObject` derived classes that don't have a Q_OBJECT macro.
-
-#### Reasons to use Q_OBJECT
-- Signals and slots
-- `QObject::inherits`
-- `qobject_cast`
-- `metaObject()->className()`
-- Use your custom widget as a selector in Qt stylesheets
-
-#### Reasons not to use Q_OBJECT
-- Templated QObjects
-- Compilation time
-
-This check can't be used with pre-compiled headers support.
-This check doesn't have false positives, but it's not included in level <= 1 because the missing Q_OBJECT might be intentional.
diff --git a/src/checks/level2/README-missing-typeinfo.md b/src/checks/level2/README-missing-typeinfo.md
deleted file mode 100644
index 675fc2f0..00000000
--- a/src/checks/level2/README-missing-typeinfo.md
+++ /dev/null
@@ -1,6 +0,0 @@
-# missing-typeinfo
-
-Suggests usage of `Q_PRIMITIVE_TYPE` or `Q_MOVABLE_TYPE` in cases where you're using `QList<T>` and `sizeof(T) > sizeof(void*)`
-or using `QVector<T>`, unless they already have a type info classification.
-
-See `Q_DECLARE_TYPEINFO` in Qt documentation for more information.
diff --git a/src/checks/level2/README-old-style-connect.md b/src/checks/level2/README-old-style-connect.md
deleted file mode 100644
index 2cad4e8d..00000000
--- a/src/checks/level2/README-old-style-connect.md
+++ /dev/null
@@ -1,18 +0,0 @@
-# old-style-connect
-
-Finds usages of old style connects.
-Connecting with old style syntax (`SIGNAL`/`SLOT`) is much slower than using pointer to member syntax (PMF).
-
-Here's however a non-exhaustive list of caveats you should be aware of:
-- You can't disconnect with new-syntax if the connect was made with old-syntax (and vice-versa)
-- You can't disconnect from a static slot with new-syntax (although connecting works)
-- Difference in behaviour when calling slots of partially destroyed objects (<https://codereview.qt-project.org/#/c/83800>)
-
-#### Fixits
-
-You can convert the most simple cases with `export CLAZY_FIXIT=fix-old-style-connect`.
-Be careful, as PMF is not a 100% drop-in replacement.
-
-#### Pitfalls
-
-Although this check doesn't have false-positives it's a level2 check, that's because some connects are tricky to convert to PMF syntax and might introduce bugs if you don't know what you're doing.
diff --git a/src/checks/level2/README-qstring-allocations.md b/src/checks/level2/README-qstring-allocations.md
deleted file mode 100644
index fda4278e..00000000
--- a/src/checks/level2/README-qstring-allocations.md
+++ /dev/null
@@ -1,49 +0,0 @@
-# qstring-unneeded-heap-allocations
-
-Finds places with unneeded memory allocations due to temporary `QString`s.
-
-Here's a summary of usages that allocate:
-
-1. `QString s = "foo"; // Allocates, use QStringLiteral("foo") instead`
-
-2. `QString s = QLatin1String("foo"); // Allocates, use QStringLiteral("foo") instead`
-
-    2.1 `QString s = QLatin1String(""); // No allocation. QString is optimized for this case, so it's safe for empty literals`
-
-3. `QString s = QStringLiteral("foo"); // No allocation`
-
-4. `QString s = QString::fromLatin1("foo"); // Allocates, use QStringLiteral`
-
-5. `QString s = QString::fromUtf8("foo"); // Allocates, use QStringLiteral`
-
-6. `s == "foo" // Allocates, use QLatin1String`
-
-7. `s == QLatin1String("foo) // No allocation`
-
-8. `s == QStringLiteral("foo") // No allocation`
-
-9. `QString {"append", "compare", "endsWith", "startsWith", "indexOf", "insert",`
-   `         "lastIndexOf", "prepend", "replace", "contains" } // They all have QLatin1String overloads, so passing a QLatin1String is ok.`
-
-10. `QString::fromLatin1("foo %1").arg(bar) // Allocates twice, replacing with QStringLiteral makes it allocate only once.`
-
-
-#### Fixits
-
-    fix-qlatin1string-allocations        // To replace QLatin1String with QStringLiteral only where it was allocating before
-    fix-fromLatin1_fromUtf8-allocations  // To replace fromLatin1() and fromUtf8() so it doesn't allocate
-    fix-fromCharPtrAllocations           // To replace raw string literals so it doesn't allocate
-
-    Example:
-        export CLAZY_FIXIT="fix-fromCharPtrAllocations"
-
-#### Pitfalls
-
-- `QStringLiteral` might make your app crash at exit if plugins are involved.
-See:
-<https://blogs.kde.org/2015/11/05/qregexp-qstringliteral-crash-exit> and
-<http://lists.qt-project.org/pipermail/development/2015-November/023681.html>
-
-- Also note that MSVC crashes when `QStringLiteral` is used inside initializer lists. For that reason no warning or fixit is emitted for this case unless you set an env variable:
-
-        export CLAZY_EXTRA_OPTIONS="qstring-allocations-no-msvc-compat"
diff --git a/src/checks/level2/README-returning-void-expression.md b/src/checks/level2/README-returning-void-expression.md
deleted file mode 100644
index bb405092..00000000
--- a/src/checks/level2/README-returning-void-expression.md
+++ /dev/null
@@ -1,14 +0,0 @@
-# returning-void-expression
-
-Warns when returning a void expression.
-
-Example:
-```
-void doStuff()
-{
-    if (cond)
-        return // Oops, forgot the ; but it still compiles since processStuff() returns void.
-
-    processStuff();
-}
-```
diff --git a/src/checks/level2/README-rule-of-three.md b/src/checks/level2/README-rule-of-three.md
deleted file mode 100644
index 9e9f64ea..00000000
--- a/src/checks/level2/README-rule-of-three.md
+++ /dev/null
@@ -1,13 +0,0 @@
-# rule-of-three
-
-Implements the rule of three:
-<https://en.wikipedia.org/wiki/Rule_of_three_%28C%2B%2B_programming%29>
-
-#### Exceptions
-To reduce the amount of warnings, these cases won't emit warnings:
-- class has a QSharedDataPointer member
-- class inherits from QSharedData
-- if only the dtor is implemented and it's protected
-- class name ends with "Private" and is defined in a .cpp, .cxx or _p.h file
-
-In some cases you're missing methods, in others you have too many methods. You'll have to judge what's the correct fix and beware of binary compatibility.
diff --git a/src/checks/level2/README-static-pmf.md b/src/checks/level2/README-static-pmf.md
deleted file mode 100644
index e27e7b6c..00000000
--- a/src/checks/level2/README-static-pmf.md
+++ /dev/null
@@ -1,9 +0,0 @@
-# static-pmf
-
-Warns when storing a pointer to QObject member function into a static variable.
-Passing such variable to a connect is known to fail when using MingW.
-
-Example:
-
-static auto pmf = &QObject::destroyed;
-QCOMPARE(pmf, &QObject::destroyed); // fails
diff --git a/src/checks/level2/README-virtual-call-ctor.md b/src/checks/level2/README-virtual-call-ctor.md
deleted file mode 100644
index 2bf715c9..00000000
--- a/src/checks/level2/README-virtual-call-ctor.md
+++ /dev/null
@@ -1,9 +0,0 @@
-# virtual-call-ctor
-
-Finds places where you're calling pure virtual functions inside a constructor or destructor.
-Compilers usually warn about this if there isn't any indirection, this check will catch cases like calling
-a non-pure virtual that calls a pure virtual.
-
-
-This check only looks for pure virtuals, ignoring non-pure, which in theory you shouldn't call,
-but seems common practice.