Private</a></li>
<li><a href="#hl_earlyexit">Use Early Exits and 'continue' to Simplify
Code</a></li>
+ <li><a href="#hl_else_after_return">Don't use "else" after a
+ return</a></li>
<li><a href="#hl_predicateloops">Turn Predicate Loops into Predicate
Functions</a></li>
</ol></li>
<li><a href="#ll_iostream"><tt>#include <iostream></tt> is
<em>forbidden</em></a></li>
<li><a href="#ll_avoidendl">Avoid <tt>std::endl</tt></a></li>
+ <li><a href="#ll_raw_ostream">Use <tt>raw_ostream</tt></a</li>
</ol></li>
<li><a href="#nano">Microscopic Details</a>
declare the symbol. This can lead to problems at link time.</p>
<p>So, the rule for LLVM is to always use the <tt>class</tt> keyword, unless
-<b>all</b> members are public, in which case <tt>struct</tt> is allowed.</p>
+<b>all</b> members are public and the type is a C++ "POD" type, in which case
+<tt>struct</tt> is allowed.</p>
</div>
</div>
+<!-- _______________________________________________________________________ -->
+<div class="doc_subsubsection">
+ <a name="hl_else_after_return">Don't use "else" after a return</a>
+</div>
+
+<div class="doc_text">
+
+<p>For similar reasons above (reduction of indentation and easier reading),
+ please do not use "else" or "else if" after something that interrupts
+ control flow like return, break, continue, goto, etc. For example, this is
+ "bad":</p>
+
+<div class="doc_code">
+<pre>
+ case 'J': {
+ if (Signed) {
+ Type = Context.getsigjmp_bufType();
+ if (Type.isNull()) {
+ Error = ASTContext::GE_Missing_sigjmp_buf;
+ return QualType();
+ } else {
+ break;
+ }
+ } else {
+ Type = Context.getjmp_bufType();
+ if (Type.isNull()) {
+ Error = ASTContext::GE_Missing_jmp_buf;
+ return QualType();
+ } else {
+ break;
+ }
+ }
+ }
+ }
+</pre>
+</div>
+
+<p>It is better to write this something like:</p>
+
+<div class="doc_code">
+<pre>
+ case 'J':
+ if (Signed) {
+ Type = Context.getsigjmp_bufType();
+ if (Type.isNull()) {
+ Error = ASTContext::GE_Missing_sigjmp_buf;
+ return QualType();
+ }
+ } else {
+ Type = Context.getjmp_bufType();
+ if (Type.isNull()) {
+ Error = ASTContext::GE_Missing_jmp_buf;
+ return QualType();
+ }
+ }
+ break;
+</pre>
+</div>
+
+<p>Or better yet (in this case), as:</p>
+
+<div class="doc_code">
+<pre>
+ case 'J':
+ if (Signed)
+ Type = Context.getsigjmp_bufType();
+ else
+ Type = Context.getjmp_bufType();
+
+ if (Type.isNull()) {
+ Error = Signed ? ASTContext::GE_Missing_sigjmp_buf :
+ ASTContext::GE_Missing_jmp_buf;
+ return QualType();
+ }
+ break;
+</pre>
+</div>
+
+<p>The idea is to reduce indentation and the amount of code you have to keep
+ track of when reading the code.</p>
+
+</div>
<!-- _______________________________________________________________________ -->
<div class="doc_subsubsection">
</ol>
<p>Note that using the other stream headers (<tt><sstream></tt> for
-example) is allowed normally, it is just <tt><iostream></tt> that is
-causing problems.</p>
+example) is not problematic in this regard (just <tt><iostream></tt>).
+However, raw_ostream provides various APIs that are better performing for almost
+every use than std::ostream style APIs, so you should just use it for new
+code.</p>
-<p>The preferred replacement for stream functionality is the
-<tt>llvm::raw_ostream</tt> class (for writing to output streams of various
-sorts) and the <tt>llvm::MemoryBuffer</tt> API (for reading in files).</p>
+<p><b>New code should always
+use <a href="#ll_raw_ostream"><tt>raw_ostream</tt></a> for writing, or
+the <tt>llvm::MemoryBuffer</tt> API for reading files.</b></p>
</div>
</div>
+<!-- _______________________________________________________________________ -->
+<div class="doc_subsubsection">
+ <a name="ll_raw_ostream">Use <tt>raw_ostream</tt></a>
+</div>
+
+<div class="doc_text">
+
+<p>LLVM includes a lightweight, simple, and efficient stream implementation
+in <tt>llvm/Support/raw_ostream.h</tt> which provides all of the common features
+of <tt>std::ostream</tt>. All new code should use <tt>raw_ostream</tt> instead
+of <tt>ostream</tt>.</p>
+
+<p>Unlike <tt>std::ostream</tt>, <tt>raw_ostream</tt> is not a template and can
+be forward declared as <tt>class raw_ostream</tt>. Public headers should
+generally not include the <tt>raw_ostream</tt> header, but use forward
+declarations and constant references to <tt>raw_ostream</tt> instances.</p>
+
+</div>
+
+
<!-- ======================================================================= -->
<div class="doc_subsection">
<a name="nano">Microscopic Details</a>