<div dir="ltr">You're right, it is now fixed.</div><div class="gmail_extra"><br><div class="gmail_quote">On Mon, Aug 17, 2015 at 3:02 AM, Benjamin Kramer <span dir="ltr"><<a href="mailto:benny.kra@gmail.com" target="_blank">benny.kra@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class=""><br>
> On 16.08.2015, at 16:53, Dylan McKay <<a href="mailto:dylanmckay34@gmail.com">dylanmckay34@gmail.com</a>> wrote:<br>
><br>
> I fixed everything you mentioned Benjamin - thanks for the review!<br>
><br>
> Attached is the new patch.<br>
<br>
</span>You don't need the size_t Count argument anymore, the MutableArrayRef already carries the size.<br>
<br>
- Ben<br>
<div><div class="h5"><br>
><br>
> On Sun, Aug 16, 2015 at 10:57 PM, Benjamin Kramer <<a href="mailto:benny.kra@gmail.com">benny.kra@gmail.com</a>> wrote:<br>
><br>
> > On 15.08.2015, at 16:23, Dylan McKay via llvm-commits <<a href="mailto:llvm-commits@lists.llvm.org">llvm-commits@lists.llvm.org</a>> wrote:<br>
> ><br>
> > From b7094fa637ffd59b10776e0a4bbf7f8c04235306 Mon Sep 17 00:00:00 2001<br>
> > From: Dylan McKay <<a href="mailto:dylanmckay34@gmail.com">dylanmckay34@gmail.com</a>><br>
> > Date: Sun, 16 Aug 2015 02:09:05 +1200<br>
> > Subject: [PATCH] Extend MCAsmLexer so that it can peek forward several tokens<br>
> ><br>
> > This commit adds a virtual `peekTokens()` function to `MCAsmLexer`<br>
> > which can peek forward an arbitrary number of tokens.<br>
> ><br>
> > It also makes the `peekTok()` method call `peekTokens()` method, but<br>
> > only requesting one token.<br>
> ><br>
> > The idea is to better support targets which more more ambiguous<br>
> > assembly syntaxes.<br>
> > ---<br>
> > include/llvm/MC/MCParser/AsmLexer.h | 3 ++-<br>
> > include/llvm/MC/MCParser/MCAsmLexer.h | 13 ++++++++++++-<br>
> > lib/MC/MCParser/AsmLexer.cpp | 16 +++++++++++++---<br>
> > 3 files changed, 27 insertions(+), 5 deletions(-)<br>
> ><br>
> > diff --git a/include/llvm/MC/MCParser/AsmLexer.h b/include/llvm/MC/MCParser/AsmLexer.h<br>
> > index 62d39b2..77023e4 100644<br>
> > --- a/include/llvm/MC/MCParser/AsmLexer.h<br>
> > +++ b/include/llvm/MC/MCParser/AsmLexer.h<br>
> > @@ -47,7 +47,8 @@ public:<br>
> > StringRef LexUntilEndOfStatement() override;<br>
> > StringRef LexUntilEndOfLine();<br>
> ><br>
> > - const AsmToken peekTok(bool ShouldSkipSpace = true) override;<br>
> > + size_t peekTokens(AsmToken *Buf, size_t Size,<br>
> > + bool ShouldSkipSpace = true) override;<br>
><br>
> This could be a MutableArrayRef<AsmToken> instead of a pointer+size pair.<br>
><br>
> ><br>
> > bool isAtStartOfComment(const char *Ptr);<br>
> > bool isAtStatementSeparator(const char *Ptr);<br>
> > diff --git a/include/llvm/MC/MCParser/MCAsmLexer.h b/include/llvm/MC/MCParser/MCAsmLexer.h<br>
> > index 71f15b3..fb9d32d 100644<br>
> > --- a/include/llvm/MC/MCParser/MCAsmLexer.h<br>
> > +++ b/include/llvm/MC/MCParser/MCAsmLexer.h<br>
> > @@ -162,7 +162,18 @@ public:<br>
> > }<br>
> ><br>
> > /// Look ahead at the next token to be lexed.<br>
> > - virtual const AsmToken peekTok(bool ShouldSkipSpace = true) = 0;<br>
> > + const AsmToken peekTok(bool ShouldSkipSpace = true) {<br>
> > + AsmToken Buf;<br>
> > +<br>
> > + size_t ReadCount = peekTokens(&Buf, 1, ShouldSkipSpace);<br>
> > + assert(ReadCount == 1);<br>
><br>
> Add (void)ReadCount; here to avoid unused variable warnings when assertions are disabled.<br>
><br>
> > +<br>
> > + return Buf;<br>
> > + }<br>
> > +<br>
> > + /// Look ahead an arbitrary number of tokens.<br>
> > + virtual size_t peekTokens(AsmToken *Buf, size_t Size,<br>
> > + bool ShouldSkipSpace = true) = 0;<br>
> ><br>
> > /// Get the current error location<br>
> > const SMLoc &getErrLoc() {<br>
> > diff --git a/lib/MC/MCParser/AsmLexer.cpp b/lib/MC/MCParser/AsmLexer.cpp<br>
> > index b983d99..fcedb42 100644<br>
> > --- a/lib/MC/MCParser/AsmLexer.cpp<br>
> > +++ b/lib/MC/MCParser/AsmLexer.cpp<br>
> > @@ -436,7 +436,8 @@ StringRef AsmLexer::LexUntilEndOfLine() {<br>
> > return StringRef(TokStart, CurPtr-TokStart);<br>
> > }<br>
> ><br>
> > -const AsmToken AsmLexer::peekTok(bool ShouldSkipSpace) {<br>
> > +size_t AsmLexer::peekTokens(AsmToken *Buf, size_t Size,<br>
> > + bool ShouldSkipSpace) {<br>
> > const char *SavedTokStart = TokStart;<br>
> > const char *SavedCurPtr = CurPtr;<br>
> > bool SavedAtStartOfLine = isAtStartOfLine;<br>
> > @@ -446,8 +447,17 @@ const AsmToken AsmLexer::peekTok(bool ShouldSkipSpace) {<br>
> > SMLoc SavedErrLoc = getErrLoc();<br>
> ><br>
> > SkipSpace = ShouldSkipSpace;<br>
> > - AsmToken Token = LexToken();<br>
> ><br>
> > + size_t ReadCount;<br>
> > + for(ReadCount = 0; ReadCount<Size; ++ReadCount) {<br>
> > + AsmToken Token = LexToken();<br>
> > +<br>
> > + Buf[ReadCount] = Token;<br>
> > +<br>
> > + if(Token.is(AsmToken::Eof))<br>
> > + break;<br>
> > +<br>
> > + }<br>
><br>
> This hunk has funky identation. clang-format?<br>
><br>
> Patch looks good when my comments are fixed.<br>
><br>
> - Ben<br>
><br>
> > SetError(SavedErrLoc, SavedErr);<br>
> ><br>
> > SkipSpace = SavedSkipSpace;<br>
> > @@ -455,7 +465,7 @@ const AsmToken AsmLexer::peekTok(bool ShouldSkipSpace) {<br>
> > CurPtr = SavedCurPtr;<br>
> > TokStart = SavedTokStart;<br>
> ><br>
> > - return Token;<br>
> > + return ReadCount;<br>
> > }<br>
> ><br>
> > bool AsmLexer::isAtStartOfComment(const char *Ptr) {<br>
> > --<br>
> > 2.4.4<br>
> ><br>
><br>
><br>
</div></div>> <0001-Extend-MCAsmLexer-so-that-it-can-peek-forward-severa.patch><br>
<br>
</blockquote></div><br></div>