From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Subject: Re: [PATCH] DTC: Add support for a C-like #include "file" mechanism. From: Jon Loeliger To: David Gibson In-Reply-To: <20070329020744.GB16270@localhost.localdomain> References: <1174681120.32390.82.camel@ld0161-tx32> <20070328062159.GJ27401@localhost.localdomain> <1175102216.32390.220.camel@ld0161-tx32> <20070329020744.GB16270@localhost.localdomain> Content-Type: text/plain Message-Id: <1175184168.32390.275.camel@ld0161-tx32> Mime-Version: 1.0 Date: Thu, 29 Mar 2007 11:02:48 -0500 Cc: "linuxppc-dev@ozlabs.org" List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Wed, 2007-03-28 at 21:07, David Gibson wrote: > > > > +\"[^"\n]*\" { > > > > + yytext[strlen(yytext) - 1] = 0; > > > > + if (!push_input_file(yytext + 1)) { > > > > + /* Some unrecoverable error.*/ > > > > > > Should be some kind of error message here. > > > > Hrm. Yeah. So I looked, and all the return 0 cases issue specific errors before returning, so that happens already. > > Well, it has to come early so that the type is seen > > prior to the builtin version. But it could be reorganized > > a bit, I suppose. I was just angling on not cluttering > > up the lexer proper. > > Err... builtin version of what? I forgot I already separated out the YYLTYPE definition to handle this issue. Never mind; it's kosher already. > > Note some people are seeing some compilation problems > > around the struct field "filenum" that I introduced. > > I don't see that problem, but a coworker here does, > > so I am now able to reproduce the issue and am working > > on a fix. Sorry. > > Hrm. Which struct did you add the field to? And the patch I applied yesterday fixed this YYLTYPE up too. jdl