<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Tahoma
}
--></style>
</head>
<body class='hmmessage'>
diff attached, more semantically significant than usual (as opposed to "structural" changes<br>that try not to change behavior, this actually doesn't really mean to change behavior either,<br>but it might)<br><br><br>> Date: Fri, 19 Nov 2010 12:48:28 +0000<br>> To: m3commit@elegosoft.com<br>> From: jkrell@elego.de<br>> Subject: [M3commit] CVS Update: cm3<br>> <br>> CVSROOT: /usr/cvs<br>> Changes by: jkrell@birch. 10/11/19 12:48:28<br>> <br>> Modified files:<br>> cm3/m3-sys/m3cc/gcc/gcc/m3cg/: parse.c m3-parse.h <br>> <br>> Log message:<br>> add an extra pass so that declare_segment can get to the data<br>> that is in bind_segment, so that the size/alignment of segments<br>> is always correct; this fixes configure -enable-checking<br>> This probably also addresses the recurring problem I have had<br>> where loads in segments were broken because the segment was<br>> smaller than the load offset, e.g. on OpenBSD/mips64, and<br>> much was broken. I had been working around by blowing up<br>> the initial segment size guess. Now we get it right up front.<br>> <br>> More work is needed here for the typing/debugging of segments (aka globals).<br>> Once we turn types back on! (once they don't break things)<br>> <br>> add a lot of casts to satisfy configure -enable-checking<br>> This could use careful review.<br>> <br>> configure -enable-checking all good now *except* for static link problem,<br>> and as such shall be enabled shortly unconditionally, at least<br>> the checks that don't affect .h files (except static link problem)<br>> (that's an extra level of checking I haven't tested recently, will soon,<br>> but might be too slow; as well, we can/should dial this back for release)<br>> <br> </body>
</html>