Matthew Woehlke wrote:
> Jim Meyering wrote:
>> It's saying that with its default settings, the post-cpp-expansion of
>> this code is too large:
>> [snip]
>
> I know what the problem is; as I was saying, I didn't find a way to
> get past this type of error except "don't do that", i.e. break down
> what cpp is seeing into manageable bites.
>
> How about the attached work-around (obviously '#if 1' should be
> changed to something more like '#if using-the-hp-ux-cc-compiler')?
>
> (Build succeeds with this, using HP-UX's cc. I'll post the test
> results when the suite is done.)
How about this instead?
>From 122169078df3a1dfa1e083c2e1bfdac83ae28c52 Mon Sep 17 00:00:00 2001
From: Jim Meyering <meyer...@redhat.com>
Date: Thu, 19 Mar 2009 20:14:26 +0100
Subject: [PATCH] dd: use a more portable definition of O_FULLBLOCK
* src/dd.c (O_FULLBLOCK): Compute its value via two separate
expressions to avoid a compiler limitation.
Reported by Matthew Woehlke.
---
src/dd.c | 33 +++++++++++++++++++--------------
1 files changed, 19 insertions(+), 14 deletions(-)
diff --git a/src/dd.c b/src/dd.c
index 9a1c875..8a814f7 100644
--- a/src/dd.c
+++ b/src/dd.c
@@ -265,20 +265,25 @@ static struct symbol_value const conversions[] =
enum
{
- /* Use a value that is larger than that of any other O_ symbol. */
- O_FULLBLOCK = ((MAX (O_APPEND,
- MAX (O_BINARY,
- MAX (O_CIO,
- MAX (O_DIRECT,
- MAX (O_DIRECTORY,
- MAX (O_DSYNC,
- MAX (O_NOATIME,
- MAX (O_NOCTTY,
- MAX (O_NOFOLLOW,
- MAX (O_NOLINKS,
- MAX (O_NONBLOCK,
- MAX (O_SYNC,
- MAX (O_TEXT, 0)))))))))))))) << 1)
+ /* Use a value that is larger than that of any other O_ symbol.
+ Do it in two steps to avoid hitting a cpp expansion size limit
+ in the vendor compiler from HP-UX 11.23. */
+ M1 = (MAX (O_APPEND,
+ MAX (O_BINARY,
+ MAX (O_CIO,
+ MAX (O_DIRECT,
+ MAX (O_DIRECTORY,
+ MAX (O_DSYNC, 0))))))),
+
+ M2 = (MAX (O_NOATIME,
+ MAX (O_NOCTTY,
+ MAX (O_NOFOLLOW,
+ MAX (O_NOLINKS,
+ MAX (O_NONBLOCK,
+ MAX (O_SYNC,
+ MAX (O_TEXT, 0)))))))),
+
+ O_FULLBLOCK = (MAX (M1, M2) << 1)
};
/* Ensure that we didn't shift it off the end. */
--
1.6.2.rc1.285.gc5f54
_______________________________________________
Bug-coreutils mailing list
Bug-coreutils@gnu.org
http://lists.gnu.org/mailman/listinfo/bug-coreutils
