[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [PATCH] GCC seriously needs to be less picky.


On 01/25/2011 10:07 PM, Chris Lumens wrote:
Don't tell me I have to catch the return value of write in a variable just
to later complain that I'm not doing anything with that variable.  That's
borderline sociopathic behavior.

Erm, although I'm no longer part of the team I cannot help but respond here.
The warnings given by gcc-4.6 for unused-but-set-variable are very valid
and very useful. We hit a few in spice to and fixed them rather then just
disabling the warning, removing quite a bit of dead code.

As for your write case, the solution there of course is to do proper
error checking, ie as a minimum do:

if (write(...) == -1)

This way in the unlikely event that a write would fail despite you
expecting it never to fail, you at least have some chance of seeing
an error message pointing to the issue.



  configure.ac |    6 +++++-
  1 files changed, 5 insertions(+), 1 deletions(-)

diff --git a/configure.ac b/configure.ac
index 6e677b5..ab37f4a 100644
--- a/configure.ac
+++ b/configure.ac
@@ -185,8 +185,12 @@ if test x$ipv6 = xyes ; then

+# GCC likes to bomb out on some ridiculous warnings.  Add your favorites
+# here.
  # Add remaining compiler flags we want to use

  # Filter CFLAGS (remove duplicate flags)
  cflags_filter() {

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]