Stephen Hahn <[EMAIL PROTECTED]> wrote:
> > This reminds me of to ask when code review will be possible by people
> > from outsude Sun.
>
> Can you be more specific about what kind of code review actions you
> mean? For instance, wearing my RTI advocate hat, I would probably
> accept RTI s
Joerg Schilling writes:
> > That would mean "the strchr function takes a pointer to character
> > argument, it does not modify the storage pointed to, and the value it
> > returns has the same storage qualifiers as the first argument."
>
> If the #pragma only affects the storage attributes (as I i
James Carlson <[EMAIL PROTECTED]> wrote:
> > >typeof(arg1)
> > >strchr(const char *, char)
> > >
> > >or something similar that would carry the actual input over to the return
> > >value.
> >
> >
> > It would soon give you C++ type overloading. Might not be bad, but you
> > know where to find
[EMAIL PROTECTED] writes:
> >typeof(arg1)
> >strchr(const char *, char)
> >
> >or something similar that would carry the actual input over to the return
> >value.
>
>
> It would soon give you C++ type overloading. Might not be bad, but you
> know where to find C++ :-)
A good subset of the func
>The question would be if the C standard commitee could define something like
>this:
>
>typeof(arg1)
>strchr(const char *, char)
>
>or something similar that would carry the actual input over to the return
>value.
It would soon give you C++ type overloading. Might not be bad, but you
know whe
[EMAIL PROTECTED] wrote:
> The C standards folks aren't that stupid. The function should not
> have their return values const qualified because in the majority
> of cases:
> - the argument isn't const
> - the return value is going to be modified.
>
> The const qualifier for the argume
>Yes, there are a number of standard library and system functions which
>could or should have their return types const-qualified but don't.
>Historical and conformance reasons, of course, but it makes
>-Wwrite-strings and similar options rather useless.
The C standards folks aren't that stupid.
On Mon 20 Mar 2006 at 03:24PM, Stephen Lau wrote:
> Joerg Schilling wrote:
> >> "The list of code reviewers should be individuals who have
> >> actually reviewed the code, not an alias for a group of people who
> >> were given the opportunity to do so."
> >
> >
> >This reminds me of to ask wh
Joerg Schilling wrote:
"The list of code reviewers should be individuals who have
actually reviewed the code, not an alias for a group of people who
were given the opportunity to do so."
This reminds me of to ask when code review will be possible by people
from outsude Sun.
Dan Pri
* Joerg Schilling <[EMAIL PROTECTED]> [2006-03-20 13:36]:
> Keith M Wesolowski <[EMAIL PROTECTED]> wrote:
> > "The list of code reviewers should be individuals who have
> > actually reviewed the code, not an alias for a group of people who
> > were given the opportunity to do so."
>
>
On Mon, Mar 20, 2006 at 10:36:23PM +0100, Joerg Schilling wrote:
> If you really like to use const char *, then it may needed to change thousands
> of lines. And even then you would not find:
Yes, I understand. I wasn't suggesting to change the code but rather
to examine it. In most cases, larg
Keith M Wesolowski <[EMAIL PROTECTED]> wrote:
> No one said you have to manually test every possible input to every
> possible program. In many cases you may be able to use semi-automated
> means to rule out a need for any manual testing at all (for example,
> if you can show that every reference
Keith M Wesolowski <[EMAIL PROTECTED]> wrote:
> bugs of this type. I believe the printing system (cmd/lp) in
> particular is actually using -fwritable-strings in a few places still,
> so it would definitely break with -xstrconst.
Is there no way to modify the related code to work with -xstrconst
On 3/18/06, Keith M Wesolowski <[EMAIL PROTECTED]> wrote:
> On Sat, Mar 18, 2006 at 09:31:27PM +0100, Roland Mainz wrote:
>
> > Ouch... every single binary with all possible combinations of input ?
> > That's almost impossible for a single person .. ;-(
>
> No one said you have to manually test eve
[EMAIL PROTECTED] wrote:
> >Are there any bug reports in the gcc bugzilla for that ?
>
> I can't tell what versions exhibited these issues and whther these
> were fixed; they were certainly well known at some point.
I recall some trouble in the gcc2.9x.x days... but do the problems apply
to gcc4.
On Sat, Mar 18, 2006 at 09:31:27PM +0100, Roland Mainz wrote:
> Ouch... every single binary with all possible combinations of input ?
> That's almost impossible for a single person .. ;-(
No one said you have to manually test every possible input to every
possible program. In many cases you may
>Are there any bug reports in the gcc bugzilla for that ?
I can't tell what versions exhibited these issues and whther these
were fixed; they were certainly well known at some point.
>> Doesn't gcc -fwritable-strings work?
>
>Erm... I think this would be the wrong way. C/C++ string literals are
Keith M Wesolowski wrote:
> On Sat, Mar 18, 2006 at 08:35:51PM +0100, Roland Mainz wrote:
> > Are there any objections that I make a patch which simply adds
> > "-xstrconst" to the tree-wide compile options ?
>
> Not at all; I'd welcome it, provided - and this is a big catch - that
> you can demon
On Sat, Mar 18, 2006 at 08:35:51PM +0100, Roland Mainz wrote:
> Are there any objections that I make a patch which simply adds
> "-xstrconst" to the tree-wide compile options ?
Not at all; I'd welcome it, provided - and this is a big catch - that
you can demonstrate that you've achieved the neces
Keith M Wesolowski wrote:
> On Sat, Mar 18, 2006 at 09:29:35AM +0100, [EMAIL PROTECTED] wrote:
>
> > Doesn't gcc -fwritable-strings work?
>
> Yes, in gcc 3. However it issues a warning about deprecation, and in
> gcc 4 the option is gone entirely.
Are there any objections that I make a patch wh
[EMAIL PROTECTED] wrote:
>
> >> Note that when building ON with gcc, we use -fno-builtin.
> >
> >Why ? I understand this is usefull for debugging - but it kills some
> >interesting opportunities for optimisation... ;-(
>
> Gcc builtins were often buggy (expected alignment were none was
> guarante
On Sat, Mar 18, 2006 at 09:29:35AM +0100, [EMAIL PROTECTED] wrote:
> Doesn't gcc -fwritable-strings work?
Yes, in gcc 3. However it issues a warning about deprecation, and in
gcc 4 the option is gone entirely.
--
Keith M Wesolowski "Sir, we're surrounded!"
Solaris Kernel Team
>> Note that when building ON with gcc, we use -fno-builtin.
>
>Why ? I understand this is usefull for debugging - but it kills some
>interesting opportunities for optimisation... ;-(
Gcc builtins were often buggy (expected alignment were none was
guaranteed) or called functions which don't exis
On Sat, Mar 18, 2006 at 12:54:36AM +0100, Roland Mainz wrote:
> Is there no way to use gcc's builtin functions except those which should
> be handled by the platform-optimized libc ? Using -fno-builtin is like
> hunting ducks with a Abrahams tank...
Yes, you can say -fno-builtin-. Unfortunately
Keith M Wesolowski wrote:
> On Fri, Mar 17, 2006 at 11:53:30PM +0100, Roland Mainz wrote:
> > > Note that when building ON with gcc, we use -fno-builtin.
> >
> > Why ? I understand this is usefull for debugging - but it kills some
> > interesting opportunities for optimisation... ;-(
>
> While I'
On Fri, Mar 17, 2006 at 11:53:30PM +0100, Roland Mainz wrote:
> > Note that when building ON with gcc, we use -fno-builtin.
>
> Why ? I understand this is usefull for debugging - but it kills some
> interesting opportunities for optimisation... ;-(
While I'm not the one who implemented this, I'
Jonathan Adams wrote:
> On Fri, Mar 17, 2006 at 11:15:40PM +0100, Roland Mainz wrote:
> > Is there any RFE yet which asks to compile the whole kernel+tools with
> > "-xstrconst -xbuiltin=%all" to improve the footprint a little bit for
> > Sun Forte/Workshop/Studio builds (I've checked the bug datab
Keith M Wesolowski wrote:
> On Fri, Mar 17, 2006 at 11:15:40PM +0100, Roland Mainz wrote:
> > Is there any RFE yet which asks to compile the whole kernel+tools with
> > "-xstrconst -xbuiltin=%all" to improve the footprint a little bit for
> > Sun Forte/Workshop/Studio builds (I've checked the bug d
On Fri, Mar 17, 2006 at 11:15:40PM +0100, Roland Mainz wrote:
> Is there any RFE yet which asks to compile the whole kernel+tools with
> "-xstrconst -xbuiltin=%all" to improve the footprint a little bit for
> Sun Forte/Workshop/Studio builds (I've checked the bug database but
> couldn't find anythi
On Fri, Mar 17, 2006 at 11:15:40PM +0100, Roland Mainz wrote:
> Is there any RFE yet which asks to compile the whole kernel+tools with
> "-xstrconst -xbuiltin=%all" to improve the footprint a little bit for
> Sun Forte/Workshop/Studio builds (I've checked the bug database but
> couldn't find anyth
Hi!
Is there any RFE yet which asks to compile the whole kernel+tools with
"-xstrconst -xbuiltin=%all" to improve the footprint a little bit for
Sun Forte/Workshop/Studio builds (I've checked the bug database but
couldn't find anything matching... ;-( ) ? gcc already does the same by
defaul
31 matches
Mail list logo
