Re: Another bit of ISO C conversion...

Re: Another bit of ISO C conversion...

[Richard Kenner]

    > My preference is the indented type.  I would do a grep for "define FOO"
    > and would normally want to see the commented-out #defines too.

    I don't see why indentation matters here.  Either way, your grep for
    "define FOO" should still show what you describe.

Sorry I was unclear.

I was making two points:

(1) I like the indented style better since I think it's easier to read.
(2) Since I grep for "define FOO", not "#define FOO", I don't see the
ability to grep for the latter as a major advantage of not indenting.

Re: Another bit of ISO C conversion...

[James Lemke]

> My preference is the indented type.  I would do a grep for "define FOO"
> and would normally want to see the commented-out #defines too.

I don't see why indentation matters here.  Either way, your grep for
"define FOO" should still show what you describe.

-- 
James Lemke   jim@wasabisystems.com   Orillia, Ontario
http://www.wasabisystems.com

Re: Another bit of ISO C conversion...

[Zack Weinberg]

"Kaveh R. Ghazi" <ghazi@caip.rutgers.edu> writes:

>
> Since you brought up this code in hwint.h, I was wondering...
>
> Didn't we used to use HOST_WIDE_INT=int in some cases?  What happened
> to that?

No clue.

zw

Re: Another bit of ISO C conversion...

[Kaveh R. Ghazi]

 > This can become
 > 
 > #if HOST_BITS_PER_LONG >= 64 || !defined NEED_64BIT_HOST_WIDE_INT
 > #   define HOST_BITS_PER_WIDE_INT HOST_BITS_PER_LONG
 > #   define HOST_WIDE_INT long
 > #elif HOST_BITS_PER_LONGLONG >= 64
 > #   define HOST_BITS_PER_WIDE_INT HOST_BITS_PER_LONGLONG
 > #   define HOST_WIDE_INT long long
 > #elif HOST_BITS_PER___INT64 >= 64
 > #   define HOST_BITS_PER_WIDE_INT HOST_BITS_PER___INT64
 > #   define HOST_WIDE_INT __int64
 > #else
 > #   error "Unable to find a suitable type for HOST_WIDE_INT"
 > #endif
 > 
 > which is simpler, easier to read, and easier to modify.  


Since you brought up this code in hwint.h, I was wondering...

Didn't we used to use HOST_WIDE_INT=int in some cases?  What happened
to that?

		Thanks,
		--Kaveh
--
Kaveh R. Ghazi			ghazi@caip.rutgers.edu

Re: Another bit of ISO C conversion...

[Zack Weinberg]

Steven Bosscher <s.bosscher@student.tudelft.nl> writes:

> Before going to convert all files to #elif, can we agree on what style we 
> should use?
>
> I would prefer to keep # in column 1 so you can still grep for ^#define etc.  
> In config/, there are quite a few files with commented out /* #defines */...

That is my preference also.

zw

Re: Another bit of ISO C conversion...

[Richard Kenner]

    I would prefer to keep # in column 1 so you can still grep for
    ^#define etc.  In config/, there are quite a few files with commented
    out /* #defines */...

My preference is the indented type.  I would do a grep for "define FOO"
and would normally want to see the commented-out #defines too.

Re: Another bit of ISO C conversion...

[Steven Bosscher]

On Monday 08 December 2003 01:27, Zack Weinberg wrote:
> We've had a lot of people put in time to convert GCC to ISO C style,
> and I want to say that I appreciate all of that effort.  There's one
> transformation that hasn't been getting done, though: use of #elif.
> K&R C didn't have this directive so we have lots of nested #if
> blocks.  For example, from hwint.h:

Hmm looks like a good cleanup.

> Note also that it is no longer necessary to indent the leading # of
> an #error directive; or, turnabout, it is no longer necessary to write
> the leading # of any directive in column 1, so this becomes a style
> decision.

Before going to convert all files to #elif, can we agree on what style we 
should use?

I would prefer to keep # in column 1 so you can still grep for ^#define etc.  
In config/, there are quite a few files with commented out /* #defines */...

Gr.
Steven

Another bit of ISO C conversion...

[Zack Weinberg]

We've had a lot of people put in time to convert GCC to ISO C style,
and I want to say that I appreciate all of that effort.  There's one
transformation that hasn't been getting done, though: use of #elif.
K&R C didn't have this directive so we have lots of nested #if
blocks.  For example, from hwint.h:

#if HOST_BITS_PER_LONG >= 64 || !defined NEED_64BIT_HOST_WIDE_INT
#   define HOST_BITS_PER_WIDE_INT HOST_BITS_PER_LONG
#   define HOST_WIDE_INT long
#else
# if HOST_BITS_PER_LONGLONG >= 64
#   define HOST_BITS_PER_WIDE_INT HOST_BITS_PER_LONGLONG
#   define HOST_WIDE_INT long long
# else
#  if HOST_BITS_PER___INT64 >= 64
#   define HOST_BITS_PER_WIDE_INT HOST_BITS_PER___INT64
#   define HOST_WIDE_INT __int64
#  else
    #error "Unable to find a suitable type for HOST_WIDE_INT"
#  endif
# endif
#endif

This can become

#if HOST_BITS_PER_LONG >= 64 || !defined NEED_64BIT_HOST_WIDE_INT
#   define HOST_BITS_PER_WIDE_INT HOST_BITS_PER_LONG
#   define HOST_WIDE_INT long
#elif HOST_BITS_PER_LONGLONG >= 64
#   define HOST_BITS_PER_WIDE_INT HOST_BITS_PER_LONGLONG
#   define HOST_WIDE_INT long long
#elif HOST_BITS_PER___INT64 >= 64
#   define HOST_BITS_PER_WIDE_INT HOST_BITS_PER___INT64
#   define HOST_WIDE_INT __int64
#else
#   error "Unable to find a suitable type for HOST_WIDE_INT"
#endif

which is simpler, easier to read, and easier to modify.  

Note also that it is no longer necessary to indent the leading # of 
an #error directive; or, turnabout, it is no longer necessary to write
the leading # of any directive in column 1, so this becomes a style
decision.

zw