Hi Joonwoo,

This one also I don't understand.  fixincludes is already supposed to be doing 
new conversion in asm operands.  And something inside quotes -- which it looks 
like your stuff checks for -- shouldn't be parsed by C++.  The code in 
2.6.35.6 doesn't look like it has "new" in it any more.  Can you send the 
example?

Eddie


On 09/25/2010 10:29 PM, Joonwoo Park wrote:
> eliminate following build error on linux 2.6.35 arch x86_64
> include/click-linuxmodule/include0/asm/uaccess_64.h: In function 
> ‘copy_user_generic’:
> include/click-linuxmodule/include0/asm/uaccess_64.h:32: error: undefined 
> named operand ‘new’
>
> Signed-off-by: Joonwoo Park<[email protected]>
> ---
>   linuxmodule/fixincludes.pl |    5 +++++
>   1 files changed, 5 insertions(+), 0 deletions(-)
>
> diff --git a/linuxmodule/fixincludes.pl b/linuxmodule/fixincludes.pl
> index 54a0813..a4c9aa7 100755
> --- a/linuxmodule/fixincludes.pl
> +++ b/linuxmodule/fixincludes.pl
> @@ -209,6 +209,11 @@ sub one_includeroot ($$) {
>           # unquote.
>           $_ = sunprotect($_);
>
> +         # "new" and other keywords in inline assembler
> +         s{(\basm\b.*)+(".*?[^\\])(\bnew\b)(.*")}{$1$2new_value$4}g;
> +         s{(\basm\b.*)+(".*?[^\\])(\band\b)(.*")}{$1$2and_value$4}g;
> +         s{(\basm\b.*)+(".*?[^\\])(\bswap\b)(.*")}{$1$2swap_value$4}g;
> +
>           # perhaps nothing has changed; avoid changing the timestamp
>           if (-f "$outputroot$dd/$d") {
>               open(F, "$outputroot$dd/$d");

_______________________________________________
click mailing list
[email protected]
https://amsterdam.lcs.mit.edu/mailman/listinfo/click

Reply via email to