On 10/22/2015 08:06 AM, Michal Marek wrote:

On 2015-10-22 13:31, Prarit Bhargava wrote:

quoted

On 10/21/2015 03:52 PM, Michal Marek wrote:

quoted

Dne 21.10.2015 v 21:27 Prarit Bhargava napsal(a):

quoted

On 10/15/2015 04:16 PM, Michal Marek wrote:

quoted

Otherwise make tags can't parse them:

ctags: Warning: arch/ia64/kernel/smp.c:60: null expansion of name pattern "\1"
ctags: Warning: drivers/xen/events/events_2l.c:41: null expansion of name pattern "\1"
ctags: Warning: drivers/acpi/processor_idle.c:64: null expansion of name pattern "\1"
ctags: Warning: kernel/locking/lockdep.c:153: null expansion of name pattern "\1"
ctags: Warning: kernel/workqueue.c:305: null expansion of name pattern "\1"
ctags: Warning: kernel/rcu/rcutorture.c:133: null expansion of name pattern "\1"
ctags: Warning: kernel/rcu/rcutorture.c:135: null expansion of name pattern "\1"
ctags: Warning: net/rds/page.c:45: null expansion of name pattern "\1"
ctags: Warning: net/ipv4/syncookies.c:53: null expansion of name pattern "\1"
ctags: Warning: net/ipv6/syncookies.c:44: null expansion of name pattern "\1"

I guarantee you're going to end up fixing this issue over and over again as more
code is added in.

This is certainly going to happen, but it should be quickly spotted by
anybody running make tags on linux-next. And 10 instances since the
beginning of git is not too many.

Not everyone uses 'make tags'.  'make cscope' exists and functions correctly ;)

cscope works, but unfortunately it cannot be extended to understand the
preprocessor constructs. But it does not suffer from the problem at
hand, obviously.

quoted

quoted

It's not ctags itself parsing the DEFINE_PER_CPU() macro, but a
user-supplied regex specified on commandline. Which can only operate on
single lines.

What's the regex?

See
https://lkml.kernel.org/r/1444940195-28272-9-git-send-email-mmarek@suse.com

It used to require a closing parenthesis, so it would not match the
multiline macro invocations at all. Now it matches them, but ctags
correctly warns that the empty string is probably not what we intended
to match.

It seems wrong to change kernel code, not for a bug, but for a userspace search.

P.

Michal