Re: [PATCH 3/3] tracing/filters: disallow newline as delimeter

[Tom Zanussi <tzanussi@gmail.com>]

On Wed, 2009-04-22 at 10:09 +0200, Ingo Molnar wrote:
> * Tom Zanussi <tzanussi@gmail.com> wrote:
> 
> > On Tue, 2009-04-21 at 12:16 +0200, Ingo Molnar wrote:
> > > * Li Zefan <lizf@cn.fujitsu.com> wrote:
> > > 
> > > > Ingo Molnar wrote:
> > > > > * Li Zefan <lizf@cn.fujitsu.com> wrote:
> > > > > 
> > > > >> Ingo Molnar wrote:
> > > > >>> * Li Zefan <lizf@cn.fujitsu.com> wrote:
> > > > >>>
> > > > >>>> I guess because user input is often ended with '\n' (like "echo 
> > > > >>>> xxx"), thus '\n' is used as a delimeter besides ' ', but we can 
> > > > >>>> just strip tailing spaces.
> > > > >>> Hm, how about:
> > > > >>>
> > > > >>> ( echo 'x'
> > > > >>>   echo '|| y' ) > filter
> > > > >>>
> > > > >>> type of scripts? Shouldnt the parser be permissive in general?
> > > > >>>
> > > > >> This patch doesn't forbid this usage: ;)
> > > > >>
> > > > >> ( echo 'parent_comm == a'
> > > > >>   echo '|| parent_comm == b' ) > filter
> > > > >>
> > > > >> This patch does forbid this usage:
> > > > >>
> > > > >> ( echo 'parent_comm'
> > > > >>   echo '=='
> > > > >>   echo 'a' ) > filter
> > > > > 
> > > > > Same argument though, no?
> > > > > 
> > > > 
> > > > Then I have no strong opinion on this. I'm fine to drop this patch.
> > > 
> > > I've applied the other two - no strong opinion either about this 
> > > patch. Tom, what do you think? (there's also some new parser in the 
> > > works i suspect)
> > 
> > I think it works ok as it is, so dropping this patch would be fine 
> > with me.
> > 
> > I am working on a new parser; I'd hoped to have it finished by 
> > now, but I seem to be continually distracted lately. :-( At this 
> > point I have a parser that basically works; the main thing it 
> > still needs and what I'm working on now is a way to allow it to be 
> > easily extended to support special-case handling for special types 
> > e.g. if a predicate field refers to something that's a dev_t, the 
> > user should be able to specify 'device == /dev/sda' or 'device == 
> > sda' or 'device == (8,1)' or 'device == 8:1', so it should be 
> > relatively easy to add that kind of support to the parser when 
> > there's a need for it.
> > 
> > Once that's done, I'll hook it all up and post it as soon as I 
> > can, at the end of the week hopefully...
> 
> Nice! :-)
> 
> If your current lineup works you might want to post that straight 
> away even without the type extensions, if you think there's value in 
> testing/reviewing those bits independently. It generally works 
> better to have gradual patches. (we can find bugs sooner, review is 
> easier, etc.)
> 

Sure, I understand.  I wasn't planning on adding any type extensions
with the initial parser, just restructuring the code so that adding them
wouldn't require a lot of restructuring later.  So, yeah, just the basic
parser once I get it into a postable state, very soon, and anything else
for later...

Tom 

> 	Ingo