linux-kernel - text processing (Re: Google's Summer of Code?)

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-Id: <E1JWixC-0003Ap-V2@flower>
Date:	Wed, 5 Mar 2008 03:01:58 +0100
From:	Oleg Verych <olecom@...wer.upol.cz>
To:	Andi Kleen <andi@...stfloor.org>,
	Linus Torvalds <torvalds@...ux-foundation.org>
Cc:	"Pekka Enberg" <penberg@...helsinki.fi>,
	linux-kernel@...r.kernel.org, mingo@...e.hu,
	vegard.nossum@...il.com, Andrew Morton <akpm@...ux-foundation.org>
Subject: text processing (Re: Google's Summer of Code?)

Andi Kleen @ Tue, 04 Mar 2008 21:38:46 +0100:
> Andrew Morton
[]
>> http://kernelnewbies.org/KernelProjects
>>
>> There are surely many more things we could put there.
>>
>> I receive a dribble of emails about the setrlimit64/getrlimit64 one, so
>> people are looking at it, and are looking to do work.  (I haven't usefully
>> responded to those emails, btw - am not sure how my name got on that one -
>> probably Ulrich would be better).
>
> http://marc.info/?l=linux-kernel&m=119981029708530&w=2
>
> is also a candidate. There are still quite a lot of unconverted 
> drivers over.

And you've saw first patch there...



I'd say, that i see similar things here (LKML, kernel), wrt shell usage
and text processing.

* checkpatch instead of hard-armed editors (*The* tools of programmers),
  with one's (linux, glibc, bash, whatever) source-friendly capabilities
  (error/coding-style highlight, easy call-graph, param checking, security
   audit(+audit scripts), etc.)

* linux-2.6/scripts/unifdef.c instead of coding style + simple script
  (reinventing of compilers is a dream of CS professors of all times :)

* much of te `make` based stuff

* text processing, which is source code processing, if we are in
  Open Source, has no place in
  + design (super-macro constructs --> C code),
  + auditing (stupid vmsplice() case *and* first ``fix'')
  + testing (writing source in parallel with constructing userspace
    test programs, based on same source; once all is done, script
    generates/constructs kernel part)

* and perl is everywhere


On my `sed` scripts i was getting (from Sam):

   "Because your shell script is unreadable by normal human beings[*]
   while the perl script for people with a bit of perl fu can read it
   and fix/modify it.

   We want tools that can be maintained and enhanced by most people.

   [*] Normal human beings are people with same level of shell
   scripting/sed skills that I have just to put that straight."

   "Linecount is down but so is maintainability / extendability."

So, no tools or perl is better than nothing?

I don't say, i will solve Andi's quest, i just lost interest. But it is
damn interesting one! One, that many script kiddies will do in minutes,
if they would knew `sed` and a bit of UNIX practice, but not perl, C,
diff, git, etc.

So, teach youngsters about "maintainability / extendability" and "Normal
human beings", or what? OTOH, Who are teachers?

Just two points to show skill mismatch, i.e. for

+ managing/manipulating source,
+ designing, writing, maintaining correct kernel code.

Latter isn't for n00bs, right?

* multi-line grep
    Andi Kleen http://mid.gmane.org/20080109000358.GF2117@one.firstfloor.org

* full and correct greping of linux style function definitions
    Linus Torvalds http://mid.gmane.org/1054519757.161606@palladium.transmeta.com

All in one shot:

# print linux-style function definitions

sed -n '
/^[^[:blank:]#/].*[),]$/{
 /,$/{
 :_start;
   N;
   s=)$=&= ; t_end ; b_start;
 :_end;
 };
p}' kernel/*c | pager

#_____

One may not know `sed` at all (but i glad to explain and share
everything, i know), *patterns* and expressions are key things.

Once you have approved and try-and-buy tested one, everything else is
pure technical thing. Maybe somebody outstanding (like Rusty) can do that
with CPP, i don't know; `sed` is just more familiar thing for this.

Silence in reply i will understand quite right, i might be wrong. I just
can't see all that stuff not having much skilled people
involved/interested.
--
-o--=O`C
 #oo'L O
<___=E M
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/