Re: [PATCH] update checkpatch.pl to version 0.03
From: Jan Engelhardt
Date: Fri Jun 08 2007 - 11:16:48 EST
On Jun 8 2007 16:42, Adrian Bunk wrote:
>On Fri, Jun 08, 2007 at 04:34:01PM +0200, Jesper Juhl wrote:
>> On 08/06/07, Adrian Bunk <bunk@xxxxxxxxx> wrote:
>>> It's not only about MODULE_AUTHOR, if you consider it rude to limit
>>> people's names to ASCII, then don't forget that we have printk's like
>>> Linux agpgart interface v0.102 (c) Dave Jones
>>> What happens if the maintainer changes and it's now
>>> Linux agpgart interface v0.103 (c) Dave Ãnes
>>> Does the console handle it correctly during boot?
>>> Can all tools that process the syslog cope with it?
>>> Perhaps the answer is in both cases "yes", but it's a completely
>>> untested area.
>>> We really must have all bugs shaken out and all users using fixed tools
>>> _before_ we can start outputting UTF-8 - limiting people's names to
>>> ASCII in not ideal, but IMHO causing breakages for users is a much
>>> bigger problem.
>> I haven't looked at it in depth yet, but it would seem we already have
>> a few files that need to be looked at with this in mind. Looks like
>> it's not exactely a new problem (although all the following could be
>> in comments of course)...
>They should all be in comments, and all UTF-8 I've ever seen in such
>files was only in comments (mostly author names). So yes, adding UTF-8
>to program code will result in new problems.
>If you find any source file that contains UTF-8 outside of comments
>please complain loudly.
I present loudly and proudly (I *don't* complain):
17:11 ichi:/ws/linux-2.6.22-rc4 > find . -iname '*.[ch]' -print0 | xargs -0 grep MODULE_AUTHOR | grep -P '[\x80-\xff]' --color=never
./arch/i386/kernel/cpu/cpufreq/e_powersaver.c:MODULE_AUTHOR("Rafaï Bilski <rafalbilski@xxxxxxxxxx>");
./drivers/char/watchdog/i6300esb.c:MODULE_AUTHOR("Ross Biro and David Hïdeman");
./drivers/char/watchdog/w83627hf_wdt.c:MODULE_AUTHOR("Pïraig Brady <P@xxxxxxxxxxxxxx>");
./drivers/hwmon/via686a.c:MODULE_AUTHOR("Kyïti Mïkki <kmalkki@xxxxxxxxx>, "
./drivers/i2c/busses/i2c-via.c:MODULE_AUTHOR("Kyïti Mïkki <kmalkki@xxxxxxxxx>");
./drivers/isdn/hisax/isdnhdlc.c:MODULE_AUTHOR("Wolfgang Mïs <wolfgang@xxxxxxxxxxxxx>, "
./drivers/mtd/devices/phram.c:MODULE_AUTHOR("JÃrn Engel <joern@xxxxxxxxxxxxxx>");
./drivers/mtd/maps/cfi_flagadm.c:MODULE_AUTHOR("KÃri DavÃÃsson <kd@xxxxxxxx>");
./drivers/mtd/maps/dbox2-flash.c:MODULE_AUTHOR("KÃri DavÃÃsson <kd@xxxxxxxx>, Bastian Blank <waldi@xxxxxxxxxx>, Alexander Wild <wild@xxxxxxxxxxxxxxxxx>");
./drivers/net/irda/kingsun-sir.c:MODULE_AUTHOR("Alex Villacïs Lasso <a_villacis@xxxxxxxxxxxxx>");
So, we had some ISO8859-1 and some UTF-8 in there already. (And as for
MODULE_AUTHOR, it should stay there - 'fix' modinfo instead.)
BTW, there's also still a ton of non-UTF-8 in the kernel; I've already
fixed that weeks ago and sent some patch to trivial@, Adrian -
did you receive it?