On Wed, Apr 09, 2008 at 08:42:01PM +0200, Jim Meyering wrote:
"Daniel P. Berrange" <berrange(a)redhat.com> wrote:
> On Wed, Apr 09, 2008 at 02:40:13PM +0100, Richard W.M. Jones wrote:
>> On Wed, Apr 09, 2008 at 02:43:22PM +0200, Jim Meyering wrote:
>> > Please don't add the "tab-width: 4" specifier.
>> > Specifying a tab-width at all in a new file with "indent-tabs-mode:
nil"
>> > is a contradiction. The latter says there should be no TABs, yet
>> > the former says "when there are, give them width 4." Coding
style
>> > guidelines are universal in their recommendations to stick with 8-byte
>> > TAB stops, independent of whether you actually use TAB or spaces.
>>
>> Agreed, good idea.
>
> ACK.
>
> Could we get a make syntax-check test to look for these bogus tabs too
I've done that.
Abbreviated patches below.
However, there are some "issues" to consider.
Any global space-changing delta like these is going to cause
trouble (conflicts) for people with pending changes and on branches.
This one isn't too bad (as these things go), since the TAB-to-space
change affects fewer than 1500 lines in 37 files. However, the
new rule enforces the coding standard only in files with an existing
"indent-tabs-mode: nil" directive. There are currently 42 .[ch] files
that don't have such a directive (excluding gnulib/).
$ git ls-files|grep -E '\.[ch]$'|xargs grep -L indent-tabs-mode|grep -v \
gnulib|wc -l
42
If I were to do the same for those remaining files,
the TAB-to-space change would modify an additional 2817 lines
in 28 files:
$ git ls-files|grep -E '\.[ch]$'|xargs grep -L indent-tabs-mode|grep -v \
gnulib|xargs grep -E '^ * '|wc -l
2817
$ git ls-files|grep -E '\.[ch]$'|xargs grep -L indent-tabs-mode|grep -v \
gnulib|xargs grep -El '^ * '|wc -l
28
My opinion is that if it's worth doing the first, it's also worth
finishing the job, ... but then I don't have any huge re-architecting
changes in my queue.
The only changes I have pending are the host device enumeration patches
which are all new code, so unaffected, and the serial device support for
QEMU which is easy enough to fix-up since its fairly isolated chunks of
code. So I think its a net win to fix up everything.
Dan.
--
|: Red Hat, Engineering, Boston -o-
http://people.redhat.com/berrange/ :|
|:
http://libvirt.org -o-
http://virt-manager.org -o-
http://ovirt.org :|
|:
http://autobuild.org -o-
http://search.cpan.org/~danberr/ :|
|: GnuPG: 7D3B9505 -o- F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 :|