Jim Meyering wrote:
On Fri, Mar 14, 2014 at 7:36 AM, Paul Eggert <[email protected]> wrote:
behoffski wrote:
One easy way to spot tab characters in C sources
This is grep, right? So we should use "grep ' '".
Jim Meyering has the strongest feeling about tabs, so I assume he'll pick
this one up.
Hi Brenton,
I do indeed have strong feelings about tabs :-)
Like remove all leading ones, and any TAB-after-space.
[rules to enforce that are in maint.mk and cfg.mk]
However, a few times when a crusade led me to do
what you're proposing and eliminate all of them, I later realized
that I found value in TAB-aligned comments, and actually regretted
performing that particular TAB->SP+ conversion.
On a code normalization front, I am inclined to fully automate the
process, and have spent time (though not recently) trying to find
the combination of settings in the uncrustify tool that recreate the
style we use in a few projects. If you feel like pursuing this, please
look at uncrustify and see if you can minimize the changes required
to move grep to something uncrustify-generated.
Bottom line, I'm not terribly keen on changing comment-aligning
tabs to spaces.
Tabs in code are indeed evil. Different users have different tab stops.
I would convert *all* tabs to spaces and add to each file:
/* -*- tab-width:8; intent-tabs-mode:nil; -*- */
/* ex: set tabstop=8 expandtab: */
A lot of people make a lot of effort to make their code look good. Then
someone makes a change and uses a different tab setting and makes it
look ugly, either in a pager or an editor.
Bottom line is to not permit patches with tabs.
-- Bruce Dubbs
linuxfromscratch.org
P.S. There was a considerable amount of research in the 1980s that
investigated cognitive interpretation of code at different tab stops.
The general conclusion was that between 2 and 4 spaces maximizes
comprehension.
