Skip to contentSkip to conversation

Tabs, spaces, indent and 80 character lines

DISCUSSION Richard Knutsson 5 participants 13 messages Feb 24, 2008 – Feb 26, 2008

lore
Share & tools
lore
#0 lore raw

Good evening

In the thread "Merging of completely unreviewed drivers" I got reminded of the "use tabs not spaces"-mentality. My question is: why?

The tab-character serves us well as a indent-indicator, but for some reason there has been focus on its relation to spaces. On the question "How long should a line at maximum be?" it is relevant (the question is not [1]). So it is set to be as wide as 8 spaces, but when did it become a replacement for 8 spaces? Why hinder a developer who prefer 2, 4, 6 or any other != 8 width? By only using tabs as indents, and changing the CodeStyle to be something like "maximum 80 characters-wide lines, with a tab-setting of 8 spaces", that is possible + easier to write code-checkers [2].

Or are we really that concerned about the disk-space? ;)

2 cents away...
/Richard Knutsson

[1] As has been pointed out by many, it is the complexity that matters (code-checker). A short line can be overly complicated and still under 80 lines and vice versa. Let the editors handle the long lines for the author (even I should be able to write a decent script who can decide where to chop-off the line).

[2] As it is now, it can look like the indentation is ex. ... 2, 2, 5, 3, 3... because the second line is split up. (I think it should have been 2, 2, 2, 3, 3 (or even better: 2, 2, 3, 3 ))

#1 lore raw
Quoted 2 lines · Richard Knutsson

Why hinder a developer who prefer
2, 4, 6 or any other != 8 width?

I guess we could use tabs only at the line start, for indentation only. Rather hard to implement, most text editors can't do that yet.

Quoted 3 lines

By only using tabs as indents, and
changing the CodeStyle to be something like "maximum 80 characters-wide lines, with a tab-setting of 8 spaces",

This changes nothing.

Quoted 2 lines

that is
possible + easier to write code-checkers [2].

I doubt it.

Quoted 1 line

Or are we really that concerned about the disk-space? ;)

Unpacked sources will be much bigger with not tabs, sure.

Signature · 1 line
Krzysztof Halasa
#2 lore raw
Quoted 10 lines · Krzysztof Halasa
Quoted 3 lines · Richard Knutsson

Why hinder a developer who prefer
2, 4, 6 or any other != 8 width?

I guess we could use tabs only at the line start, for indentation only. Rather hard to implement, most text editors can't do that yet.

You mean for split lines? Hopefully there won't be that many, so there is just to delete the tabs it added and replace it with spaces.

Quoted 8 lines
Quoted 5 lines

By only using tabs as indents, and
changing the CodeStyle to be something like "maximum 80 characters-wide lines, with a tab-setting of 8 spaces",

This changes nothing.

Exactly! But then we can remove the "we use 8 wide tabs in the kernel" in CodeStyle.

Quoted 7 lines
Quoted 4 lines

that is
possible + easier to write code-checkers [2].

I doubt it.

Easier to write code-checkers? OK, maybe not. Just that I got hit by this problem at a time when I wrote a simple checker (don't remember its purpose).

Quoted 6 lines
Quoted 3 lines

Or are we really that concerned about the disk-space? ;)

Unpacked sources will be much bigger with not tabs, sure.

Without no tabs at all, you mean? Don't want to think about that scenario, but with this suggestion, I would estimate maybe 0,5 - 1% bigger.

Thanks for your input

#3 lore raw
Quoted 4 lines · Richard Knutsson
Quoted 3 lines · Krzysztof Halasa

Richard Knutsson <[email protected]> writes:

Quoted 8 lines
Quoted 6 lines
Quoted 3 lines

Why hinder a developer who prefer
2, 4, 6 or any other != 8 width?

I guess we could use tabs only at the line start, for indentation only. Rather hard to implement, most text editors can't do that yet.

You mean for split lines? Hopefully there won't be that many, so there is just to delete the tabs it added and replace it with spaces.

IMO, tabs SHOULD be used for syntactic indentation and spaces for decoration purpose only. I.e. a line should start with a number of tabs equal to its nesting level and after that only spaces should be used. for example, the following code

for (i = 0; i < n; i++) printk("a very long format string", some, parameters);

should be formatted like this:

<tabs...>for (i = 0; i < n; i++)
<tabs...><tab>printk("a very long format string",
<tabs...><tab> some, parameters);

this will show exactly right regardless of your editor's tab expansion setting as long as you use fixed-width fonts - where the screen width of the space character is equal to all other characters. Once you start using tabs instead of spaces to push text right so it appears exactly below some other text on the line above you make a dependency on *your* editor's tab expansion policy and that's not very considerate for folks who prefer a different one.

Quoted 33 lines
Quoted 7 lines
Quoted 5 lines

By only using tabs as indents, and
changing the CodeStyle to be something like "maximum 80 characters-wide lines, with a tab-setting of 8 spaces",

This changes nothing.

Exactly! But then we can remove the "we use 8 wide tabs in the kernel" in CodeStyle.

Quoted 6 lines
Quoted 4 lines

that is
possible + easier to write code-checkers [2].

I doubt it.

Easier to write code-checkers? OK, maybe not. Just that I got hit by this problem at a time when I wrote a simple checker (don't remember its purpose).

Quoted 5 lines
Quoted 3 lines

Or are we really that concerned about the disk-space? ;)

Unpacked sources will be much bigger with not tabs, sure.

Without no tabs at all, you mean? Don't want to think about that scenario, but with this suggestion, I would estimate maybe 0,5 - 1% bigger.

Thanks for your input

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

#4 lore raw
Quoted 32 lines · Benny Halevy
Quoted 10 lines · Richard Knutsson
Quoted 8 lines · Krzysztof Halasa
Quoted 4 lines · Richard Knutsson

Why hinder a developer who prefer
2, 4, 6 or any other != 8 width?

Quoted 6 lines
Quoted 4 lines

I guess we could use tabs only at the line start, for indentation only. Rather hard to implement, most text editors can't do that yet.

You mean for split lines? Hopefully there won't be that many, so there is just to delete the tabs it added and replace it with spaces.

IMO, tabs SHOULD be used for syntactic indentation and spaces for decoration purpose only. I.e. a line should start with a number of tabs equal to its nesting level and after that only spaces should be used. for example, the following code

for (i = 0; i < n; i++) printk("a very long format string", some, parameters);

should be formatted like this:

<tabs...>for (i = 0; i < n; i++)
<tabs...><tab>printk("a very long format string",
<tabs...><tab> some, parameters);

Quoted 7 lines

this will show exactly right regardless of your editor's tab expansion setting as long as you use fixed-width fonts - where the screen width of the space character is equal to all other characters. Once you start using tabs instead of spaces to push text right so it appears exactly below some other text on the line above you make a dependency on *your* editor's tab expansion policy and that's not very considerate for folks who prefer a different one.

Don't know what to say more then: Yup! :)

But the CodeStyle-document and checkpatch.pl does not agree with that.

#5 lore raw
Quoted 40 lines · Richard Knutsson
Quoted 32 lines · Benny Halevy
Quoted 10 lines · Richard Knutsson
Quoted 8 lines · Krzysztof Halasa
Quoted 4 lines · Richard Knutsson

Why hinder a developer who prefer
2, 4, 6 or any other != 8 width?

Quoted 7 lines
Quoted 4 lines

I guess we could use tabs only at the line start, for indentation only. Rather hard to implement, most text editors can't do that yet.

You mean for split lines? Hopefully there won't be that many, so there is just to delete the tabs it added and replace it with spaces.

IMO, tabs SHOULD be used for syntactic indentation and spaces for decoration purpose only. I.e. a line should start with a number of tabs equal to its nesting level and after that only spaces should be used. for example, the following code

for (i = 0; i < n; i++) printk("a very long format string", some, parameters);

should be formatted like this:

<tabs...>for (i = 0; i < n; i++)
<tabs...><tab>printk("a very long format string",
<tabs...><tab> some, parameters);

Quoted 7 lines

this will show exactly right regardless of your editor's tab expansion setting as long as you use fixed-width fonts - where the screen width of the space character is equal to all other characters. Once you start using tabs instead of spaces to push text right so it appears exactly below some other text on the line above you make a dependency on *your* editor's tab expansion policy and that's not very considerate for folks who prefer a different one.

Quoted 4 lines

Don't know what to say more then: Yup! :)

But the CodeStyle-document and checkpatch.pl does not agree with that.

I know :(
If there's enough interest I can take a stab at seeing what it'd take to implement such a check in checkpatch.pl

Benny

#6 lore raw
In reply to Richard Knutsson
Quoted 4 lines · Richard Knutsson
Quoted 3 lines

I guess we could use tabs only at the line start, for indentation only. Rather hard to implement, most text editors can't do that yet.

You mean for split lines?

Syntactic indentation vs alignment (including comments after non-blank, values for struct initialization etc, split lines too).

Quoted 2 lines

Hopefully there won't be that many, so there
is just to delete the tabs it added and replace it with spaces.

Actually tabs "should" be used for indentation at start of the line, then spaces. "Ideally" :-)

I.e., something like
<TAB> if (cond && (cond2 ||
<TAB> _____________cond3))
<TAB> <TAB> do_something();

Underline = space.

Perhaps some day...

Quoted 2 lines

Exactly! But then we can remove the "we use 8 wide tabs in the kernel" in CodeStyle.

I'm not sure it's practically possible now.

Quoted 3 lines
Quoted 2 lines

Unpacked sources will be much bigger with not tabs, sure.

Without no tabs at all, you mean?

With spaces in place of all tabs.

All tabs converted to spaces = 20% more?
"Alignment" tabs converted to spaces? How cares how much more would it take if it's the correct thing. Except that it's not very practical at this point.

Signature · 1 line
Krzysztof Halasa
#7 lore raw
Quoted 7 lines · Krzysztof Halasa
Quoted 4 lines · Richard Knutsson
Quoted 4 lines

I guess we could use tabs only at the line start, for indentation only. Rather hard to implement, most text editors can't do that yet.

Quoted 7 lines
Quoted 2 lines

You mean for split lines?

Syntactic indentation vs alignment (including comments after non-blank, values for struct initialization etc, split lines too).

'alignment', that's the word, thanks!

Quoted 14 lines
Quoted 3 lines

Hopefully there won't be that many, so there
is just to delete the tabs it added and replace it with spaces.

Actually tabs "should" be used for indentation at start of the line, then spaces. "Ideally" :-)

I.e., something like
<TAB> if (cond && (cond2 ||
<TAB> _____________cond3))
<TAB> <TAB> do_something();

Underline = space.

Quoted 9 lines

Perhaps some day...

Quoted 3 lines

Exactly! But then we can remove the "we use 8 wide tabs in the kernel" in CodeStyle.

I'm not sure it's practically possible now.

Well, can always patch CodeStyle and checkpatch.pl and then see what the rest thinks.

Think checkpatch is fairly easy to "fix". It seems to have all the lines in an array so there is only a need to check the indent-depth of the line above and see if the current is the same with only spaces afterwards.

The idea was to reply with a patch, but it is bit late now.

Quoted 7 lines
Quoted 6 lines
Quoted 4 lines

Unpacked sources will be much bigger with not tabs, sure.

Without no tabs at all, you mean?

Quoted 7 lines

With spaces in place of all tabs.

All tabs converted to spaces = 20% more?
"Alignment" tabs converted to spaces? How cares how much more would it take if it's the correct thing. Except that it's not very practical at this point.

Would guess more...
The reason I reacted were because checkpatch.pl complained about a:
<tab><tab>foo(int a,
<tab><tab>____int b)
with the "use tabs not spaces". There is little point in patching this stuff but lets not make it worse (IMHO).

#9 lore raw
Quoted 4 lines · Miles Bader

Why do people even respond to these trolls...?

-Miles

Obviously, this must to have been discussed before, with a clear conclusion. Unfortunately, I have not, during my ~2 years on the list, seen anything of the sort and would like to, in that case, request a small direction and apologize for the noise.

Richard "_not_ all knowing" Knutsson

BTW, your question:
* they like to be helpful
* they find the question reasonable
* they like furry animals
...
Sorry, am tired, but if you like I can keep you posted when I come up with more reasons.

#10 lore raw
Quoted 4 lines · Richard Knutsson
Quoted 2 lines · Miles Bader

Why do people even respond to these trolls...?

Obviously, this must to have been discussed before, with a clear conclusion.

It has been discussed before, at http://lkml.org/lkml/2007/11/12/19 .

What is really frustrating is that some of the people which _do_ enforce one style of the two (tabs-only or tabs-spaces) have only indirectly voiced their preferred style, as in
http://lkml.org/lkml/2008/1/19/67 .

Often it was just (1)"checkpatch flagged your spaces, hence they are wrong" instead of (2)"in my tree, I only take tabs-only patches".

Now back to coding, oh and don't forget send a patch for CodingStyle since a mail without one is often taken even less seriously.

#11 lore raw
Quoted 2 lines · Jan Engelhardt

Now back to coding, oh and don't forget send a patch for CodingStyle since a mail without one is often taken even less seriously.

Someone with a patch to Emacs to use tabs for ident + spaces for alignment maybe? :-)

Signature · 1 line
Krzysztof Halasa
#12 lore raw
In reply to Jan Engelhardt
Quoted 12 lines · Jan Engelhardt
Quoted 7 lines · Richard Knutsson
Quoted 3 lines · Miles Bader

Why do people even respond to these trolls...?

Obviously, this must to have been discussed before, with a clear conclusion.

It has been discussed before, at http://lkml.org/lkml/2007/11/12/19 .

Quoted 9 lines

What is really frustrating is that some of the people which _do_ enforce one style of the two (tabs-only or tabs-spaces) have only indirectly voiced their preferred style, as in
http://lkml.org/lkml/2008/1/19/67 .

Often it was just (1)"checkpatch flagged your spaces, hence they are wrong" instead of (2)"in my tree, I only take tabs-only patches".

Now back to coding, oh and don't forget send a patch for CodingStyle since a mail without one is often taken even less seriously.

What can I say? Doh! Am a bit surprised I had totally missed it during the mailbox-cleanups (to prevent lock-down).

Thank you! for the heads-up.
Richard Knutsson