1 |
>>>>> On Sat, 6 Jun 2015, Duncan wrote: |
2 |
|
3 |
>> *If* we should agree on using tabs, then we should also standardise |
4 |
>> the tab width. Using the same rules for indenting and whitespace as |
5 |
>> for ebuilds (i.e., tab stops every four positions) suggests itself: |
6 |
>> https://devmanual.gentoo.org/ebuild-writing/file-format/index.html#indenting-and-whitespace |
7 |
|
8 |
> (Somewhat) More seriously, standardizing the tab size defeats the |
9 |
> purpose, letting people decide for themselves, particularly when |
10 |
> it's to be the declared horizontal spacing standard in a file such |
11 |
> as this, where mixed spaces and tabs can be avoided, so someone's |
12 |
> personal setting shouldn't be mixed up by someone using spaces |
13 |
> instead |
14 |
|
15 |
It plays a role when at the same time there is a policy about the line |
16 |
width. For example, the devmanual has this (about _ebuilds_, not about |
17 |
metadata.xml): |
18 |
|
19 |
# Where possible, try to keep lines no wider than 80 positions. |
20 |
# A 'position' is generally the same as a character — tabs are four |
21 |
# positions wide, and multibyte characters are just one position wide. |
22 |
|
23 |
This would make no sense with the width of a tab being arbitrary. |
24 |
|
25 |
> (and if it is, the non-standard spaces in place of tabs is simply |
26 |
> much more obvious, allowing easier detection /due/ to the |
27 |
> non-standardized tabsize, and replacing with tabs as appropriate). |
28 |
|
29 |
I don't understand this part. We would have either spaces or tabs, but |
30 |
not both. And e.g. Emacs can highlight tabs (with whitespace-mode) so |
31 |
there's no problem seeing them. |
32 |
|
33 |
> But IMO it's all simply bikeshedding, regardless. |
34 |
|
35 |
Maybe. But standardising it could simplify life when updating metadata |
36 |
files with a script. |
37 |
|
38 |
Ulrich |