Re: [Qemu-devel] [PATCH] readline: avoid memcpy() of overlapping regions

[Markus Armbruster]

Blue Swirl <address@hidden> writes:

> On Thu, Jan 24, 2013 at 5:17 PM, Andreas Färber <address@hidden> wrote:
>> Am 17.01.2013 21:13, schrieb Blue Swirl:
>>> On Mon, Jan 14, 2013 at 9:09 AM, Stefan Hajnoczi <address@hidden> wrote:
>>>> Tell me what you consider the "correct" tab width for readers and I'll
>>>> find a piece of QEMU code that was authored for a different tab width
>>>> :).
>>>
>>> 8.
>>
>> So FWIW one exception is target-cris/helper.c, which seems to use 4. :)
>
> I don't think so:
>       miss = cris_mmu_translate(&res, env, address & TARGET_PAGE_MASK,
>                                 rw, mmu_idx, 0);
> The first line has one tab (with tab equal to 8 spaces, column 8) and
> the second, four tabs plus two spaces to reach column 34 (4 * 8 + 2)
> so 'rw' is nicely aligned to just after '(' in the first line. If we
> instead assumed a tab size of 4, the first line would have indent of 4
> spaces but the second 4 * 4 + 2 = 10 which would mess the indentation.
>
>> Many Windows and Eclipse-based editors use 4.
>>
>> My personal opinion is that tabs don't have any fixed width and, when
>> used properly, that works fairly well (i.e., indent the block with tabs
>> and do any parenthesis alignment etc. with blanks from block level).
>
> If the indentation did not make any assumptions about the tab
> settings, that could actually work. But using the above example, the
> second line should use only one tab and 26 spaces, then both lines
> would be aligned with any tab settings. Does anyone use tabs like
> this?

Wrong question.  The correct question is "Does everybody[*] use tabs
like this?"


[*] Everybody hacking on a particular body of code.  For a personal
project, that may be just that one person.