svn commit: r326658 - head/lib/libefivar

Bruce Evans brde at optusnet.com.au
Fri Dec 8 01:52:20 UTC 2017


On Thu, 7 Dec 2017, Warner Losh wrote:

> On Thu, Dec 7, 2017 at 3:50 PM, Bruce Evans <brde at optusnet.com.au> wrote:
>
>> On Thu, 7 Dec 2017, Mark Johnston wrote:
>>
>> Log:
>>>  Ensure that "out" is initialized in all error paths.
>>> ...
>>> Modified: head/lib/libefivar/efivar-dp-xlate.c
>>> ============================================================
>>> ==================
>>> --- head/lib/libefivar/efivar-dp-xlate.c        Thu Dec  7 09:05:34
>>> 2017        (r326657)
>>> +++ head/lib/libefivar/efivar-dp-xlate.c        Thu Dec  7 15:16:17
>>> 2017        (r326658)
>>> @@ -529,7 +529,7 @@ build_dp(const char *efimedia, const char *relpath, ef
>>> {
>>>         char *fp, *dptxt = NULL;
>>>         int rv = 0;
>>> -       efidp out;
>>> +       efidp out = NULL;
>>>         size_t len;
>>>
>>>         fp = path_to_file_dp(relpath);
>>>
>>
>> This is a worse style of initializations in declararations than usual.
>> 1 initialization in a non-declaration is mixed with many initializations
>> in declarations, using the C99 misfeature of allowing initialiations
>> after statements, without even any blank lines to keep the declarations
>> separate.
>
> This matches the current style of the file though...

The rest of the file, and other libefivar files with 1-tab indentation,
are not very far from KNF.  Some even have a blank line after null
declarations.  This file even has parentheses around return values in
most places.

>> Most libefivar code is much uglier than this, with 2-space indentation
>> and worse.
>
> The code you are complaining about is vendor code.... The 2-space
> indentation is the least of the issues with that code....

If it were in the vendor tree, then I could not-see it more easily.

Bruce


More information about the svn-src-all mailing list