threads / patch / 44911

patchattr: mark a file-local symbol as static

Subject: [PATCH] attr: mark a file-local symbol as static

## tl;dr

3 messages between Jan 18, 2017 and Jan 19, 2017. Diffs are folded; open one to read it.

replies: 2people: 2as markdown or json

Ramsay Jones· Jan 18, 2017, 22:27 UTC · lore

Signed-off-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
---

Hi Brandon,

If you need to re-roll your 'bw/attr' branch, could you please squash this into the relevant patch (commit 8908457159, "attr: use hashmap for attribute dictionary", 12-01-2017).

Also, I note that, although they are declared as part of the public attr api, attr_check_clear() and attr_check_reset() are also not called outside of attr.c. Are these functions part of the public api?

Also, a minor point, but attr_check_reset() is called (line 1050) before it's definition (line 1114). This is not a problem, given the declaration in attr.h, but I prefer definitions to come before use, where possible.

Thanks!

ATB, Ramsay Jones

 attr.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Show changes to attr.c +1 −1

diff --git a/attr.c b/attr.c
index f5cc68b67..e68c4688f 100644
--- a/attr.c
+++ b/attr.c
@@ -83,7 +83,7 @@ static int attr_hash_entry_cmp(const struct attr_hash_entry *a,
 }
 
 /* Initialize an 'attr_hashmap' object */
-void attr_hashmap_init(struct attr_hashmap *map)
+static void attr_hashmap_init(struct attr_hashmap *map)
 {
 	hashmap_init(&map->map, (hashmap_cmp_fn) attr_hash_entry_cmp, 0);
 }

-- 
2.11.0

Brandon Williams· Jan 18, 2017, 23:05 UTC · re: Ramsay Jones · lore

Re: [PATCH] attr: mark a file-local symbol as static

On 01/18, Ramsay Jones wrote:

Show 24 quoted lines

> 
> Signed-off-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
> ---
> 
> Hi Brandon,
> 
> If you need to re-roll your 'bw/attr' branch, could you please
> squash this into the relevant patch (commit 8908457159,
> "attr: use hashmap for attribute dictionary", 12-01-2017).
> 
> Also, I note that, although they are declared as part of the
> public attr api, attr_check_clear() and attr_check_reset() are
> also not called outside of attr.c. Are these functions part of
> the public api?
> 
> Also, a minor point, but attr_check_reset() is called (line 1050)
> before it's definition (line 1114). This is not a problem, given
> the declaration in attr.h, but I prefer definitions to come before
> use, where possible.
> 
> Thanks!
> 
> ATB,
> Ramsay Jones

Yes of course, I believe Stefan also pointed that out earlier today so I have it fixed locally.

For attr_check_clear() and attr_check_reset() the intent is that they are the accepted way to either clear or reset the attr_check structure. Currently most users of the attribute system don't have a need to clear or reset the structure but there could be future callers who need that functionality. If you feel like they shouldn't be part of the api right now then I'm open to changing that for this series.

Thanks!

-- 
Brandon Williams

Ramsay Jones· Jan 19, 2017, 01:22 UTC · re: Brandon Williams · lore

Re: [PATCH] attr: mark a file-local symbol as static

On 18/01/17 23:05, Brandon Williams wrote:

Show 28 quoted lines

> On 01/18, Ramsay Jones wrote:
>>
>> Signed-off-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
>> ---
>>
>> Hi Brandon,
>>
>> If you need to re-roll your 'bw/attr' branch, could you please
>> squash this into the relevant patch (commit 8908457159,
>> "attr: use hashmap for attribute dictionary", 12-01-2017).
>>
>> Also, I note that, although they are declared as part of the
>> public attr api, attr_check_clear() and attr_check_reset() are
>> also not called outside of attr.c. Are these functions part of
>> the public api?
>>
>> Also, a minor point, but attr_check_reset() is called (line 1050)
>> before it's definition (line 1114). This is not a problem, given
>> the declaration in attr.h, but I prefer definitions to come before
>> use, where possible.
>>
>> Thanks!
>>
>> ATB,
>> Ramsay Jones
> 
> Yes of course, I believe Stefan also pointed that out earlier today so I
> have it fixed locally.

Yep, I noticed Stefan's email just a few minutes after I hit send! ;-)

Show 6 quoted lines

> For attr_check_clear() and attr_check_reset() the intent is that they
> are the accepted way to either clear or reset the attr_check structure.
> Currently most users of the attribute system don't have a need to clear
> or reset the structure but there could be future callers who need that
> functionality.  If you feel like they shouldn't be part of the api right
> now then I'm open to changing that for this series.

No, I just wanted to check that they were intended to be part of the public api and that you anticipate additional callers in the future.

[I would still prefer definitions before use, but many people would not agree with me, so ...]

ATB, Ramsay Jones