Re: [PATCH hail] const-correctness tweaks

[Jim Meyering <jim@meyering.net>]

Jeff Garzik wrote:

> On 10/06/2010 08:07 AM, Jim Meyering wrote:
>>
>> Make write_cb callback's buffer parameter const, like all write-like
>> functions.
>> Give a few "char *" parameters the "const" attribute.
>>
>> Signed-off-by: Jim Meyering<meyering@redhat.com>
>> ---
>>
>> It looks like most of hail's interfaces are const-correct,
>> but one stood out because it provokes a warning when I tried to
>> pass a const-correct write_cb function to hstor_get from iwhd:
>>
>>      proxy.c:382: warning: passing argument 4 of 'hstor_get' from \
>>        incompatible pointer type
>>      /usr/include/hstor.h:173: note: expected \
>>        'size_t (*)(void *, size_t, size_t,  void *)' but argument is of type \
>>        'size_t (*)(const void *, size_t,  size_t,  void *)'
>>
>> In case you feel comfortable fixing this, here's a patch:
>
>
> Sorry for not getting back to this; I had hoped to solve some
> additional problems that cropped up, but didn't have time.  So, to
> forestall further delay,
>
> libtool: compile:  gcc -DHAVE_CONFIG_H -I. -I.. -I../include -pthread
> -I/usr/include/glib-2.0 -I/usr/lib64/glib-2.0/include
> -I/usr/include/libxml2 -O2 -Wall -Wshadow -g -MT hutil.lo -MD -MP -MF
> .deps/hutil.Tpo -c hutil.c -o hutil.o
> hutil.c: In function ‘hreq_hdr_push’:
> hutil.c:145: warning: assignment discards qualifiers from pointer
> target type
> hutil.c:146: warning: assignment discards qualifiers from pointer
> target type
>
> warnings appear after this patch.  When solving these warnings with
> const' markers, it quickly becomes a bit of a rat's nest.
>
> At a minimum, the write_cb callback signature must match libcurl's,
> which does not use 'const'.  I can see this makes sense from libcurl
> implementation's perspective, even if it does not really match the
> constness one expects from a foo-get function.

Hi Jeff.

Sorry I didn't notice that the first time.
I built with ./autogen.sh && ./configure && make.
It looks like you recommend -Wall -Wshadow.

The two warnings above are the only ones I see with the patch,
and they're easy to fix.  When storing const pointer params into
a struct like that, I've found that it's best to cast away the "const",
which really does reflect the semantics: by using "const" on the
parameter, I view it as promising not to deref through the pointer
*in that function*.  Since it's usually not reasonable to make
the struct member "const" (as you saw, it propagates too far
and often ends up being contradictory), the lesser evil of the cast
is preferable here.

If you're still game, the following incremental patch seems to be
enough for me:  Let me know and I'll resubmit the full one.

diff --git a/lib/hutil.c b/lib/hutil.c
index 13a8d5e..b74460b 100644
--- a/lib/hutil.c
+++ b/lib/hutil.c
@@ -142,8 +142,8 @@ int hreq_hdr_push(struct http_req *req, const char *key, const char *val)
 		val++;

 	hdr = &req->hdr[req->n_hdr++];
-	hdr->key = key;
-	hdr->val = val;
+	hdr->key = (char *) key;
+	hdr->val = (char *) val;

 	return 0;
 }