Is an empty initializer list valid C code?
Solution 1
No, an empty initializer list is not allowed. This can also be shown by GCC when compiling with -std=c99 -pedantic
:
a.c:4: warning: ISO C forbids empty initializer braces
The reason is the way the grammar is defined in §6.7.9 of the 2011 ISO C Standard:
initializer:
assignment-expression
{ initializer-list }
{ initializer-list , }
initializer-list:
designation(opt) initializer
initializer-list , designation(opt) initializer
According to that definition, an initializer-list must contain at least one initializer.
Solution 2
According to the C99 standard, array creation with an empty initializer list is forbidden. In a previous answer, you can see that grammar does not describe this case.
But what happens if you declare an array without initialization? Well, it depends on the compiler which you use. Let's take a look at this simple example: int arr[5] = {}
.
GCC
By default gcc
does not produce any warnings/errors when you try to compile this code. Not even -Wall
, but -Wpedantic
does.
warning: ISO C forbids empty initializer braces
But anyway gcc
fill members of an array with 0's exactly as if you specify it explicitly int arr[5] = {0}
see assembly output godbolt.
CLANG
But default not showing warnings about this case, but with option -Wgnu-empty-initializer
does:
warning: use of GNU empty initializer extension
Clang generates different assembly code godbolt but behaves the same.
Related videos on Youtube
Comments
-
Marcus Ahlberg almost 2 years
It is common to use
{0}
to initialize astruct
or anarray
but consider the case when the first field isn't a scalar type. If the first field ofstruct Person
is anotherstruct
or array, then this line will result in an error (error: missing braces around initializer
).struct Person person = {0};
At least GCC allows me to use an empty initializer list to accomplish the same thing
struct Person person = {};
But is this valid C code?
Also: Is this line guaranteed to give the same behavior, i.e. a zero-initialized
struct
?struct Person person;
-
Montre almost 11 yearsC shouldn't guarantee initialization of anything except
static
variables, but I'm going off a single sentence in Wikipedia here. -
Marcus Ahlberg almost 11 yearsI might add that I built my code with GCC options pedantic and warnings-as-errors. Other compilers or configurations might allow the first syntax.
-
NREZ almost 11 yearsCan you have a look at this: stackoverflow.com/questions/755305/empty-structure-in-c
-
Hasturkun almost 11 years@NREZ: That's a different issue, being a struct with no members.
-
ams almost 11 years@millimoose: variables that are explicitly initialized, even with empty braces, better had be guaranteed or something is badly broken! Local (auto) variables with an initializer are a different story though.
-
NREZ almost 11 years@Hasturkun I was pointing out the different behavior of compilers for structures in C...
-
Grijesh Chauhan almost 11 yearsAlthough this is a valid code
struct X{}; int main(){ struct X x = {}; return 1; }
Give it a try. -
Gorpik almost 11 years@Bentoy13 VC++ does not support C99, so it is not a good compiler for this specific question.
-
Montre almost 11 years@ams I was referring to the last line. I should've said "shouldn't guarantee implicit initialization" to be more specific. Obviously skipping explicit initialization would be ludicrous
-
Bentoy13 almost 11 years@millimoose & Gorpik: comment deleted; for all reasons, my comment is not pertinent
-
Gabriel Staples about 4 yearsRelated, and very useful: How to initialize all members of an array to the same value?.
-
-
torek almost 11 yearsThis is the correct answer (that an empty initializer-list is forbidden by the syntax). It's a bit unfortunate and it would be nice if the standards allowed it, but they don't really allow empty objects in general anyway, so it is not needed. The answer to the remaining part of the question (do you get as-if-zero if there is no
= {...}
part) is "yes if the object has static duration, no if not". -
Marcus Ahlberg almost 11 yearsThis answers my question. I still don't have a fool-proof way of initializing the struct on a single line if I don't know the type of the first member (as is the case if the struct declaration is being auto-generated), but that is a different question.
-
Pavel Šimerda almost 9 years@MarcusAhlberg: Indeed the way the standard is set is very silly in this case. The author of that must have forgotten logic at home to allow specifying 1 to N initializers instead of 0 to N or, looking from the opposite direction omitting 0 to N-1 initializers instead of 0 to N. GCC does a great job of not enforcing it.
-
Pavel Šimerda almost 9 years@torek: Not supporting empty objects (because you can't allocate and address them) and setting an artificial limit to omit only up to N-1 initializers instead of N resulting in absence of universal zero initializer are two very different things.
-
torek almost 9 years@PavelŠimerda: I'm confused by your comment: I didn't say empty initializers were not useful, just that they're not required to exist. If empty objects existed, then either they would need a special case ruling out initializers entirely (e.g., "empty objects may not be initialized") or they would require empty initializers. For that matter, allocating and addressing empty objects works fine in nonstandard GNU C, it just means that
&a==&b
is possible. -
undercat about 3 yearsThe assembly code for Clang looks identical as well, perhaps it's something that has changed recently.