| 1 |
On 06/27/2010 08:14 PM, Harald van Dijk wrote: |
| 2 |
> On Sun, Jun 27, 2010 at 05:46:28PM +0300, Nikos Chantziaras wrote: |
| 3 |
>> On 06/27/2010 03:23 PM, Harald van Dijk wrote: |
| 4 |
>>> The compiler is not totally free to ignore the register keyword. |
| 5 |
>>> Both the C and the C++ standards require that the compiler complain |
| 6 |
>>> when taking the address of a register variable. Other compilers will |
| 7 |
>>> issue a hard error for it. Fixing the code to not declare the |
| 8 |
>>> variable as register would be the correct thing to do. |
| 9 |
>> |
| 10 |
>> No, it would not be the correct thing to do, because of the following. |
| 11 |
>> (This is part of a discussion between me and someone quite smarter than |
| 12 |
>> me, who explained the issue in detail.) |
| 13 |
>> |
| 14 |
>> [snip] |
| 15 |
> |
| 16 |
> That explanation seems to be written by someone who does not know that |
| 17 |
> taking the address of a register variable is simply not allowed. |
| 18 |
|
| 19 |
It is allowed. Section 7.1.1, Paragraphs 2 and 3 of the C++ standard: |
| 20 |
|
| 21 |
The register specifier [...] specifies that the named variable has |
| 22 |
automatic storage duration (3.7.3). A variable declared without a |
| 23 |
storage-class-specifier at block scope or declared as a function |
| 24 |
parameter has automatic storage duration by default. |
| 25 |
|
| 26 |
A register specifier is a hint to the implementation that the variable |
| 27 |
so declared will be heavily used. [ Note: the hint can be ignored and in |
| 28 |
most implementations it will be ignored if the address of the variable |
| 29 |
is taken. This use is deprecated (see D.4). — end note ] |
| 30 |
|
| 31 |
|
| 32 |
>> OK, long read, but the the conclusion is that "fixing the code to not |
| 33 |
>> declare the variable as register would be the correct thing to do" it |
| 34 |
>> *not* the correct thing to do. The correct thing to do is to ignore the |
| 35 |
>> warning, which is not possible if warnings are turned into errors. |
| 36 |
> |
| 37 |
> And which is not possible if the warning is a hard error in the first place. |
| 38 |
> |
| 39 |
>> You also mentioned that "other compilers will issue a hard error for |
| 40 |
>> it." That sounds rather strange, and I wonder which compilers that |
| 41 |
>> might be; someone should file a bug report against them ;) |
| 42 |
> |
| 43 |
> Well, let's start with gcc; that's quite an important one for Gentoo... |
| 44 |
|
| 45 |
That code compiles just fine. I don't know of any GCC version that |
| 46 |
issues an error for this rather than just a warning. |
Archives