| 1 |
On 06/27/2010 01:47 PM, Enrico Weigelt wrote: |
| 2 |
> * Nikos Chantziaras<realnc@×××××.de> schrieb: |
| 3 |
> |
| 4 |
>> Did it actually occur to anyone that warnings are not errors? You can |
| 5 |
>> have them for correct code. A warning means you might want to look at |
| 6 |
>> the code to check whether there's some real error there. It doesn't |
| 7 |
>> mean the code is broken. |
| 8 |
> |
| 9 |
> In my personal experience, most times a warning comes it, the |
| 10 |
> code *is* broken (but *might* work in most situations). |
| 11 |
|
| 12 |
That's the key to it: most times. Granted, without -Wall (or any other |
| 13 |
options that tweaks the default warning level) we can be very sure that |
| 14 |
the warning is the result of a mistake by the developer. But with |
| 15 |
-Wall, many warnings are totally not interesting ("unused parameter") |
| 16 |
and some even try to outsmart the programmer even though he/she knows |
| 17 |
better ("taking address of variable declared register"). In that last |
| 18 |
example, fixing it would even be wrong when you consider the optimizer |
| 19 |
and the fuzzy meaning of "register" which the compiler is totally free |
| 20 |
to ignore. |
Archives