Message84600
| Author |
stutzbach |
| Recipients |
alexandre.vassalotti, donmez, stutzbach |
| Date |
2009年03月30日.18:09:43 |
| SpamBayes Score |
2.835266e-07 |
| Marked as misclassified |
No |
| Message-id |
<1238436588.35.0.0314902769013.issue3326@psf.upfronthosting.co.za> |
| In-reply-to |
| Content |
I'm using gcc 3.4.4 (cygwin) and I get the sames warnings as Alexandre.
I examined a random sampling of the code generating the warnings, all of
which followed this pattern:
some_function((some_type **) &var_of_some_other_type);
Since the variable isn't actually being dereferenced in the calling
function, the code isn't violating the strict aliasing rules. I guess
gcc 4.3 is smart enough to suppress the warning in cases like this.
We could make the warning go away by replacing "(some_type **)" with
"(void *)", though that perhaps decreases readability. Alternately we
can just disregard the warnings on older versions of gcc. |
|
History
|
|---|
| Date |
User |
Action |
Args |
| 2009年03月30日 18:09:48 | stutzbach | set | recipients:
+ stutzbach, alexandre.vassalotti, donmez |
| 2009年03月30日 18:09:48 | stutzbach | set | messageid: <1238436588.35.0.0314902769013.issue3326@psf.upfronthosting.co.za> |
| 2009年03月30日 18:09:44 | stutzbach | link | issue3326 messages |
| 2009年03月30日 18:09:43 | stutzbach | create |
|