1 |
On Thu, Sep 13, 2018 at 7:20 PM Fabian Groffen <grobian@g.o> wrote: |
2 |
> > > To illustrate harmless: |
3 |
> > > warning: this statement may fall through [-Wimplicit-fallthrough=] |
4 |
> > > The warning message already has it in it that it's just a pure guess. |
5 |
> > |
6 |
> > One that exposed a lot of unintentional fallthoughs which were fixed |
7 |
> > when reporting to upstream. |
8 |
> |
9 |
> Sure that's why the warning is there. But you ignore the point that the |
10 |
> same code compiled fine and ran fine for years without problems. |
11 |
|
12 |
The fact that something is compiling and running fine meaning there |
13 |
are no issues (bugs) within code? |
14 |
Seriously? |
15 |
Even after no-warning with multiple compiler vendors, code coverage, |
16 |
unit testing, test on various of architecture developer has access to, |
17 |
static code analysis and running for years, bugs are there. Any method |
18 |
to help detect suspicious code, even if it produces amount of false |
19 |
positive, must be embraced of those who care about quality. New |
20 |
toolchains, new scanners, new architectures all can help to improve |
21 |
quality to make sure great service is provided to users. |
22 |
|
23 |
In Gentoo language, all these issues should be detected for selected |
24 |
packages by non-stable users, on architecture and permutations that |
25 |
upstream do not have access to, and to help upstream to filter false |
26 |
positives and find the positives ones. Even one case of funding real |
27 |
issue is sufficient to justify the maintenance costs, once again for |
28 |
selected packages in which upstream following strict quality policy |
29 |
and downstream follows. Once policy is applied, the amount of noise is |
30 |
very little, toolchain evolution is not as it was 10 years ago. |
31 |
|
32 |
Regards, |
33 |
Alon |