+1

The lack of code documentation is not a free ticket to just change the code behaviour without proper analysis.

The right process should be

1. discuss

2. provide a patch

3. let others review/comment

4. decide

5. commit

It is really dangerous to easily change code like this.

Jacques, please be not so hasty with committing stuff. We have had a lot of similar cases with reverts, committing half done solutions and such lately. And please be aware that others might not have so much time to follow every commit in detail, analyze and comment promptly.

It really worries me because we lose quality and it's not easy to detect errors and changed functionality in such a complex project. And don't rely too much on the tests as we don't have such a high test coverage.

Thanks for some more patience,

Michael


Am 24.03.17 um 14:13 schrieb Jacopo Cappellato:
On Fri, Mar 24, 2017 at 10:56 AM, Jacques Le Roux <
[email protected]> wrote:

[...]
If we (both and All) agree on collaborating to document on purpose
swallowed exceptions, even when you are not directly concerned, then I
agree to revert my changes, deal?

We are not negotiating: I have simply asked you to revert the changes in
which you have changed the functional behavior of the system without
testing OR test the new behavior and confirm it is working fine.

In general I like the effort of improving this old code containing
swallowed exceptions by providing more comments, documentation etc... or
completely refactoring it; but this has to be done with proper testing.

I hope this clarifies my request.

Jacopo



Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to