Fixes to 10 coverity defect reports #480
Conversation
merging with root master.
|
Could you resolve the conflicts? |
hist/hist/src/TH1Merger.cxx
Outdated
| for (Int_t i = 0; i < nbentries; i++) | ||
| h2->Fill(hist->fBuffer[3*i + 2], hist->fBuffer[3*i + 3],hist->fBuffer[3*i + 1] ); | ||
| if (auto h2 = dynamic_cast<TH2*>(fH0)){ | ||
| R__ASSERT(h2); |
There was a problem hiding this comment.
This seems like a false positive in coverity, i.e. if h2 is nullptr we will issue a fatal error and abort. The assert in the if-stmt here becomes dead code.
There was a problem hiding this comment.
Probably coverity does not know that you will never comeback from R__ASSERT with a nullptr.
There was a problem hiding this comment.
This is indeed very likely. Is there a way to teach coverity about the semantic of that macro?
There was a problem hiding this comment.
Coverity believes that ::Fatal() might return. As a matter of fact I can simply set gAbortLevel to kFatal + 1 and Fatal() will not abort! I.e. indeed and in principle: Coverity is correct here :-)
For all practical purposes though we should annotate Fatal(); see https://coverity.cern.ch/docs/en/cov_checker_ref.html#killpath
I'll take care of it.
There was a problem hiding this comment.
Told Coverity about Fatal() in 7c8286d - that should get rid if this "defect".
hist/hist/src/TH1Merger.cxx
Outdated
| for (Int_t i = 0; i < nbentries; i++) | ||
| h3->Fill(hist->fBuffer[4*i + 2], hist->fBuffer[4*i + 3],hist->fBuffer[4*i + 4], hist->fBuffer[4*i + 1] ); | ||
| if (auto h3 = dynamic_cast<TH3*>(fH0)){ | ||
| R__ASSERT(h3); |
This reverts commit c0cf732.
|
@phsft-bot build |
|
Starting build on |
Fix warning
|
@phsft-bot build |
|
Starting build on |
|
One failure of the system in one cell. |
| Float_t xcoord[n], ycoord[n]; | ||
| for ( double xi = 0; xi < 10.01; xi += 0.2) { | ||
|
|
||
| for ( i = 0; i < 50; ++i) { |
There was a problem hiding this comment.
Hm, was it intentional to increment i twice: once in the loop and once in the loop's body (see last line).
There was a problem hiding this comment.
I don't understand these changes too. The old code looked perfectly fine to me
7 participants
No description provided.