[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: Proposal for deviations in static analyser findings
On 18.10.2022 17:49, Luca Fancellu wrote: >> On 18 Oct 2022, at 16:29, Jan Beulich <jbeulich@xxxxxxxx> wrote: >> On 18.10.2022 17:17, Luca Fancellu wrote: >>>> On 13 Oct 2022, at 12:34, Jan Beulich <jbeulich@xxxxxxxx> wrote: >>>> On 13.10.2022 12:11, Luca Fancellu wrote: >>>>>> On 13 Oct 2022, at 08:50, Jan Beulich <jbeulich@xxxxxxxx> wrote: >>>>>> Iirc at least Coverity ignores certain instances of what it might >>>>>> consider >>>>>> violations (fall-through in switch() statements in particular) in case >>>>>> _any_ comment is present. Therefore may I suggest that such comments be >>>>>> deleted (really: replaced by a blank line, to maintain correct line >>>>>> numbering) if there's no matching key-value pair? >>>>> >>>>> Yes the line won’t be altered if there is no match. This to ensure the >>>>> correct line >>>>> numbering is not affected. >>>> >>>> "won't be altered" is the opposite of what I've been asking to consider: >>>> Observing that comments _regardless_ of their contents may silence >>>> findings, >>>> the suggestion is to remove comments (leaving a blank line) when there's no >>>> entry for the targeted tool in the table entry. >>> >>> Why? The tag comment won’t do anything, it would act as a blank line from >>> the analyser >>> perspective. >> >> The _tag_ won't do anything, but as said any _comment_ may have an effect. > > Yes, any comment that is using a proprietary syntax for the tools we use: > > /* cppcheck-suppress[proprietary_ID] */ > /* coverity[proprietary_ID] */ > /* -E> hide proprietary_ID 1 “" */ > > May have an effect. > > If an entry in the database has no match with the used tool, then it would > stay as (for example): > > /* SAF-X-safe [blablabla] */ > > Which has no effect on any tool, hence I don’t see the needs to replace it > with a blank line. One more try: May I ask that you re-read the very beginning of the context I've left in place? I did call out a case where, from all I know, a tool would be silenced by the mere presence of a comment, regardless of comment text. Hence by not purging SAF-X-... comments you may suppress findings. Jan
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |