Re: [Valgrind-developers] Proposed valgrind division patch

[Paul F. <pj...@wa...>]

On 26-08-24 14:29, Mark Wielaard wrote:
> Hi,

I also agree that this sounds very interesting. I quickly skimmed over 
the paper. I haven't had time to  look at the code yet.

Do you have any feeling as to how many people would make use of this?

>> The patch is designed to be off by default. The user can start scanning
>> for divisions using --variable-latency-errors=yes on the command line,
>> VALGRIND_CLO_CHANGE("--variable-latency-errors=yes") from the program
>> under test, or, easiest to use, a new environment variable
>> VALGRIND_BESTEFFORT_VARIABLE_LATENCY_ERRORS=yes.
> 
> The environment variable trick is nice. I wonder if we can generalize
> that.

I'm not keen on more environment variables. We already have 3 mechanisms 
for passing options (command line, .valgrindrc and VALGRIND_OPTS). 
There's plenty to confuse users already (which options are command line 
only, what is the precedence between .valgrindrs and VALGRIND_OPTS). It 
would just confuse matters to have more environment variables.

Using VALGRIND_OPTS is clunky but I think that we should stick to that.

Looking at the competition, Address Sanitizer just uses ASAN_OPTIONS.

A+
Paul