local-time-multiplier

[mzuenni]

This adds the possibility to set local_time_multiplier in local config files. This might be useful if hardcoded time limits are set, but you are working on a machine which is much faster/slower than the machine used for judging.

I don't know if we really want to add such a feature though @RagnarGrootKoerkamp @mpsijm ?

[RagnarGrootKoerkamp]

Could you also add some docs on that this feature exists and can be enabled via local_time_multiplier in .bapctools.yaml in the contest dir?

[RagnarGrootKoerkamp]

This just needs some docs (specifically an instruction where/how to add it to either repo or personal config), and then it's probably good to merge?

[mzuenni]

should probably be merged after the typing thing... and may needs more input if this is really something we want?

[RagnarGrootKoerkamp]

At BAPC someone said that clang on mac has slow iostream, so they wanted to correct for that. A local language-specific multiplier would be more precise, but this could be a first approximation.

[mzuenni]

A local language-specific multiplier would be more precise

Yes but i don't think we want that? I mean we don't allow language specific multipliers and i think that is a good.

[RagnarGrootKoerkamp]

Agreed, but then it might be nice to at least merge this PR :)

[mpsijm]

My 2 cents:

• Agreed that this needs documentation, with warnings (because this could be abused as well).

• don't know if we want this

  This feature is not in the way of people who don't use it, so I'm fine with merging. I may even use this in CI after adjusting the time limits after a system test on contest hardware, if the CI is significantly slower/faster than the contest hardware.

• Adding language-specific local multipliers feels like a can of worms. The clang thing sounds to me like a bug that should be fixed (either by fiddling with compiler parameters, or upstream in Clang, or by switching to g++ to avoid comparing apples to pears).