Nlnet funding for transitioning out of SHA-1

Pierre-Yves David pierre-yves.david at ens-lyon.org
Sun Jan 26 13:33:45 EST 2020



On 1/24/20 6:34 PM, Augie Fackler wrote:
> Honestly
> 
>> On Jan 24, 2020, at 11:29, Pierre-Yves David <pierre-yves.david at ens-lyon.org> wrote:
>>
>> On 1/24/20 2:37 PM, Raphaël Gomès wrote:
>>> On 1/16/20 9:38 PM, Augie Fackler wrote:
>>>> I think overall that looks pretty good. Are you planning on making a submission?
>>> Thanks for answer, sorry I took so long to get back to you.
>>> It seems like we agree on the broad strokes, so let's go ahead.
>>
>> The more I am thinking about it (helping Raphaël with the document), the more I thinking that one of the most major roadblock we need to deal with is the test suite. We have hundreds of thousand of line of test file all relying on specific hash to be produced and consumed. We will need to design a good way to express the used of different hash function in tests. Otherwise we won't be able to seriously test the various hash using options.
> 
> Honestly I think I'd start by having a few small tests that exercise the basis with blake2b, then at some point we can set some tests to blake2b and some to sha1 and that'll be fine.

Yeah, starting with a smaller state to get the basic to run is probably 
a good approach. However, soon enought we will need to exercise the full 
test suite.

-- 
Pierre-Yves David


More information about the Mercurial-devel mailing list