Claudio André

You will have to create an account on Travis in order to enable it. But, it is very easy, no more than 3 mouse clicks needed, and you'll use your...

Also emit warnings for `--enable-openmp-for-fast-formats --enable-simd=avx2`

> Apparently @claudioandre-br changed it in 462f2b4 so it can return -1 in some other situations Reading the commit message, I would say it was bailing out with error on...

Thank you @witee-star. Please, create a PR to add V4 samples to https://github.com/openwall/john-samples repository. Solar will commit these changes in the john's main repository sooner or later.

Solar added something else at https://github.com/openwall/john/issues/5078#issuecomment-1047177149

> ideally we'd have *2john tools tested in CI, which we currently don't We have something, but there are limitations: 1. computing power (we saw in a commit today that...

Maybe the `2john` tool does not handle all existing files types. See also #4558. Please try to create a **demo "2013 wallet" file** that you can share so a developer...

Thinking out loud: I would say Nvidia did what they did for some reason; well, since a "theoretical" fix seems easy. Perhaps on some hardware it is hard to "sleep...

> Instead of clFinish, ideally we'd wait for completion of only the transfer before return from crypt_all, but of the kernel before starting the next call's transfer. How to do...

> It's not that simple. In the code I posted, clFinish was invoked after clEnqueueReadBuffer of the previous call's kernel invocation's computation result and a new clEnqueueNDRangeKernel, so the transfer...