Dima Pasechnik

Results 290 comments of Dima Pasechnik
trafficstars

It's all quite messy. While talking about Debian 11's fplll, it's version 5.4.0, with FPLLL_DEFAULT_STRATEGY and FPLLL_DEFAULT_STRATEGY_PATH set to the correct path, and to `default.json`, respectively, in `fplll/fplll_config.h`. But on...

I'm trying to compile and run (with `g++ foo.C -lfplll && ./a.out`) ```C++ /* foo.C */ #include #include #include int main () { std::cout

It seems to be a good patch, why isn't it upstreamed? The current situation with the headers in 5.4.1 is confusing at best.

One way or another, I don't see the point of having these two macros set to `""` in the header. Either they should be removed there completely, or kept with...

> libfplll6 and libfplll7 are co-installable, and each can find its correct strategies file. yes, but FPLLL_DEFAULT_STRATEGY* macros will only be correct for one of them.

why are these macros needed in the header file at all?

From C++ it's all available via concatenating `fplll::default_strategy_path()` and `fplll::default_strategy()`. The values are baked in at build time.

> to fix this in Sage, I opened https://trac.sagemath.org/ticket/33302 it's ready for review, please check.

> is there anything else we should do? as long as you don't break what you have now regarding `BKZ.DEFAULT_STRATEGY_PATH` and `BKZ.DEFAULT_STRATEGY`, all is good and taken care of by...

handling of `default.json` was seriously messed up, by too eager changes done by fpylll devs to fix upstream a Sage issue. I recall untangling it a while ago.