On 27/11/19 7:29 am, Ethan Sommer wrote: > Previously, to determine which command we should use to extract an > archive, we would run file and match the output against our list of > possible extraction commands > > Instead, run the archive through each extraction command's -t (--test) > flag, if this succeeds then we know that the command is able to extract > the file and is the one to use >
Apart from the issues Eli brought up, there is a large performance regression on large source files. $ time file -bizL libreoffice-6.3.4.1.tar.xz application/x-tar; charset=binary compressed-encoding=application/x-xz; charset=binary real 0m0.034s user 0m0.031s sys 0m0.004s $ time xz -t libreoffice-6.3.4.1.tar.xz real 0m11.970s user 0m11.898s sys 0m0.057s Allan
