Missing dependency (or other quirk?) in Makefile

[llimeht]

In an effort to shake out Makefile errors that can be exposed by parallel builds on multiprocessor systems, make's new --shuffle options are being explored - these reorder the targets within the Makefile which should be a no-op if everything is good, but can cause a failure to build from source if there are missing dependencies or other quirks.

My Debian colleagues built with --shuffle=reverse (that reorders the targets in the reverse order) to compromise on at least a reproducible reordering. In the test, pyxplot fails to build: Debian bug #1105638 and build log.

I suspect that the relevant line in the Makefile (L 178) should simply not be there - there should be no need to make a note about the debug builds being slow when creating the non-debug build. If that's the case, then it's the easiest possible fix :)

I'm not sure why this bug isn't exposed in the regular builds. Presumably something makes the bin/debug directory earlier in the normal builds.

[dcf21]

Hi Stuart,

Great to hear from you again - I hope you are well!

Well spotted - There's a typo here that places the README file in the wrong directory. I think this should be fixed by commit 38c7041. I hope it builds now.

Best wishes,

Dominic