Possible GNU Make 4.3 bug

[Eldred HABERT]

Good evening,

I found what seems like a bug, or at least an odd behavior, in GNU Make 4.3.

Attached is a ZIP file from a project that was sent to me, which seems to have trouble compiling with GNU Make, both on his Windows machine (which I'm assuming is using GnuWin32's Make 3.81) *and* on my Arch Linux box, whose Make is currently at 4.3.

To build this, you will need RGBDS (v0.4.2 was used by both of us, and reproduces the bug consistently) and Python 3; it may also be necessary to edit line 21 of the Makefile to your distribution's Python 3 binary. (The files were created on Windows, and I've had trouble with CRLF.)

The bug is the following two lines:

   Must remake target 'res/tpnk_dev_logo_tiles.2bpp.pb16.size'.
   Successfully remade target file 'res/tpnk_dev_logo_tiles.2bpp.pb16.size'.

(from `make -d`, with `-r` added to make the output easier to read, though it doesn't seem to impact anything).

It seems that Make does *not* run the recipe for that file at all (I even tried adding `false` to that rule, but this didn't cause failure). `mape -p` also prints matching information. I am not aware of anything that could cause this behavior, so I'm assuming this is a bug.

Please feel free to ask any questions related to this setup. Note that I am RGBDS' main maintainer, so I am also able to answer questions regarding that, though the Make interface (`-M` etc.) was modeled after GCC's, so it should be familiar.

Regards,
~ Eldred HABERT

shiba-beach-volleyball.zip
Description: Zip archive