Depend on empy python module

[srevinsaju]

Fix makefile dependency on /usr/bin/empy for Linux distributions
which do not package a empy binary executable.

[srevinsaju]

@chimosky @quozl

[quozl]

Thanks. Reviewed. Good progress, but unfinished. Please change the remaining references to $(EMPY).

[srevinsaju]

@quozl fixed it. No more EMPY remains

[quozl]

@srevinsaju, thanks. Reviewed.

• test on Fedora, thanks @chimosky.

[quozl]

Build stalls because python is reading from terminal. Reproducer;

$ cd gtk3/theme
$ python -m em - -p $ -D scaling=72 -D gtk=3.24.5 gtk-widgets.css.em

[quozl]

Hey, @chimosky, looks like we may have lost forward motion on this. Would you be able to fix it and test on Fedora? In removing < $(EMPY) the standalone - should also be removed. I'll test on Debian and Ubuntu as part of release.

[quozl]

On Ubuntu 19.04, tested 7a4935f; from a clean repository, with autogen, make, make install, and start Sugar.

What of test result on Fedora?

[chimosky]

Tested, builds as expected.

Logs

• When empy isn't installed.

There's a downstream issue but it doesn't affect this.

[quozl]

That's odd. Your logs don't match this pull request; the output "checking for empy..." and "checking for empy3..." should have been removed as a result of AC_PATH_PROGS being removed. Could you please check again that you are testing with 7a4935f?

[srevinsaju]

I am extremely sorry, I am away for a few days from programming. I will get back to this PR on Friday / Saturday.

I assume there was a hyphen character between em and the -p on the sugar PR too. I guess that should have also failed. This did not fail particularly, but I didn't test this line explicitly

python -m em - -p $ -D scaling=72 -D gtk=3.24.5 gtk-widgets.css.em

I don't have a desktop to test this. Seems like bash is too vast,

[chimosky]

That's odd. Your logs don't match this pull request; the output "checking for empy..." and "checking for empy3..." should have been removed as a result of AC_PATH_PROGS being removed. Could you please check again that you are testing with 7a4935f?

Yeah I was testing with 7a4935f.

Tested again; builds as expected.

Logs;

• When empy is installed.
• When empy isn't installed.

[chimosky]

I am extremely sorry, I am away for a few days from programming. I will get back to this PR on Friday / Saturday.

It's okay, it's been fixed.

I assume there was a hyphen character between em and the -p on the sugar PR too. I guess that should have also failed. This did not fail particularly, but I didn't test this line explicitly

There's quite a few differences between both changes.

[quozl]

Thanks for testing. I've fixed it in Sugar again, as make by itself outside any packaging wrapping did hang. sugarlabs/sugar@5d5c001

[quozl]

74d983e merged by hand.

[srevinsaju]

Ok, I came back to continue my work on this PR, and found out this PR's closed.
Just for my knowledge, is the - sign force empy to read from the terminal?
I can see that when - was used with empy and python3 -m` has the same effect.
I may be wrong, but it might help me in future. Thanks

[quozl]

Sure. - as an option means to read from stdin. Unless stdin has been redirected, that means the stdin of the parent shell, such as bash. This will usually be a terminal, unless make is being run by some other script or program. In removing the stdin redirect < $(EMPY) it is also necessary to remove the - that causes it to be used. It would have been obvious in testing using make, because it would stall waiting for an end of file mark from the terminal, which is a ctrl+d in Linux terminals.