Makefile change to responsefiles: 1 entry per line

[KageKirin]

Hi,
This is a small change to the gmake responsefile creation to have 1 entry per line, instead of 1 single line having space-separated entries.

Main reason: 1 of my projects has 2000+ source files, and exploded the command line limits when trying to echo all the object names into the responsefile (this happended on Mac).

Cheers.

CC: @rhoot for review.

@rhoot: Btw, you were right in the previous PR: there's no need for extra printing the object files/libs going into the responsefile since they'll be printed anyway in verbose mode, and the output is unwanted in silent mode.

[bkaradzic]

Terminate macro shell calls by semicolon

Why is this needed?

[KageKirin]

Terminate macro shell calls by semicolon

Why is this needed?

$(foreach) chaining of copy, mkdir etc. macros.
i.e. in the case of having a pre/postbuildcommand iterate over a number files to copy with $(foreach), the code will look like this:

without the added semicolon:

macro = $(SILENT) echo "$(1)"
$(foreach v,$(inputs),$(call macro,$(v));)

and now with the trailing semicolon in the macro definition:

macro = $(SILENT) echo "$(1)";
$(foreach v,$(inputs),$(call macro,$(v)))

IMHO, the 2nd one is easier to read and less confusing.
Also, I added the semicolon to have the regular macros be similar to the newly added FPRINT macro, i.e. with trailing semicolon.

I hope this answer is kind of satisfactory. The long story is that I was blocked by the lack of trailing semicolon for a while.

EDIT: I agree that 41ab1a1 does not really need to be part of this PR, and I can remove it if required.
My main reason for adding it was to have the same, semicolon-trailing form for all macros defined in the makefile, in order to avoid someone removing it from FPRINT by mistake b/c it looks different.

[rhoot]

This does not work on Windows. The object response file for genie (after adding the flag) turns out to be this:

"" ;
"obj/Release/src/host/lua-5.3.0/src/lapi.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lauxlib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lbaselib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lbitlib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lcode.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lcorolib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lctype.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/ldblib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/ldebug.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/ldo.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/ldump.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lfunc.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lgc.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/linit.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/liolib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/llex.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lmathlib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lmem.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/loadlib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lobject.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lopcodes.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/loslib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lparser.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lstate.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lstring.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lstrlib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/ltable.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/ltablib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/ltm.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lundump.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lutf8lib.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lvm.o" ;  echo "obj/Release/src/host/lua-5.3.0/src/lzio.o" ;  echo "obj/Release/src/host/os_chdir.o" ;  echo "obj/Release/src/host/os_copyfile.o" ;  echo "obj/Release/src/host/os_getcwd.o" ;  echo "obj/Release/src/host/os_is64bit.o" ;  echo "obj/Release/src/host/os_isdir.o" ;  echo "obj/Release/src/host/os_isfile.o" ;  echo "obj/Release/src/host/os_match.o" ;  echo "obj/Release/src/host/os_mkdir.o" ;  echo "obj/Release/src/host/os_pathsearch.o" ;  echo "obj/Release/src/host/os_rmdir.o" ;  echo "obj/Release/src/host/os_stat.o" ;  echo "obj/Release/src/host/os_ticks.o" ;  echo "obj/Release/src/host/os_uuid.o" ;  echo "obj/Release/src/host/path_getabsolute.o" ;  echo "obj/Release/src/host/path_getrelative.o" ;  echo "obj/Release/src/host/path_helpers.o" ;  echo "obj/Release/src/host/path_isabsolute.o"
;  echo "obj/Release/src/host/premake.o" ;  echo "obj/Release/src/host/premake_main.o" ;  echo "obj/Release/src/host/scripts.o" ;  echo "obj/Release/src/host/string_endswith.o" ;  echo "obj/Release/src/host/string_hash.o" ;

[KageKirin]

Thanks for reviewing, and sorry wasting your time with a broken commit.
I just remembered that when I tested it on Windows, I ran make from a bash shell on Windows, not a Windows Command prompt... my bad for not being thourough.

Anyway, I have an idea for a fix that I will implement tomorrow, and then test on Windows (DOS shell and Powershell) as well.
I'll update this PR after testing.

[KageKirin]

@rhoot @bkaradzic
Ok, so I fixed the issue and completely changed the PR branch's history.

I also removed 41ab1a1 b/c it did not make any sense after my fixes.

The FPRINT macro (daaad72) is now a $(shell) call, with the net benefit of being executed once per $(foreach ) iteration.
And the Windows variant of FPRINT consists in echo.$(1), which with $(1) empty results in just an empty line.

Please not that I have a follow up commit already prepared here:
https://github.comKageKirin/GENie/tree/makefile-wholearchive-attfile
If you prefer, I can also close this PR and submit the for the branch above.

[KageKirin]

Ok, closing this one in favor just merging both changes with #458 instead.