Duke4.net Forums: EDuke32 / Mapster32 Compilation Issues - Duke4.net Forums

Jump to content

  • 2 Pages +
  • 1
  • 2
  • You cannot start a new topic
  • You cannot reply to this topic

EDuke32 / Mapster32 Compilation Issues

User is offline   LeoD 

  • Duke4.net topic/3513

#1

Compilation on Raspbian is broken as of r5445
Spoiler

r5448 introduces one more error
Spoiler

Compiler warnings from r5444:
Spoiler
Spoiler


This post has been edited by LeoD: 28 November 2015 - 04:39 PM

1

User is online   Hendricks266 

  • Weaponized Autism

  #2

Derp. I'll fix these now and test on my Pi 2 when I return home from holiday.
-1

User is online   Hendricks266 

  • Weaponized Autism

  #3

The errors should be fixed. I haven't touched the GTK warnings because it's something they did on their end that made it downstream when you updated libgtk, and I would need to (1) look up and implement what the non-"deprecated" way to do it would be, and (2) make the codepath conditional on the GTK version that introduces the new way, so that we don't break the build when using older versions.
1

User is offline   LeoD 

  • Duke4.net topic/3513

#4

View PostHendricks266, on 28 November 2015 - 01:57 AM, said:

The errors should be fixed. I haven't touched the GTK warnings because it's something they did on their end that made it downstream when you updated libgtk, and I would need to (1) look up and implement what the non-"deprecated" way to do it would be, and (2) make the codepath conditional on the GTK version that introduces the new way, so that we don't break the build when using older versions.
Fixed. Thanks for explaining.
0

User is online   Hendricks266 

  • Weaponized Autism

  #5

How have you configured your Pi to produce the gdk_pixbuf_from_pixdata warning? I ran sudo apt-get update and sudo apt-get upgrade and even then I get no warnings or errors, regardless of RELEASE, CPLUSPLUS, or USE_OPENGL*.

Maybe the Pi 2 is different from the Pi model you have.

*Besides an unused variable in menus.c with USE_OPENGL=0.
0

User is offline   LeoD 

  • Duke4.net topic/3513

#6

View PostHendricks266, on 29 November 2015 - 01:05 AM, said:

How have you configured your Pi to produce the gdk_pixbuf_from_pixdata warning? I ran sudo apt-get update and sudo apt-get upgrade and even then I get no warnings or errors, regardless of RELEASE, CPLUSPLUS, or USE_OPENGL*.
Seems to be the point to mention that I'm on testing, not stable. Otherwise I have no idea what type of settings/configuration I should look for.
1

User is online   Hendricks266 

  • Weaponized Autism

  #7

View PostLeoD, on 29 November 2015 - 05:42 AM, said:

Seems to be the point to mention that I'm on testing, not stable.

I bet that would do it. :)

I looked at the GTK documentation anyway.

Quote

gdk_pixbuf_from_pixdata has been deprecated since version 2.32 and should not be used in newly-written code.

Use GResource instead.

Not only that,

Quote

GdkPixdata is deprecated and should not be used in newly-written code.

Apparently GResource is the output of a new GTK program run during compilation in order to import images and text files into a program easily. It looks like a replacement for what we already run:

$(DUKE3D_RSRC)/%_banner.c: $(DUKE3D_RSRC)/%.bmp
	echo "#include \"gtkpixdata_shim.h\"" > $@
	gdk-pixbuf-csource --extern --struct --raw --name=startbanner_pixdata $^ | sed 's/load_inc//' >> $@


Since the code is only marked deprecated and isn't going to be removed, I'm not going to bother right now. I'll give it another look when I refactor the startup windows.

This post has been edited by Hendricks266: 30 November 2015 - 08:52 PM

0

User is offline   Steve 64 

#8

I notice that EDuke32 on Steam got an update what was in it?

This post has been edited by Demon Duke: 01 December 2015 - 01:02 AM

0

User is online   Hendricks266 

  • Weaponized Autism

  #9

Wrong thread, Steve.

Nothing. It may have had to do with recent troubles users are having with the Re-rockestrated Soundtrack. I only got an update to Alien Carnage.
0

User is offline   Steve 64 

#10

View PostHendricks266, on 01 December 2015 - 08:33 AM, said:

Wrong thread, Steve.

Nothing. It may have had to do with recent troubles users are having with the Re-rockestrated Soundtrack. I only got an update to Alien Carnage.


Sorry about that didn't know where to ask the question
1

User is offline   LeoD 

  • Duke4.net topic/3513

#11

Raspbian compilation broken:
Spoiler


RENDERTYPE=WIN warnings:
Spoiler


This post has been edited by LeoD: 11 January 2016 - 07:27 AM

2

User is online   Hendricks266 

  • Weaponized Autism

  #12

Should be fixed now.
1

User is offline   LeoD 

  • Duke4.net topic/3513

#13

make RENDERTYPE=WIN broken as of r5612:
Spoiler

0

User is offline   LeoD 

  • Duke4.net topic/3513

#14

View PostLeoD, on 08 February 2016 - 12:36 PM, said:

make RENDERTYPE=WIN broken as of r5612
Fixed.
Spoiler


Now, automatic synthesis builds seem to be broken. (r5627 should be available.) -> r5628

This post has been edited by LeoD: 14 February 2016 - 07:47 AM

0

User is offline   LeoD 

  • Duke4.net topic/3513

#15

make POLYMER=0 broken as of r5700:
Spoiler

1

User is online   Hendricks266 

  • Weaponized Autism

  #16

I already have a fix in my tree.
0

User is offline   Spiker 

#17

View PostHendricks266, on 23 April 2016 - 08:18 AM, said:

I already have a fix in my tree.
Anything else you have in your tree?
1

User is offline   sigpaw1 

#18

View PostHendricks266, on 23 April 2016 - 08:18 AM, said:

I already have a fix in my tree.


I just commented out both references to smoothratio since they didn't appear to be referenced anywhere (I preseume this may have been a variable supplied in the function parameter list some time in the past). ?
0

User is online   Hendricks266 

  • Weaponized Autism

  #19

The fix is pushed now.
0

User is offline   sigpaw1 

#20

View PostHendricks266, on 02 May 2016 - 01:24 PM, said:

The fix is pushed now.


Thansk! I'll go grab it...
0

User is offline   LeoD 

  • Duke4.net topic/3513

#21

MSYS2/MinGW building leaves a zero byte file named "NUL" which cannot be deleted by Windows (but from MSYS2). I think we had a related issue some time ago, caused by Windows' nasty DOS backward compatibility features.
0

User is online   Hendricks266 

  • Weaponized Autism

  #22

Hmm, that's supposed to get caught by this:

ifeq ($(HOSTPLATFORM),WINDOWS)
# MSYS2 lets you create files named NUL but has a /dev/null. Go figure.
    ifeq (,$(wildcard /dev/null))
        DONT_PRINT = > NUL 2>&1
    endif
endif

0

User is offline   LeoD 

  • Duke4.net topic/3513

#23

View PostHendricks266, on 04 May 2016 - 01:29 PM, said:

Hmm, that's supposed to get caught by this:
Took a while to figure out ... my build script does a better job than"'make veryclean". The first build after a clean checkout creates a couple of object file directories which aren't deleted by "make veryclean".
Spoiler
To reproduce the creation of "NUL", run:
> svn checkout http://svn.eduke32.com/eduke32/polymer/eduke32
> cd eduke32/
> make -j LTO=0 eduke32.exe


This post has been edited by LeoD: 04 May 2016 - 02:57 PM

0

User is online   Hendricks266 

  • Weaponized Autism

  #24

This should be fixed in the latest SVN.
0

User is offline   LeoD 

  • Duke4.net topic/3513

#25

View PostHendricks266, on 22 May 2016 - 08:57 PM, said:

This should be fixed in the latest SVN.
Nope (r5756). NUL is still created under MSYS2, though not under older MSYS (which I hadn't checked before).
0

User is online   Hendricks266 

  • Weaponized Autism

  #26

Very strange. I ran `pacman -Syu` on my old msys2 install before testing. Could you find the bit in Makefile.common that I quoted above, and before the "ifeq ($(HOSTPLATFORM),WINDOWS)", add the following and tell me what prints out?

$(warning $(OS), $(strip $(shell uname -s)), $(HOSTPLATFORM), $(PLATFORM), $(wildcard /dev/null))
0

User is offline   LeoD 

  • Duke4.net topic/3513

#27

View PostHendricks266, on 24 May 2016 - 03:53 PM, said:

Very strange. I ran `pacman -Syu` on my old msys2 install before testing.
Concerning older MSYS2: The 'correct' update procedure seems to have changed recently..

View PostHendricks266, on 24 May 2016 - 03:53 PM, said:

Could you find the bit in Makefile.common that I quoted above, and before the "ifeq ($(HOSTPLATFORM),WINDOWS)", add the following and tell me what prints out?
Makefile.common:301: Windows_NT, MINGW64_NT-6.1, WINDOWS, WINDOWS,
0

User is online   Hendricks266 

  • Weaponized Autism

  #28

Does `echo test >/dev/null` work as expected in the same shell from which you invoke make? If so, we may need a better way to detect its presence than $(wildcard)...
0

User is online   Hendricks266 

  • Weaponized Autism

  #29

I followed the instructions you linked to make sure my msys2 is fully updated, and in all three shells (MSYS, MINGW32, MINGW64) the "$(wildcard /dev/null)" returns "/dev/null" (instead of nothing, as you pasted). Does `echo test >/dev/null` work as you would expect? In other words, does piping something to /dev/null create a C:\msys64\dev\null file?
0

User is offline   LeoD 

  • Duke4.net topic/3513

#30

View PostHendricks266, on 28 May 2016 - 07:32 PM, said:

In other words, does piping something to /dev/null create a C:\msys64\dev\null file?
No, but /dev/null is already present:
Spoiler


This post has been edited by LeoD: 29 May 2016 - 08:48 AM

0

Share this topic:


  • 2 Pages +
  • 1
  • 2
  • You cannot start a new topic
  • You cannot reply to this topic


All copyrights and trademarks not owned by Voidpoint, LLC are the sole property of their respective owners. Play Ion Fury! ;) © Voidpoint, LLC

Enter your sign in name and password


Sign in options