Login
Username:

Password:

Remember me



Lost Password?

Register now!
Main Menu
Who is Online
7 user(s) are online (7 user(s) are browsing Forum)

Members: 0
Guests: 7

more...


Browsing this Thread:   1 Anonymous Users




« 1 (2) 3 4 5 »


Re: Error linking binary - Milkytracker for AROS ABI-v1
#39
Home away from home
Home away from home


See User information
what i mean is that a better option is if forks are not spread all around the world for users to investigate and thereore would be best to keep it common where possible. but thx. looking into it.

Posted on: 2018/5/12 8:37
 Top  Twitter  Facebook  Google Plus  Linkedin  Del.icio.us  Digg  Reddit  Mr. Wong 


Re: Error linking binary - Milkytracker for AROS ABI-v1
#38
Home away from home
Home away from home


See User information
@wawa:
nothing special there as it is the way to cooperate using git. see also this gist.

Please don't confuse a git fork with svn forking. They are named the same but are different things in practice.

Posted on: 2018/5/11 22:37
 Top  Twitter  Facebook  Google Plus  Linkedin  Del.icio.us  Digg  Reddit  Mr. Wong 


Re: Error linking binary - Milkytracker for AROS ABI-v1
#37
Home away from home
Home away from home


See User information
after all why not. i might fork the repo. there might be better ways of cooperation and keeping it all in sync but whatever.

Posted on: 2018/5/11 20:04
 Top  Twitter  Facebook  Google Plus  Linkedin  Del.icio.us  Digg  Reddit  Mr. Wong 


Re: Error linking binary - Milkytracker for AROS ABI-v1
#36
Home away from home
Home away from home


See User information
Quote:

Marlon wrote:

And also nested ones. it'll certainly break for an AROS compile if you don't define AMIGA and __AMIGA__ alongside AROS and __AROS__.

oh man, have you red my post? no problem here. i have not added AROS ifdefs to any of your source files. (yes i ifdeffed some pragmas for gcc, as proposal) and the very few changes i did to the source itself was in order to avoid some errors. it is all just a proposal on my part, for your review.

Quote:

I suggest you make a fork of the code, push it to your repo and do a pull request to the AmigaPorts repo. It's more work using the patch-files.


why using the patch files? why forking if you can just merge it in? why not use cmake if it is used upstream? it adds up to maintenance effort.

you asked me to write an aros makefile for the source and i did. whats wrong now?

Posted on: 2018/5/11 19:04
 Top  Twitter  Facebook  Google Plus  Linkedin  Del.icio.us  Digg  Reddit  Mr. Wong 


Re: Error linking binary - Milkytracker for AROS ABI-v1
#35
Home away from home
Home away from home


See User information
Quote:

Marlon wrote:
All I'm saying is that I've personally done no changes to CMakeList.txt to accommodate for any of my changes.


yes. i noticed. im not discarding your changes, im building upon that. again, the changes to cmakefiles are very slight and should affect only AROS so far. however you can make them affect amiga if you add AMIGA to the if sections "if(AROS OR AMIGA)". all ths is rather easy.

Quote:

And even if you compile for AROS you need to define AMIGA (same goes for os4, morphos, warpos etc).

And my premake code accommodates for all that.


i expect that. AROS build system defineds AMIGA as well i think. you do not need to do that explicitely, no problem here.

you can continue to use premake, but i have no idea of it and i frear it will not cope with building with aros build system and meta makefiles. therefore i added changes to lists. imho its easier, but take it or leave it. you could simply review my diff. or apply it locally to test, you can always revert.

Posted on: 2018/5/11 18:56
 Top  Twitter  Facebook  Google Plus  Linkedin  Del.icio.us  Digg  Reddit  Mr. Wong 


Re: Error linking binary - Milkytracker for AROS ABI-v1
#34
Not too shy to talk
Not too shy to talk


See User information
Quote:

wawa wrote:
Quote:

Marlon wrote:
All Amiga based fixes in my Milkytracker source falls under the define __AMIGA__ and AMIGA. And I add define AROS and __AROS__ to that for AROS specific code.

One does not exclude the other.


in cmake files its AROS, like in "if(AROS)" otherwise you are right. we can probably simplify joining AROS and AMIGA where possible. i didnt want to change too much at this point. the changes in cmakelists shouldnt affect your way of compiling. but isnt that more convenient to simply use cmake with my changes and have it compile in one go?

My long term plan is to sort out the CMake support for all platforms. My initial fixes on Milkytracker was under the __AMIGA__ define. However as more work was made there's lines of code like:
#if defined(__AMIGA__) && !defined(AROS)

And also nested ones. it'll certainly break for an AROS compile if you don't define AMIGA and __AMIGA__ alongside AROS and __AROS__.

Quote:

configuration "i386-aros"
buildoptions "-O3 -fomit-frame-pointer -fno-exceptions -static-libstdc++ -s"
linkoptions { "-lpthread", "-static-libstdc++" }
includedirs { "/opt/x86-aros/bin/linux-i386/AROS/Extras/Developer/include", "/opt/x86-aros/bin/linux-x86_64/tools/crosstools/lib/gcc/i386-aros/6.3.0/include-fixed", "/opt/aros/sdk/include/SDL", "./src/ppui/osinterface/amiga", "./src/ppui/osinterface/sdl-1.2", "./src/ppui/osinterface/posix" }
libdirs { "" }
defines { "AMIGA", "__AMIGA__", "AROS", "aros", "__AROS__", "__aros__" }


I suggest you make a fork of the code, push it to your repo and do a pull request to the AmigaPorts repo. It's more work using the patch-files.

Posted on: 2018/5/11 18:55
 Top  Twitter  Facebook  Google Plus  Linkedin  Del.icio.us  Digg  Reddit  Mr. Wong 


Re: Error linking binary - Milkytracker for AROS ABI-v1
#33
Home away from home
Home away from home


See User information
Quote:

Marlon wrote:
All Amiga based fixes in my Milkytracker source falls under the define __AMIGA__ and AMIGA. And I add define AROS and __AROS__ to that for AROS specific code.

One does not exclude the other.


in cmake files its AROS, like in "if(AROS)" otherwise you are right. we can probably simplify joining AROS and AMIGA where possible. i didnt want to change too much at this point. the changes in cmakelists shouldnt affect your way of compiling. but isnt that more convenient to simply use cmake with my changes and have it compile in one go?

Posted on: 2018/5/11 18:46
 Top  Twitter  Facebook  Google Plus  Linkedin  Del.icio.us  Digg  Reddit  Mr. Wong 


Re: Error linking binary - Milkytracker for AROS ABI-v1
#32
Not too shy to talk
Not too shy to talk


See User information
Quote:

wawa wrote:
Quote:

Marlon wrote:
And there's a reason I use premake instead of cmake... Cmake relies in SDL2. I made lots of edits to the source to keep the SDL1.2 compatibility, and CMake does not account for that at all.


not necessarily. look at my cmakelist.txt changes in main directory. you can disable sdl2 check. you jjust need to provide the link liib when necessary. also look at the other cmakelists changes around the source. im definitely including your sdl-1.2 headers and linking against aros sdl 1.2 library. there is no 2.0.

yes i have noticed the ahi code is based on sdl and not ready. im using the sdl of course.


All I'm saying is that I've personally done no changes to CMakeList.txt to accommodate for any of my changes. And even if you compile for AROS you need to define AMIGA (same goes for os4, morphos, warpos etc).

And my premake code accommodates for all that.

Posted on: 2018/5/11 18:44
 Top  Twitter  Facebook  Google Plus  Linkedin  Del.icio.us  Digg  Reddit  Mr. Wong 


Re: Error linking binary - Milkytracker for AROS ABI-v1
#31
Home away from home
Home away from home


See User information
Quote:

Marlon wrote:
And there's a reason I use premake instead of cmake... Cmake relies in SDL2. I made lots of edits to the source to keep the SDL1.2 compatibility, and CMake does not account for that at all.


not necessarily. look at my cmakelist.txt changes in main directory. you can disable sdl2 check. you jjust need to provide the link liib when necessary. also look at the other cmakelists changes around the source. im definitely including your sdl-1.2 headers and linking against aros sdl 1.2 library. there is no 2.0.

yes i have noticed the ahi code is based on sdl and not ready. im using the sdl of course.

Posted on: 2018/5/11 18:41
 Top  Twitter  Facebook  Google Plus  Linkedin  Del.icio.us  Digg  Reddit  Mr. Wong 


Re: Error linking binary - Milkytracker for AROS ABI-v1
#30
Not too shy to talk
Not too shy to talk


See User information
All Amiga based fixes in my Milkytracker source falls under the define __AMIGA__ and AMIGA. And I add define AROS and __AROS__ to that for AROS specific code.

One does not exclude the other.

Posted on: 2018/5/11 18:16
 Top  Twitter  Facebook  Google Plus  Linkedin  Del.icio.us  Digg  Reddit  Mr. Wong 




« 1 (2) 3 4 5 »



You can view topic.
You cannot start a new topic.
You cannot reply to posts.
You cannot edit your posts.
You cannot delete your posts.
You cannot add new polls.
You cannot vote in polls.
You cannot attach files to posts.
You cannot post without approval.
You cannot use topic type.
You cannot use HTML syntax.
You cannot use signature.
You cannot create PDF files.
You cannot get print page.

[Advanced Search]


Search
Top Posters
1 paolone
paolone
4462
2 nikolaos
nikolaos
4206
3 magorium
magorium
4095
4 phoenixkonsole
phoenixkonsole
3942
5 deadwood
deadwood
2917
6 ncafferkey
ncafferkey
2810
7 mazze
mazze
2222
8 Kalamatee
Kalamatee
2212
9 clusteruk
clusteruk
2114
Powered by XOOPS © 2001-2025 The XOOPS Project