|
Notes |
|
|
|
Patch to solve this issue (looks like somebody have mistaken '/' and '\'; '/' is standard path separator in C, and '\' is used only with printf and company):
--------------------------------------------------------------------------------
--- lib/CModHandler.cpp (revision 2897)
+++ lib/CModHandler.cpp (working copy)
@@ -1,8 +1,8 @@
#include "StdInc.h"
#include "CModHandler.h"
#include "JsonNode.h"
-#include "Filesystem\CResourceLoader.h"
-#include "Filesystem\ISimpleResourceLoader.h"
+#include "Filesystem/CResourceLoader.h"
+#include "Filesystem/ISimpleResourceLoader.h"
/*
* CModHandler.h, part of VCMI engine
* |
|
|
|
(0003022)
|
|
Ivan
|
|
2012-09-20 16:15
|
|
|
|
|
|
Another error:
make[2]: Entering directory `/home/username/Development/VCMI/vcmi-svn/client'
CXX vcmiclient-CCallback.o
CXX vcmiclient-CBattleAnimations.o
In file included from ./BattleInterface/CBattleAnimations.cpp:5:0:
./BattleInterface/../CMusicHandler.h:47:13: error: extra qualification ‘CSoundHandler::’ on member ‘GetSoundChunk’ [-fpermissive]
make[2]: *** [vcmiclient-CBattleAnimations.o] Error 1
Patch:
--- client/CMusicHandler.h (revision 2897)
+++ client/CMusicHandler.h (working copy)
@@ -44,7 +44,7 @@
std::map<soundBase::soundID, Mix_Chunk *> soundChunks;
Mix_Chunk *GetSoundChunk(soundBase::soundID soundID);
- Mix_Chunk *CSoundHandler::GetSoundChunk(std::string &sound);
+ Mix_Chunk *GetSoundChunk(std::string &sound);
//have entry for every currently active channel
//boost::function will be NULL if callback was not set |
|
|
|
|
> fixed, rev 2898
You are too fast :) |
|
|
|
|
GNU C++ compiler introduced an interesting feature in 4.1 version, released in May 2006: no longer you can compile code like
class a {
void a::hello(void) {}
};
Yes it is redundant and probably against C++ specs to redefine the a:: at function declaration.
----------------------------------------------------------------
^^^^ I dunno how such easy-to-find-just-compile-me gotchas can be made now... |
|
|
|
|
|
|
|
(0003027)
|
|
Ivan
|
2012-09-20 16:33
(edited on: 2012-09-20 16:36) |
|
Haven't I fixed that too? clang compiles fine, Haven't check gcc yet.
>> I dunno how such easy-to-find-just-compile-me gotchas can be made now
Because MSVC accepts that code and only I (from devs) use gcc?
|
|
|
|
|
> Haven't I fixed that too?
I dunno, I reported against 2897... With the patches I mentioned here, VCMI finishes its build.
> clang compiles fine
`clang' is so broken on my architecture...
> MSVC accepts that code and only I (from devs) use gcc
Oh, is that real? Are GCC, make, and autotools so hard to use?
Well, time to test http://bugs.vcmi.eu/view.php?id=1078 [^] (that's why I decided to build fresh checkout). |
|
|
|
(0003029)
|
|
Ivan
|
|
2012-09-20 16:53
|
|
If somebody reports compile errors I always check for any subsequent errors - no need to report them.
clang is faster so I use it most of time. Usually code that works in clang works in gcc and vice versa. I also have gcc-4.5, 4.6, 4.7 and snapshot to test them from time to time.
>> Oh, is that real? Are GCC, make, and autotools so hard to use?
People are too used with Win + Studio. Mingw toolchain was tested only a few days ago.
So this is more like classic Windows vs Linux rather than hard to use gcc or make. |
|
|
|
(0003030)
|
|
Ivan
|
|
2012-09-20 16:54
|
|
|
Oh. And back to resolved. |
|
|
|
|
> If somebody reports compile errors I always check for any subsequent errors - no need to report them.
Hmm, Warmonger asked me to "report only one issue at a time", but yes, I see that related errors shall be reported inside only one issue.
Well, I updated to rev2899. And it looks okay. |
|
|
|
(0003032)
|
|
Ivan
|
|
2012-09-20 17:05
|
|
|
That's for gameplay errors. For compile issues - just let me now that something is wrong. |
|
|
|
(0003039)
|
|
Tow
|
|
2012-09-20 19:48
|
|
> Oh, is that real? Are GCC, make, and autotools so hard to use?
Long story short: yes, they are.
I could write a thriller story about getting GDB pretty-printing to work on Windows. Or am I just spoiled? :P |
|