Problems compiling on os x 10.9

unelg · October 24, 2013, 8:39pm

Checking for libXpm …
libraries to check for: libXpm.so libXpm.sl libXpm.dylib libXpm.dll.a libXpm.a libXpm.lib libXpm
Checking in directories /opt/local/lib for libXpm.so libXpm.sl libXpm.dylib libXpm.dll.a libXpm.a libXpm.lib libXpm
Checking for library libXpm.so in directory /opt/local/lib
Checking /opt/local/lib/libXpm.so
libXpm.so not found in /opt/local/lib
Checking for library libXpm.sl in directory /opt/local/lib
Checking /opt/local/lib/libXpm.sl
libXpm.sl not found in /opt/local/lib
Checking for library libXpm.dylib in directory /opt/local/lib
Checking /opt/local/lib/libXpm.dylib
Found file /opt/local/lib/libXpm.dylib
Checking if /opt/local/lib/libXpm.dylib contains compatible code
lipo -info /opt/local/lib/libXpm.dylib | grep ’ x86_64’
result: 1
is a compatible library
Checking if /opt/local/lib/libXpm.dylib is linked against compatible C++ run-time
otool -L /opt/local/lib/libXpm.dylib | grep libstdc++
result: 0
is linked against compatible C++ run-time library

tpochep · October 24, 2013, 8:41pm

[quote=“unelg”]picnam5a: /Users/ngu/jroot/abc/icons/arrow_down.xpm
picnam5b: /Users/ngu/jroot/abc/icons/arrow_down.xpm 0x7fe84b34e4f0 0
pic: arrow_down.xpm
getting xpms
picnam3a: /Users/ngu/jroot/abc/icons/arrow_up.xpm
picnam3b: /Users/ngu/jroot/abc/icons/arrow_up.xpm
picnam5a: /Users/ngu/jroot/abc/icons/arrow_up.xpm
picnam5b: /Users/ngu/jroot/abc/icons/arrow_up.xpm 0x7fe84b34f7d0 0

 std::cout<<"picnam5b: "<<picnam<< " "<<pic<< "  "<< pic->fPic <<std::endl;[/quote]

Ok, thanks! I'd say, this TGPicturePool is quite obviously a very sloppy code: its method was called TWICE with argument "arrow_down.xpm" (or any xpm file - looks like something is broken under the hood with xpms/images)and the first time it returns a valid pointer, but the second time - 0, though it should return 0 the first time already not obscuring and hiding the problem till the next call. I'll investigate why fPic is 0 and let you know later (well, unless somebody else already knows the reason and can help before, also, have a look at previous topics with arrow_*.xpm, may be they already had a recipe).

tpochep · October 24, 2013, 9:15pm

What about libAfterImage (I do not know the exact name).
In my configure.log I can see something like:
…

Checking whether to build included libAfterImage Result: yes
…

unelg · October 24, 2013, 9:27pm

I think I don’t have that.
but config.log says:
checking whether to build included libAfterImage
Result: yes

Here it is in the root distro:
graf2d/asimage/src/libAfterImage

tpochep · October 24, 2013, 9:31pm

[quote=“unelg”]I think I don’t have that.
but config.log says:
checking whether to build included libAfterImage
Result: yes[/quote]

Good, it’s much better if the problem is not complicated by some external library.

Well, fPic == kNone can actually come from TASImage.
So, next In graf2d/asimage/src/TASImage.cxx, method TASImage::GetPixmap, line number ~ 2212:

if (x11) { ret = ..... //can you print this 'ret' here? }

We’ll have a lot of prints now (I guess so), but it’s ok.

unelg · October 24, 2013, 9:38pm

zero all the time

Path: /Users/ngu/icons:/Users/ngu/jroot/abc/icons:.:/Users/ngu/icons:/Users/ngu/jroot/abc/icons:.
picnam: /Users/ngu/jroot/abc/icons/arrow_down.xpm
picnam3a: /Users/ngu/jroot/abc/icons/arrow_down.xpm
picnam3b: /Users/ngu/jroot/abc/icons/arrow_down.xpm
picnam5a: /Users/ngu/jroot/abc/icons/arrow_down.xpm
  TASI:0
picnam5b: /Users/ngu/jroot/abc/icons/arrow_down.xpm 0x7faa63b6ca20  0
pic: arrow_down.xpm
getting xpms
picnam3a: /Users/ngu/jroot/abc/icons/arrow_up.xpm
picnam3b: /Users/ngu/jroot/abc/icons/arrow_up.xpm
picnam5a: /Users/ngu/jroot/abc/icons/arrow_up.xpm
  TASI:0
picnam5b: /Users/ngu/jroot/abc/icons/arrow_up.xpm 0x7faa63a6da10  0
pic: arrow_up.xpm
got up xpm
not scaled???
????
got down xpm

tpochep · October 24, 2013, 9:47pm

[quote=“unelg”]zero all the time

Path: /Users/ngu/icons:/Users/ngu/jroot/abc/icons:.:/Users/ngu/icons:/Users/ngu/jroot/abc/icons:.
picnam: /Users/ngu/jroot/abc/icons/arrow_down.xpm
picnam3a: /Users/ngu/jroot/abc/icons/arrow_down.xpm
picnam3b: /Users/ngu/jroot/abc/icons/arrow_down.xpm
picnam5a: /Users/ngu/jroot/abc/icons/arrow_down.xpm
  TASI:0
picnam5b: /Users/ngu/jroot/abc/icons/arrow_down.xpm 0x7faa63b6ca20  0
pic: arrow_down.xpm
getting xpms
picnam3a: /Users/ngu/jroot/abc/icons/arrow_up.xpm
picnam3b: /Users/ngu/jroot/abc/icons/arrow_up.xpm
picnam5a: /Users/ngu/jroot/abc/icons/arrow_up.xpm
  TASI:0
picnam5b: /Users/ngu/jroot/abc/icons/arrow_up.xpm 0x7faa63a6da10  0
pic: arrow_up.xpm
got up xpm
not scaled???
????
got down xpm[/quote]

hmm, it looks like asimage2pixmap _always_ fails.
This function is in graf2d/asimage/src/libAfterImage/ximage.c. Can you either add some printfs there (since it's a C code) or set a breakpoint in asimage2pixmap function (if you have a debug build).
I have to recompile my ROOT now, since it's not an X11 version.

unelg · October 24, 2013, 9:59pm

Who defines X_DISPLAY_MISSING ?
I never see “1 b”

printf ("1\n");
#ifndef X_DISPLAY_MISSING
printf ("1 b\n");
        if( im )

tpochep · October 24, 2013, 10:01pm

[quote=“unelg”]Who defines X_DISPLAY_MISSING ?

printf ("1\n");
#ifndef X_DISPLAY_MISSING
printf ("1 b\n");
        if( im )[/quote]

Is it defined in your case? In my - it's not.

unelg · October 24, 2013, 10:12pm

config.h has
/* Define to 1 if the X Window System is missing or not being used. */
#define X_DISPLAY_MISSING 1

unelg · October 24, 2013, 10:14pm

remove the damn define and finally I got my damn arrow XPM.
what a waste of time !!!

tpochep · October 24, 2013, 10:15pm

[quote=“unelg”]config.h has
/* Define to 1 if the X Window System is missing or not being used. */
#define X_DISPLAY_MISSING 1[/quote]

Aha, this is much better - so none of functions in ximage.c actually works?
Ha, can you add #undef X_DISPLAY_MISSING at the beginning of ximage.c and will it change anything?

tpochep · October 24, 2013, 10:17pm

[quote=“unelg”]remove the damn define and finally I got my damn arrow XPM.
what a waste of time !!![/quote]

Yes, you are right, it’s a waste of time, if I had this problem on my mac it would be much faster, it’s quite stupid to “debug in a forum”. Still there is a bug in libAfterImage configuration we can fix so people in future will not have the similar problem on Mac OS and other platforms. While searching on this forum, I found the topic with the same problem and people … were using g++ 2.95 - so you can imagine, how old this bug is.

unelg · October 24, 2013, 10:24pm

graf2d/asimage/src/libAfterImage/configure has a bug to define validity of X,
or not.
Remember this thing came up w/ osx 10.9 fresh install.
Who knows what the guys at cupertino did to X11 standards … They might have moved around
some files to fool the configure utility…

tpochep · October 24, 2013, 10:27pm

[quote=“unelg”]graf2d/asimage/src/libAfterImage/configure has a bug to define validity of X,
or not.
Remember this thing came up w/ osx 10.9 fresh install.
Who knows what the guys at cupertino did to X11 standards … They might have moved around
some files to fool the configure utility…[/quote]

Yes, probably this script has a bug.
I do not have this problems with 10.9 and other people in this topic did not. Also, when ROOT crashes on linux with the same warning/error message, it can be the similar problem. And I have no idea about your XQuartz installation. For example, I’ve installed XQuartz from the scratch.

unelg · October 24, 2013, 10:34pm

you know what?
make clean; ./configure ==> X11 is now defined…
Where did my bug go? is this real ? or a very very bad nightmare ?

SongofHeaven · October 24, 2013, 10:53pm

I’m afraid this is not what you need to compile (this is not the branch I’ve mentioned in this topic).

Do:
git clone root.cern.ch/git/root.git any-name-you-want
cd any-name-you-want
git checkout -t origin/v5-34-00-patches
./configure
make[/quote]
Hi, all
I used this method to install root again in my updated mac os x 10.9, finally what i get is a error message below:

but if i do not use the v5-34-00-patches ,i.e. use the default branch, the error message I got is :

need your help,thanks!

remitche · October 24, 2013, 11:32pm

Hi,

I find the same thing after downloading v5-34-00-patches and trying to compile:

Error: cannot open file “AvailabilityMacros.h” include/RConfig.h:376:
Warning: Error occurred during reading source files
Warning: Error occurred during dictionary source generation
!!!Removing core/base/src/G__Base1.cxx core/base/src/G__Base1.h !!!
Error: core/utils/src/rootcint_tmp: error loading headers…
make: *** [core/base/src/G__Base1.cxx] Error 1

Any ideas?

Ryan

aphecetche · October 25, 2013, 8:50am

Hi all,

I also had some issues with libstd++ and I guess it might be due to an improper install of the XCode command line tools upon initial XCode install.

Could you try “xcode-select --install” and answer yes when prompted to install the command line tools and then retry the root compilation ?

It worked for me (well, at least those “etc/cling/Interpreter/ValuePrinter.h:11:10: fatal error: ‘string’ file not found” messages disappeared).

Regards,

dinbach · October 25, 2013, 12:28pm

[quote=“aphecetche”]Hi all,

I also had some issues with libstd++ and I guess it might be due to an improper install of the XCode command line tools upon initial XCode install.

Could you try “xcode-select --install” and answer yes when prompted to install the command line tools and then retry the root compilation ?

It worked for me (well, at least those “etc/cling/Interpreter/ValuePrinter.h:11:10: fatal error: ‘string’ file not found” messages disappeared).

Regards,[/quote]

Hi,
I had exactly the same issues with the compilation and the “xcode-select --install” indeed fixed everything.
Many thanks!