SBo scripts not building on current (read 1st post, pls)

55020 · 03-01-2018, 03:22 PM

Don't worry, continuity plans for gcc-java have been ready for a while now, thanks to Alien Bob.
(It'll need a tweak for version 5.5 eventually, but there's plenty of time for that)

kjhambrick · 03-01-2018, 03:37 PM

Quote:

Originally Posted by 55020

Don't worry, continuity plans for gcc-java have been ready for a while now, thanks to Alien Bob.
(It'll need a tweak for version 5.5 eventually, but there's plenty of time for that)

Thanks for the Link 55020 !

All you 'real' SBo Devs are awesome

-- kjh( ! whew ! )

Alien Bob · 03-01-2018, 04:57 PM

Good catch, I have now updated that script to gcc 5.5.0: https://git.slackbuilds.org/slackbui...57fcf39b8a793c

Bindestreck · 03-02-2018, 03:46 AM

opencv fails during compilation, same error as here: https://github.com/opencv/opencv/issues/10862

Reverting back to doxygen 1.8.9.1, compilation succeeds.

brodo · 03-02-2018, 03:50 AM

Quote:

Originally Posted by ponce

have you tried launching FreeCAD from the command line in a terminal and see if there's any output there when you try to open the examples?
Does this happen with all the examples or just with some of them?
Does this happen also on slackware stable?
maybe it's better if you open a dedicated topic on the matter here or on the SBo mailing list as, IMHO, this doesn't look related to the argument of this topic.

It is the result of command line action:

<<<<
bash-4.4$ FreeCAD
FreeCAD 0.16, Libs: 0.16RUnknown
© Juergen Riegel, Werner Mayer, Yorik van Havre 2001-2015
##### #### ### ####
# # # # # #
# ## #### #### # # # # #
#### # # # # # # # ##### # #
# # #### #### # # # # #
# # # # # # # # # ## ## ##
# # #### #### ### # # #### ## ## ##

** (FreeCAD:13312): WARNING **: invalid source position for horizontal gradient
** (FreeCAD:13312): WARNING **: invalid source position for horizontal gradient
>>>>>>

Q: Does this happen with all the examples or just with some of them?

With most of them. Especially STEP files.

Q: Does this happen also on slackware stable?

I use only current Slackware so I don't know.

ponce · 03-02-2018, 04:25 AM

Quote:

Originally Posted by Bindestreck

opencv fails during compilation, same error as here: https://github.com/opencv/opencv/issues/10862

Reverting back to doxygen 1.8.9.1, compilation succeeds.

strange, this should already be workarounded since a couple of weeks

http://cgit.ponce.cc/slackbuilds/commit/?h=opencv

are you sure you got the latest version of the repository?

Bindestreck · 03-02-2018, 05:24 AM

Quote:

Originally Posted by ponce

strange, this should already be workarounded since a couple of weeks

http://cgit.ponce.cc/slackbuilds/commit/?h=opencv

are you sure you got the latest version of the repository?

Hmm... I use SBo to sync, but looking at the changelog, the latest sync seems to be from 20180210:

Code:

HEAD is now at fa5bve6cc6 20180210.1 global branch merge.

Repository update complete.

I am using the repository "SBo-git (current)".

Am i doing something wrong?

ponce · 03-02-2018, 05:50 AM

Quote:

Originally Posted by Bindestreck

Hmm... I use SBo to sync, but looking at the changelog, the latest sync seems to be from 20180210:

Code:

HEAD is now at fa5bve6cc6 20180210.1 global branch merge.

Repository update complete.

I am using the repository "SBo-git (current)".

Am i doing something wrong?

I suspect you are not syncing correctly

https://github.com/Ponce/slackbuilds...ry-with-sbopkg (it's linked also in the first post of this topic)

Bindestreck · 03-02-2018, 06:00 AM

Quote:

Originally Posted by ponce

I suspect you are not syncing correctly

https://github.com/Ponce/slackbuilds...ry-with-sbopkg (it's linked also in the first post of this topic)

Hmm... it helped doing this:

Code:

rm -fR /var/lib/sbopkg/SBo-git

and then resyncing. So git got a little confused...

Thanks.

Qury · 03-02-2018, 06:32 AM

Hi ponce

FreeImage is not building on current. Output is below. This is required for the MEGASync client.

Code:

Source/LibRawLite/./internal/dcraw_common.cpp:5546:14: note: candidate 2: float powf64(float, float)
 static float powf64(float a, float b)
              ^~~~~~
Source/LibRawLite/./internal/dcraw_common.cpp:10186:85: warning: ISO C++ says that these are ambiguous, even though the worst conversion for the first is better than the worst conversion for the second:
         imgdata.lens.makernotes.MinAp4CurFocal = powf64(2.0f, (getreal(type) / 2.0f));
                                                                                     ^
In file included from /usr/include/c++/7.3.0/cmath:45:0,
                 from /usr/include/c++/7.3.0/math.h:36,
                 from Source/LibRawLite/./internal/dcraw_common.cpp:24:
/usr/include/bits/mathcalls.h:140:1: note: candidate 1: _Float64 powf64(_Float64, _Float64)
 __MATHCALL_VEC (pow,, (_Mdouble_ __x, _Mdouble_ __y));
 ^
Source/LibRawLite/./internal/dcraw_common.cpp:5546:14: note: candidate 2: float powf64(float, float)
 static float powf64(float a, float b)
              ^~~~~~
make[1]: *** [Makefile.gnu:64: Source/LibRawLite/./internal/dcraw_common.o] Error 1
make[1]: Leaving directory '/tmp/SBo/FreeImage'
make: *** [Makefile:21: default] Error 2

orbea · 03-02-2018, 08:20 AM

Try using clang, iirc it needs clang to build correctly on current, but I haven't tried in a while since it was an erroneous dependency in my case... (Openmw does not require freeimage nor did it ever require it, upstream confirmed so...)

kjhambrick · 03-02-2018, 08:24 AM

Quote:

Originally Posted by qunying

pdftk could not be built due to GCC no longer support Java.
Some of the functionality of it could be done with mutool from mupdf though.

Quote:

Originally Posted by 55020

Don't worry, continuity plans for gcc-java have been ready for a while now, thanks to Alien Bob.
(It'll need a tweak for version 5.5 eventually, but there's plenty of time for that)

Quote:

Originally Posted by Alien Bob

Good catch, I have now updated that script to gcc 5.5.0: https://git.slackbuilds.org/slackbui...57fcf39b8a793c

Thanks Alien Bob !

And thanks 55020 and gunying for the head's up !

I downloaded ; built and installed gcc5 from the link you posted.

Also installed the latest Slackware64 -current updates and booted the new 4.14.23 Generic Kernel.

After rebooting, I tried to build pdftk in the gcc5 Environment( my pdftk.SlackBuild wrapper is below my .sig ).

These are the first few lines in the .log file -- they do show that the environment is set for gcc5:

Code:

pdftk.SlackBuild startup  at Fri Mar  2 07:59:52 CST 2018

building pdftk with gcc5:

CC     = gcc-5
CPP    = cpp-5
CXX    = g++-5
AR     = gcc-ar-5
NM     = gcc-nm-5
RANLIB = gcc-ranlib-5

Archive:  /home/dld/slackbuilds/SBo/pdftk/pdftk/pdftk-2.02-src.zip
<<snip>>

The MakeFile fails when trying to invoke g++ on pdftk.cc due to unresolved #include <gcj/cni.h>

Code:

<<snip>>

gcj -O2 -fPIC -Wall -fsource=1.3 --encoding=UTF-8 --classpath="/usr/share/java:/tmp/SBo/pdftk-2.02-dist/java:." -c java_lib.jar
make[1]: Leaving directory '/tmp/SBo/pdftk-2.02-dist/java'
g++ -O2 -fPIC -DPATH_DELIM=0x2f -DASK_ABOUT_WARNINGS=false -DUNBLOCK_SIGNALS -fdollars-in-identifiers -DPDFTK_VER=\"2.02\" -O2 -fPIC -Wall -Wextra -Weffc++ -I/tmp/SBo/pdftk-2.02-dist/pdftk/../java pdftk.cc -c
pdftk.cc:31:0: warning: ignoring #pragma GCC java_exceptions [-Wunknown-pragmas]
 #pragma GCC java_exceptions
 
pdftk.cc:33:10: fatal error: gcj/cni.h: No such file or directory
 #include <gcj/cni.h>
          ^~~~~~~~~~~
compilation terminated.
make: *** [Makefile.Base:46: pdftk.o] Error 1

I tried a q&d patch to the pdftk-1.44-add_custom_optflags.patch file, appending a pair of -I's to the CPPFLAGS:

Code:

+export CPPFLAGS= $(OPT) -DPATH_DELIM=0x2f -DASK_ABOUT_WARNINGS=false -DUNBLOCK_SIGNALS -fdollars-in-identifiers -I/usr/lib64/gcc/x86_64-slackware-linux/5.5.0/include/c++ -I /usr/lib64/gcc/x86_64-slackware-linux/5.5.0/include

The -I's got me past the initial error but the build failed anyway with numerous errors, apparently related to the #includes.

This is the tail of the log:

Code:

pdftk.cc:3780:4: error: 'JvCreateJavaVM' was not declared in this scope
    JvCreateJavaVM(NULL);
    ^~~~~~~~~~~~~~
pdftk.cc:3876:17: error: 'class java::lang::String' has no member named 'indexOf'
    if( message->indexOf( JvNewStringUTF( "com.lowagie.text.pdf.PdfDictionary" ) )>= 0 &&
                 ^~~~~~~
pdftk.cc:3877:15: error: 'class java::lang::String' has no member named 'indexOf'
      message->indexOf( JvNewStringUTF( "com.lowagie.text.pdf.PRIndirectReference" ) )>= 0 )
               ^~~~~~~
pdftk.cc:3896:4: error: 'JvDetachCurrentThread' was not declared in this scope
    JvDetachCurrentThread();
    ^~~~~~~~~~~~~~~~~~~~~
pdftk.cc:3896:4: note: suggested alternative: 'JvAttachCurrentThread'
    JvDetachCurrentThread();
    ^~~~~~~~~~~~~~~~~~~~~
    JvAttachCurrentThread
make: *** [Makefile.Base:46: pdftk.o] Error 1

pdftk.SlackBuild FAILED at Fri Mar  2 08:01:49 CST 2018 ... RetCode = 2

Looks like I've got some more work to do on pdftk

-- kjh

This is the pdftk.SlackBuild wrapper ( .do-pdftk.SlackBuild.sh )

Code:

#!/bin/sh

DoName="pdftk.SlackBuild"

MyVersion=""

[ "$VERSION"  != "" ] && MyVersion="$VERSION"

. ./pdftk.info

[ "$MyVersion" = "" ] && MyVersion="$VERSION"

LogNam="$DoName-$MyVersion.log"

unset PRGNAM
unset VERSION
unset HOMEPAGE
unset DOWNLOAD
unset MD5SUM
unset DOWNLOAD_x86_64
unset MD5SUM_x86_64
unset REQUIRES
unset MAINTAINER
unset EMAIL
#
# set and export any extra variables here
#
source /etc/profile.d/gcc5.sh || ( echo "ERROR:  the pdftk.SlackBuild requires gcc5" >&2 ; exit 1 ; )

export VERSION="$MyVersion"
#
# README  here
#
# If PDF is electronic paper, then pdftk is an electronic staple-remover, 
# hole-punch, binder, secret-decoder-ring, and X-Ray-glasses. Pdftk is a 
# command-line tool for doing everyday things with PDF documents.  Keep one 
# in the top drawer of your desktop and use it to merge PDF documents, 
# split PDF documents, apply watermarks, repair corrupt PDF, update PDF 
# metadata, and more.  pdftk does not require Acrobat and is free software.
#
# README there
#
# do the deed
#
echo "$DoName startup  at `date`"                        |tee    $LogNam
echo ""                                                  |tee -a $LogNam
echo "building pdftk with gcc5:"                         |tee -a $LogNam
echo ""                                                  |tee -a $LogNam
echo "CC     = $CC"                                      |tee -a $LogNam
echo "CPP    = $CPP"                                     |tee -a $LogNam
echo "CXX    = $CXX"                                     |tee -a $LogNam
echo "AR     = $AR"                                      |tee -a $LogNam
echo "NM     = $NM"                                      |tee -a $LogNam
echo "RANLIB = $RANLIB"                                  |tee -a $LogNam
echo ""                                                  |tee -a $LogNam

./$DoName 2>&1                                           |tee -a $LogNam
RetCode=${PIPESTATUS[0]}

[ "$RetCode" = "0" ] && RetStr="success" || RetStr="FAILED"
#
# log your results
#
echo ""                                                 |tee -a $LogNam
echo "$DoName $RetStr at `date` ... RetCode = $RetCode" |tee -a $LogNam
exit $RetCode

Qury · 03-02-2018, 10:29 AM

Quote:

Originally Posted by orbea

Try using clang, iirc it needs clang to build correctly on current, but I haven't tried in a while since it was an erroneous dependency in my case... (Openmw does not require freeimage nor did it ever require it, upstream confirmed so...)

hi orbea

FreeImage did not work with clang either. Also it is not a requirement for OpenMW, but MEGASync.

Code:

              ^
/usr/include/math.h:399:34: note: expanded from macro '__MATH_PRECNAME'
# define __MATH_PRECNAME(name,r) name##f64##r
                                 ^
<scratch space>:236:1: note: expanded from here
powf64
^
Source/LibRawLite/./internal/dcraw_common.cpp:5546:14: note: candidate function
static float powf64(float a, float b)
             ^
Source/LibRawLite/./internal/dcraw_common.cpp:12033:30: warning: use of logical '&&' with constant operand [-Wconstant-logical-operand]
      is_raw += (i = get4()) && 1;
                             ^  ~
Source/LibRawLite/./internal/dcraw_common.cpp:12033:30: note: use '&' for a bitwise operation
      is_raw += (i = get4()) && 1;
                             ^~
                             &
Source/LibRawLite/./internal/dcraw_common.cpp:12033:30: note: remove constant to silence this warning
      is_raw += (i = get4()) && 1;
                            ~^~~~
7 warnings and 12 errors generated.
make[1]: *** [Makefile.gnu:64: Source/LibRawLite/./internal/dcraw_common.o] Error 1
make[1]: Leaving directory '/tmp/SBo/FreeImage'
make: *** [Makefile:21: default] Error 2

ponce · 03-02-2018, 10:56 AM

Quote:

Originally Posted by Qury

Hi ponce

FreeImage is not building on current. Output is below. This is required for the MEGASync client.

Code:

Source/LibRawLite/./internal/dcraw_common.cpp:5546:14: note: candidate 2: float powf64(float, float)
 static float powf64(float a, float b)
              ^~~~~~
Source/LibRawLite/./internal/dcraw_common.cpp:10186:85: warning: ISO C++ says that these are ambiguous, even though the worst conversion for the first is better than the worst conversion for the second:
         imgdata.lens.makernotes.MinAp4CurFocal = powf64(2.0f, (getreal(type) / 2.0f));
                                                                                     ^
In file included from /usr/include/c++/7.3.0/cmath:45:0,
                 from /usr/include/c++/7.3.0/math.h:36,
                 from Source/LibRawLite/./internal/dcraw_common.cpp:24:
/usr/include/bits/mathcalls.h:140:1: note: candidate 1: _Float64 powf64(_Float64, _Float64)
 __MATHCALL_VEC (pow,, (_Mdouble_ __x, _Mdouble_ __y));
 ^
Source/LibRawLite/./internal/dcraw_common.cpp:5546:14: note: candidate 2: float powf64(float, float)
 static float powf64(float a, float b)
              ^~~~~~
make[1]: *** [Makefile.gnu:64: Source/LibRawLite/./internal/dcraw_common.o] Error 1
make[1]: Leaving directory '/tmp/SBo/FreeImage'
make: *** [Makefile:21: default] Error 2

try to get in touch with the maintainer, in case he might have a fix...

USUARIONUEVO · 03-02-2018, 11:28 PM

Hi , beignet fails to build on current , cause LLVM is too new for beignet-1.3.1 , upgrade version to 1.3.2 , solve the problem.

Thanks!