graphics/cuneiform exits compilation in port with: Linking CXX executable ../../cuneiform CMakeFiles/cuneiform-cli.dir/cuneiform-cli.cpp.o: In function `main': /usr/ports/graphics/cuneiform/work/cuneiform-linux-1.1.0/cuneiform_src/cli/cuneiform-cli.cpp:(.text+0x5de): undefined reference to `Magick::Image::Image(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&)' /usr/ports/graphics/cuneiform/work/cuneiform-linux-1.1.0/cuneiform_src/cli/cuneiform-cli.cpp:(.text+0x644): undefined reference to `Magick::Image::write(Magick::Blob*, std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&)' c++: error: linker command failed with exit code 1 (use -v to see invocation) --- cuneiform --- *** [cuneiform] Error code 1 but compiles with poudriere. If I do this patch -- cuneiform_src/cli/cuneiform-cli.cpp.orig 2015-07-08 09:26:05 UTC +++ cuneiform_src/cli/cuneiform-cli.cpp @@ -152,7 +152,8 @@ static string supported_formats() { */ static char* read_file(const char *fname); -#ifdef USE_MAGICK +#ifndef USE_MAGICK +#define USE_MAGICK #include <Magick++.h> using namespace Magick; @@ -337,7 +338,9 @@ int main(int argc, char **argv) { return 0; } -#ifdef USE_MAGICK +#ifndef USE_MAGICK #define USE_MAGICK +#include <Magick++.h> +using namespace Magick; Magick::InitializeMagick(""); #endif it compiles fine. Other errors appears with gcc48, but this patch --- cuneiform_src/Kern/hrkint/stdprt.h.orig 2015-07-08 11:24:29 UTC +++ cuneiform_src/Kern/hrkint/stdprt.h @@ -58,6 +58,7 @@ OF THIS SOFTWARE, EVEN IF ADVISED OF THE #define __STDPRT_H #include "std.h" +#include <cstdarg> typedef int32_t HEvent32; typedef struct tagStdPrtEvent helps.
10.1-RELEASE-p13 FreeBSD 10.1-RELEASE-p13 #0: Wed Jul 1 04:03:12 CEST 2015 root@user:/usr/obj/usr/src/sys/GENERIC amd64
please provide your patch in a proper patch file.
Created attachment 161087 [details] patch-stdprt_h
Created attachment 161088 [details] patch-cuneiform-cli-cpp
I set USE_GCC=4.8+ in the Makefile. Clang throws a lot of warnings. Don't know if this is necessairely.
Created attachment 161574 [details] cuneiform.diff
Created attachment 162064 [details] cuniform patch with mandatory ImageMagick Hello, Walter asked for my help, because there is no visible progress in this ticket. I did some changes to his patch: - add the missing pump of PORTREVISION - remove the option to choose IMAGEMAGICK With the patch the port compiles fine as long as ImageMagick is installed. Without the software it is not possible to compile - whether via make or poudriere. Therefore i make ImageMagick a mandatory LIB_DEPENDS and remove all conditionals to check this. No it builds fine for 9.3, 10.1 and 10.2 - all amd64. Patch and Logs attached.
An example poudriere log created for FreeBSD 10.2 AMD64 can be found here. Attention it has a size of 3.6 MB: http://ports.toco-domains.de/poudriere-test-logs/cuneiform-1.1.0_5.log
Created attachment 162185 [details] cuneiform.diff
Correct Makefile, cause of portlint. Also tested on 9.3 and 10.2 i386.
@walter, I see two patches here, which makes how to progress this issue ambiguous for committers. We need one patch, that is approved by the maintainer and QA tested. Please provide instructions.
Make that "I see 3 patches here"
Created attachment 162187 [details] cuneiform.diff
Pending maintainer-approval + on one of the patches. @torsten, walter, can you please coordinate among yourselves which patch should be approved/obsoleted, leaving one patch to commit, thanks!
I found no way to obsolete Thorstens diff. Take the last one.
@walter, Torsten will need to obsolete it
I mailed him.
(In reply to Walter Schwarzenfeld from comment #18) As requested i obsoleted my patch. :) @Walter: Your patch removed the LIB_DEPENDENCY. In your mail you suggest that the LIB_DEPENDENCY now is mandatory and not removed. Can you check this again please? Thanks, Torsten
Created attachment 162197 [details] cuneiform.diff
LIB_DEPENCY was not meant. Only the conditions.
Created attachment 162198 [details] cuneiform.diff
Is it now ok?
Seems we run into a maintainer approval timeout.
Sorry for delay with reply, please commit and thank you for the patch
Created attachment 165398 [details] svn-diff_graphics_cuneiform
Created attachment 165399 [details] portlint_graphics_cuneiform
Add USE_GCC=4.8+ (to silence a bunch of error messages with clang).
Approval was for attachment 165398 [details], which was replaced with attachment 165399 [details] after approval. Maintainer, this needs your review and approval once more. @Walter: Feel free to omit the -v flag for portlint(8) to only include warnings and failures. Please also confirm this passes build and package testing with poudriere, or if not yet installed, attach the output of: make stage && make check-plist && make stage-qa && make package Thanks!
Created attachment 165406 [details] poudriere_graphics_cuneiform-1.1.0.5.log
After nearly 3 month we can consider this as maintainer timeout and go ahead.
Hello Walter, the patch did not apply any more. Can you please provide a new one? I will take care of the PR. Greetings, Torsten
Created attachment 171156 [details] svn-diff_graphics_cuneiform_new
Silenced some warnings. Can't obsolete old diff, I don't see it under obsolete patch, only the new.
Hello Walter, the patch did not apply: $ make patch ===> License BSD3CLAUSE accepted by the user ===> cuneiform-1.1.0_5 depends on file: /usr/local/sbin/pkg - found ===> Fetching all distfiles required by cuneiform-1.1.0_5 for building ===> Extracting for cuneiform-1.1.0_5 => SHA256 Checksum OK for cuneiform-linux-1.1.0.tar.bz2. ===> Patching for cuneiform-1.1.0_5 ===> Applying FreeBSD patches for cuneiform-1.1.0_5 I can't seem to find a patch in there anywhere. => Patch patch-cuneiform__src_Kern_rimage_sources_main_cricontrol.cpp failed to apply cleanly. => Patch(es) patch-CTDIB.cpp patch-c_trans_h patch-clu_test.c patch-cor_spel_c patch-cuneiform-cli-cpp applied cleanly. *** Error code 1 Stop. make: stopped in /usr/ports/graphics/cuneiform
Created attachment 171263 [details] patch-cuneiform
I have some problems with svn with this port, I hope the new patch will work now.
Patch applies now - start testbuilds. :)
Patch obsolete, should reworked. I close this PR.