summaryrefslogtreecommitdiffstats
path: root/CMakeScripts/DefineDependsandFlags.cmake (follow)
Commit message (Collapse)AuthorAgeFilesLines
* add draft janet interface for livecodings-ol2019-12-021-1/+2
|
* CMake: enable -fstack-protector-strong by defaultPatrick Storz2019-10-301-2/+2
|
* gcc has AddressSanitizer support (allow WITH_ASAN)Thomas Holder2019-10-291-3/+0
|
* CMake/MSYS2: Link against libssp (required for _FORTIFY_SOURCE)Patrick Storz2019-10-181-0/+1
| | | | | | | | | | | | mingw-w64 does not seem to have native support for fortification, libssp offers it, so let's try this for now. Not sure yet if disabling _FORTIFY_SOURCE would be better/worse. See also https://sourceforge.net/p/mingw-w64/mailman/message/36764708/ https://github.com/msys2/MINGW-packages/issues/5803 https://github.com/msys2/MINGW-packages/issues/5868
* CMake: Avoid reordering CMAKE_CXX_FLAGSPatrick Storz2019-10-121-11/+17
| | | | | | | | | | | Order matters; the reordering caused the _FORTIFY_SOURCE flag to be always undefined (as we flipped a "-U" and "-D") Also skip removing duplicates to avoid similar pitfalls (we kept only the first occurrence but should've kept the last) It did not work properly anyway (we still had duplicates in the end) and was only required as we wrote the final CMAKE_CXX_FLAGS back to cache and consequently appended new flags with every run of CMAKE.
* Update DefineDependsandFlags.cmakeMarc Jeanmougin2019-10-121-1/+1
|
* Initialize some uninitialized variablesMarc Jeanmougin2019-10-111-1/+1
|
* Move LPE's into 3 groups normal, experimental y testing. Testing is only ↵Jabier Arraiza2019-10-081-6/+0
| | | | usable by setting up a flag. Must set off for release. I think also for LPETool
* remove HAVE_POTRACEMarc Jeanmougin2019-08-221-2/+0
|
* Make potrace a dep, remove previous codeMarc Jeanmougin2019-08-211-10/+6
|
* CMake: Sync states of WITH_NLS and ENABLE_NLS if gettext isn't foundPatrick Storz2019-07-111-0/+1
|
* WITH_ASAN AddressSanitizer compile optionThomas Holder2019-07-021-2/+17
|
* fix "ld: framework not found -lCocoa"Thomas Holder2019-06-191-0/+18
|
* Remove unused libyaml dependencyPatrick Storz2019-06-051-12/+0
| | | | | This was required by xverbs, which were removed in db05b842cba28f01b431eee890537959aa2d8fe3
* Fix linking for Homebrew (macOS)Thomas Holder2019-06-021-1/+1
| | | | Supersedes !710
* Include more feature librariesRené de Hesselle2019-05-311-2/+2
| | | | | | | | | | - gtkspell - libyaml - cppunit - librevenge - libcdr - ImageMagick 6 - OpenMP
* macOS: gtk-mac-integrationThomas Holder2019-05-291-0/+6
|
* Merge branch 'xgettextMigration'Marc Jeanmougin2019-05-271-5/+4
|\
| * Improve xgettext processMarc Jeanmougin2019-05-271-1/+1
| |
| * Mix init-tool and xgettext to allow inx and glade filesJabiertxof2019-02-091-0/+5
| |
| * Add xgettext instead intl-toolJabiertxof2019-02-091-9/+3
| |
* | Remove support for LCMS 1Karl Cheng2019-05-121-10/+1
| | | | | | | | | | | | | | Now that all platforms (including macOS) now support LCMS 2, we no longer need to support LCMS 1. Fixes: https://bugs.launchpad.net/inkscape/+bug/1133014
* | add _GLIBCXX_ASSERTIONS flag to detect OOB accesses in debug buildsMarc Jeanmougin2019-04-241-0/+1
| |
* | Add new lib2geom dependency: "double-conversion"Patrick Storz2019-04-021-0/+2
| |
* | CMake: Use CMAKE_CXX_STANDARD and don't set it in three placesPatrick Storz2019-03-311-8/+1
| | | | | | | | | | | | | | | | This requires bumping cmake version requirement to 3.1 (which should long be available on supported platforms, though) TODO: We actually depend on GNU compiler extensions right now (i.e. *don't* build with -std=c++11 but require -std=gnu++11)
* | CMake: Avoid defining _FORTIFY_SOURCE twice in separate placesPatrick Storz2019-03-311-6/+6
| |
* | GTK-3.22 dependencyThomas Holder2019-03-201-4/+4
| | | | | | | | Closes #147
* | Improve Debug buildingThomas Holder2019-03-061-0/+1
| | | | | | | | | | | | | | - add -O1 to make FORTIFY_SOURCE happy (eliminate compiler warnings) - canonicalize CMAKE_CXX_FLAGS_DEBUG to avoid unnecessary recompilation Both issues were introduced with f7fc3d3a7f8f51a01cd1dfb261c3e234617e092e
* | Add gcc/version checkMarc Jeanmougin2019-02-111-2/+7
| |
* | Fix buildMarc Jeanmougin2019-02-111-0/+4
|/
* Fix linking for MacPortsThomas Holder2019-01-051-1/+2
|
* CMake: Specifically enable Microsoft bit-field layout for structsEduard Braun2018-11-211-0/+1
| | | | (silences warnings in clang / already default for gcc)
* CMake: Remove popt from dependenciesEduard Braun2018-11-211-5/+0
| | | | Not required anymore after 408cb49b5559a81ea803df64bf58457a5dd4bf16
* -Werror=format -Werror=format-securityThomas Holder2018-11-211-0/+3
|
* Fix install location of .mo filesEduard Braun2018-10-121-1/+1
| | | | | | | .mo files were copied to e.g. "en.mo" instead of "inkscape.mo" This was broken after e939fcccf36f2a89435b0f8328db9be32ce3b600 as there was one tiny customization in our copy of FindGettext.cmake
* Merge branch 'cmake-cxx11' of gitlab.com:xzcvczx/inkscapeMarc Jeanmougin2018-10-081-0/+2
|\
| * Add CMAKE_CXX_STANDARD for cmake 3.1Simon Wells2018-10-081-0/+2
| | | | | | | | add note saying can remove -std=c++11 when minimum cmake is 3.1
* | Prefer LDFLAGS over LIBRARIES for pkgconfigSimon Wells2018-10-061-12/+2
|/ | | | | LDFLAGS include LIBRARIES, therefore remove duplicates Replaced LIBRARIES with LDFLAGS for gtk and gtkspell so they work even if the libraries are not in a standard path
* CMake: Allow to build with GraphicsMagick instead of ImageMagickEduard Braun2018-10-021-8/+20
| | | | | | | | | For now we still default to ImageMagick, but we can fall-back to GraphicsMagick (e.g. if only ImageMagick 7 is available; add a check for this). To prefer GraphicsMagick despite a compatible version of ImageMagick use -DWITH_IMAGE_MAGICK=OFF when running CMake
* Avoid duplicating defines in config.h and on command lineEduard Braun2018-10-011-4/+2
|
* config.h - remove all defines we never useEduard Braun2018-10-011-6/+0
| | | | | (strictly speaking PACKAGE and PACKAGE_NAME are unused, too, but let's keep them for now)
* Move GTKMM_CHECK_VERSION from config.h to it's own headerEduard Braun2018-10-011-9/+0
| | | | | Avoids having to recompile most of the Inkscape codebase whenever gtkmm is updated.
* CMake/po: Check for presence of intltool for creating inkscape.potEduard Braun2018-09-291-2/+9
| | | | | Also remove FindGettext module and use the one that ships with CMake (bumps minimum version to CMake 2.8.8 for GETTEXT_VERSION_STRING)
* CMake: Whitespace cleanupEduard Braun2018-09-291-80/+78
|
* fix 1789208 poppler 0.69Thomas Holder2018-09-281-12/+0
|
* CMake: Set WITH_GTKSPELL to OFF if not found for consistencyEduard Braun2018-06-301-46/+36
| | | | | Also remove (unused) legacy variable "TRY_GTKSPELL", remove an outdated comment, fix indentation and some cleanup
* make yaml optional even if presentMarc Jeanmougin2018-03-201-9/+11
|
* make jemalloc optional at compile timeMarc Jeanmougin2018-03-171-3/+7
|
* Fix GTKMM version check macroAlexander Valavanis2018-03-041-0/+9
|
* Explicitly link against libX11 for ld.goldTobias Ellinghaus2018-01-101-3/+11
| | | | | | Some linkers, like gold, don't find symbols in libraries not explicitly linked against. So we have to add libX11. On systems not using X11 this shouldn't hurt.