CMake and Zlib

[Neustradamus]

@madler: It is possible to solve all problems and merge some PRs?

Thanks in advance.

Issues:

• Missing zconf.h when using zlib as a git submodule (with CMake) #133
• CMake skip building examples/tests/running tests and installing files? (As submodule) #147
• Make sure the zlib cmake do not pollute the working directly. #162
• CMake: including current src directory into include path #218
• missing ZLIB configuration support on Windows [feature request] #259
• CMake: Make tests optional #308
• CMake option to build/install only static library #359
• cmake comes with FindZLIB and zlib ZLIBConfig and the targets they import are different. #366
• CMake build giving invalid output or seg fault on some inputs using Z_FIXED #485
• cmake and configure builds are not bit-for-bit identical #497
• configure includes wrong zconf.h in out-of-tree builds #499
• Never ever edit files in source directory #600
• Build files: consistency of binaries names on Windows between CMakeLists and Makefiles #652
• Add FetchContent compatibility #759
• Add package export to installation #762
• error: unknown type name ‘z_size_t’ from test/minigzip.c #772
• Can't get zlib1 #816
• CMake build would delete the source file zconf.h #781
• Dirty source repository after build #812
• zlib 1.3 forces /MD although I build it with /MT on Windows #849
• Use GNUInstallDirs #884
• Missing zlib.h in CMake built directory #899
• cmake: Target "zlibstatic" INTERFACE_INCLUDE_DIRECTORIES property contains path issue #915
• [Bug] ZLib 1.3.1 does not build on Windows #919
• snprintf warning with ubuntu + c89 + clang + cmake #927
• Compiler warnings when building on Windows with clang + cmake #932
• Any plan to release 1.3.2 with cmake build directory interference? #937
• Problem with CMake`s check_type_size #972
• zlib.pc vs windows #988
• cmake config files #1009
• Release 1.3.2 tag to support libpng >1.6.40 #1019
• Build error with emscripten #1026
  --
• Add the Imported Target ZLIB::ZLIB for CMake. zlib-ng/zlib-ng#1320
• cmake --build build 时报错 fastcws/fastcws#5 + fix zlib cmake issues fastcws/fastcws#6

PRs:

• CMakeLists.txt fixes for linux #15
• Adding support for position independent code when compiling in static #21
• Linking fails if source path contains spaces #25
• Mac OS X linker does not support --version-script #26
• Support cmake building i686 and amd64 asm implementation. #36
• Support cmake building i686 and amd64 asm for develop branch #37
• CMake: file(RENAME) since 2.8 #43
• Simplify installation paths #67
• Fix cmake build for MSVC with AMD64 option #72
• Added the option to disable the building of example binaries in the cmake build system. #91
• Fix CMakeLists.txt for windows compilation #116
• Include zconf.h in example binary in order to fix CMake's add_subdirectory() #132
• CMake option for not building the example programs #138
• Add ZLIB_TESTS command line option to disable tests if needed #139
• CMake: GNUInstallDirs, export target zlib::zlib, CPack configuration, install PDB file #148
• Cmake install fix #170
• More cmake options #177
• cmake: Add ld-version-script option #193
• Fix hard-coded offsets in AMD64 assembly. #196
• CMake: use CMAKE_CURRENT_SOURCE_DIR, *not* top-level src dir. Fixes #218 #219
• Added option To build Examples #224
• NEON implementation for Adler32 #251
• CMake: Do not specify absolute paths but relative ones. #265
• Use CMAKE_CURRENT_SOURCE_DIR in CMakeLists.txt #271
• CMake on MinGW: Add VERBATIM to add_custom_command(). #277
• cmake improvements #281
• add option to BUILD_EXAMPLES #283
• add export statement for cmake config #285
• CMake: Use configuration package for zlib install target #295
• Removed deleted assembler stuff also from CMake. #310
• cmake build enhancements, mostly for windows, with some documentation #315
• Add continuous integration support for GitHub #317
• Add a cmake switch to disable example binaries. #321
• Fix CMakeLists.txt for Visual Studio and AMD64 option #336
• CMake Update #337
• Added target_include_directories to build goals #343
• fixed for out-of-source build #344
• Feature/improve cmake #347
• Fix build on latest cygwin64 #367
• Ms VisualStudio - Assembler speedups on x64 #376
• cmake: allow user to override CMAKE_DEBUG_POSTFIX #385
• CMake: Set version in command 'project' #386
• remove leftover cmake code #392
• Don't specify --version-script on Solaris / Illumos in CMakeLists.txt. #395
• Add missing include files to cmake install step #407
• Remove unused sys/types.h cmake configuration. #432
• Make building examples optional #441
• Build zlib against windows static C runtime with CMake #448
• Add optimized slide_hash for Power processors #457
• Add optimization for Adler32 checksum for Power processors #458
• Add optimized longest_match for Power processors #459
• Adding CPU features detection code #468
• Use PROJECT_SOURCE_DIR for cmake include_directories #471
• CMakeLists.txt for examples #472
• cmake: allow skip building examples #487
• cmake: mimic behavior of Makefile a bit better; mostly fixes issue #497? #498
• Fix configure buglet; verify bit-for-bit identical installs; enhance junit output; add Linux/bsd/mac builders #500
• Make the rename of zconf a CMake option #519
• Add support for building only static library with CMake. #526
• Option to build zlib against Windows static C runtime #535
• Update cmake files to latest CMake. #543
• Stop cmake deprecation warning about using minimum versions than 2.8.12 #581
• [cmake][build] Remove ASM configure options #614
• cmake: respect custom RC flags with mingw #677
• fix: Make tests optional #691
• Re enable ASM speedups on Win #701
• Remove source modification from CMake configuration #713
• Make it build on AIX using Open XLC 17.1.1 #714
• CMake: Rename example{,64} -> zlib_example{,64} #718
• CMake package created during installation #721
• build: update cmake version to new syntax 👷 #774
• Updated CMakeLists.txt to minimum cmake 3.0.0 #792
• CMakeLists.txt: Fix add_subdirectory build #818
• Allows example binaries skipping #824
• [mini_zip][build] Add minizip build target #835 (not merged)
• Allow SKIP_EXAMPLE build Update CMakeLists.txt - default is still build example  #861
• Fix compilation issues when added as a subdirectory to a CMake project. #863
• Replace tabs from CMakeLists.txt with spaces #866
• Sdonela/disable examples (#1) #870
• Add CMake option to link MSVC runtime statically #878
• Add iOS headless building #879
• add workflow to build with all available C standards #925
• Compiler warning with fileno when building with non-gnu C extensions #926
• add Werror/Wall/Wextra #933
• Fix for issue of 32-bit z_off64_t with MinGW{-w64} while _LARGEFILE64_SOURCE=1 & _LFS64_LARGEFILE=1. #934
• change CI CMake build to use Open Watcom toolchain #955
• correct Open Watcom OS/2 build #958
• cmake: don't use --version-script on static libraries #966
• Allow overriding cmake target names, install pdb symbols on msvc compiler #976
• Fix llvm cross build for win on arm #983
• Ensure possible padding bytes are initialized #1002
• Modernize cmake file and do cmake-format #1004
• configure: Add --mandir to override default directory for manuals. #1012

I have not finished:

• https://github.com/madler/zlib/pulls?q=is%3Apr+is%3Aopen+cmake
• https://github.com/madler/zlib/issues?q=is%3Aissue+is%3Aopen+cmake

Linked to:
@Mizux, @albrematt, @johnb003, @miurahr, @WildCard65, @wrowe, @ScatteredRay, @jnhyatt, @Jihadist, @hwhsu1231, @mtl1979, @ClausKlein, @sum01, @craigscott-crascit, @hjmallon, @mattparks, @puneetmatharu, @vszakats, @tbeu, @fredgan, @RobinGeffroy, @jfern2011, @kirankotari, @etiennearnal, @omegacoleman, @yyzworker, @yalcinerbora, @niparx, @navidR, @lperron, @vguen, @tr1cks, @AndrewAtAvenza, @gvollant, @hopeless-programmer-online, @glebm, @jheaff1, @SpaceIm, @AraHaan, @pzychotic, @siepkes, @schultetwin1, @Bagira80, @wantehchang, @dankegel, @chanphil, @MichaelGoulding, @stkruse, next after.

@nmoinvaz

[Neustradamus]

Other people linked to CMake and Zlib.

@malytomas, @PiotrSikora, @avoidik, @jeking3, @JonasVautherin, @sav-ix, @01e9, @henryborchers, @zdenop, @rgpublic, @Slackadays, @bear24rw, @kgerheiser, @paulsapps, @danielalves, @rocammo, @bmarques1995, @rafael-santiago, @Andreas-Schniertshauer, @Togtja, @joeyparrish, @robiwano, @jeking3, @henryborchers, @aneeshpb-amazon, @kavionic, @cloudwindy

[madler]

There are too many cmake issues and pull requests for me to go through and disposition them, and I wouldn't know what's good or bad anyway. I very much like @Mizux 's suggestion to have a group render on, coordinate, and most importantly verify proposed cmake changes. My main concern are proposed changes that work in one case, but messes up a bunch of others. Ideally this group would be able to cover a good set of platforms.

[nmoinvaz]

You might want to give interested parties Triage access to the repository to help manage all the GitHub issues.

You could decide to merge in PRs based on the number of GitHub code review approvals.

I’d say stick to only merging changes that are absolutely necessary or ones that fix problems that lead to many duplicate issues.

[vszakats]

It would help a lot if someone with the necessary rights could approve the CI for PRs. Mine has been pending for 1+ year with no test run.

[ClausKlein]

My Feature/improve cmake #347 is ready to merge.
And it past CI https://github.com/ClausKlein/zlib/actions

[vszakats]

@ClausKlein Neat workaround for the blocked CI.

My #677 passes green, and ready to merge too.

[Neustradamus]

@ClausKlein: Nice :)

I hope it will be a good beginning!

@madler, can you look?

[Neustradamus]

@vszakats: Can you look improvements from @ClausKlein to adapt your PR?

[vszakats]

@Neustradamus Both PRs touch the buggy if(NOT MINGW) line enabling zlib1.rc. My patch fixes it for all targets (it needs the other changes in my PR), including MSYS builds. #347 fixes it for MSYS.

Can't see anything else I could adapt from the other PR, they are targeting different issues.

UPDATE: to my RC comment above: missed a NOT when reading it up earlier today, so both patches are correct, and if(NOT MINGW) should end up as if(WIN32) to fix it proper.

[ClausKlein]

I see this diff on my branch to #667, should I apply it?

diff --git a/CMakeLists.txt b/CMakeLists.txt
index d4eb7cb..19ac18b 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -121,7 +121,7 @@ set(ZLIB_SRCS
     zutil.c
 )
 
-if(NOT MINGW AND NOT MSYS)
+if(WIN32)
     set(ZLIB_DLL_SRCS
         win32/zlib1.rc # If present will override custom build rule below.
     )
diff --git a/win32/Makefile.gcc b/win32/Makefile.gcc
index 081e391..35172fb 100644
--- a/win32/Makefile.gcc
+++ b/win32/Makefile.gcc
@@ -50,7 +50,7 @@ AR = $(PREFIX)ar
 ARFLAGS = rcs
 
 RC = $(PREFIX)windres
-RCFLAGS = --define GCC_WINDRES
+RCFLAGS =
 
 STRIP = $(PREFIX)strip
 
diff --git a/win32/zlib1.rc b/win32/zlib1.rc
index ceb4ee5..5253b6a 100644
--- a/win32/zlib1.rc
+++ b/win32/zlib1.rc
@@ -1,11 +1,7 @@
 #include <winver.h>
 #include "../zlib.h"
 
-#ifdef GCC_WINDRES
 VS_VERSION_INFO		VERSIONINFO
-#else
-VS_VERSION_INFO		VERSIONINFO	MOVEABLE IMPURE LOADONCALL DISCARDABLE
-#endif
   FILEVERSION		ZLIB_VER_MAJOR,ZLIB_VER_MINOR,ZLIB_VER_REVISION,0
   PRODUCTVERSION	ZLIB_VER_MAJOR,ZLIB_VER_MINOR,ZLIB_VER_REVISION,0
   FILEFLAGSMASK		VS_FFI_FILEFLAGSMASK

[vszakats]

I think we are better keeping PRs granular and each fixing distinct problems, so I'd keep them separate. I'm also fine with rebasing if necessary. This makes it easier to later bisect where a problem was introduced or track down what changed and why. Our issue IMO is that none of the CMake patches are actually merged. (Unless we want to somehow offer one "megapatch" solving all issues at once, which is possible but needs a great amount of work and testing.)

[mtl1979]

I see this diff on my branch to #667, should I apply it?

diff --git a/CMakeLists.txt b/CMakeLists.txt
index d4eb7cb..19ac18b 100644
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -121,7 +121,7 @@ set(ZLIB_SRCS
     zutil.c
 )
 
-if(NOT MINGW AND NOT MSYS)
+if(WIN32)
     set(ZLIB_DLL_SRCS
         win32/zlib1.rc # If present will override custom build rule below.
     )

As far as I know, MINGW and MSYS are defined only with few most recent versions of cmake, so testing them not being defined makes very little sense. I know some people think that we all should use very latest version of cmake, but the reality is that some people still use older versions of cmake for own valid reasons. The core idea of zlib has been to support very old systems, not the very latest like some zlib forks do.

[vszakats]

My patch (#677) fixes that by deleting all MINGW uses (even though zlib has been relying on it since 2011). Also MSYS is not necessary there after that.

[mtl1979]

MINGW

MINGW was added in cmake 3.2 and MSYS was added in cmake 3.14. My own work PC has cmake 3.10.2 installed in /usr/bin/.

[vszakats]

CMake documented it in 3.2 (2015-03-04) indeed, but zlib's use predates that: dc5a43e (2011-09-09) by almost 4 years. Even more interestingly, the first introduction of MINGW in CMake happened on 2003-08-21: Kitware/CMake@a413160, predating its 2.4 (2006-04-18) release (their first tagged one in Git) by 3 years. (Git log goes back to year 2000.)

MINGW seems safe to use, but we don't need this macro here and it causes bugs even when defined correctly by CMake.