Port build system to CMake (2/2 from #5)

[aaron-sierra]

This is the second half of #5 and, therefore, depends on #6 which ends with the version 1.8 bump.

These changes include:

• providing a cross-build compatible alternative to iotools --make-links during install.
• porting the build system to CMake for future dependencies and more trivial integration into Buildroot's package build framework.

Buildroot package reference files

Config.in

config BR2_PACKAGE_IOTOOLS
        bool "iotools"
        help
          The iotools package provides a set of simple command line
          tools which allow access to hardware device registers.
          Supported register interfaces include PCI, IO, memory mapped
          IO, SMBus, CPUID, and MSR. Also included are some utilities
          which allow for simple arithmetic, logical, and other
          operations.

          https://github.com/aaron-sierra/iotools/

if BR2_PACKAGE_IOTOOLS

config BR2_PACKAGE_IOTOOLS_NOLINKS
        bool "Exclude symlinks"
        default n

endif

iotools.mk

###############################################################################
#
# iotools
#
###############################################################################

IOTOOLS_VERSION = 3e0166d9cc832544279f27688279d6060c81b4b0
IOTOOLS_SITE = $(call github,aaron-sierra,iotools,$(IOTOOLS_VERSION))
IOTOOLS_LICENSE = GPLv2
IOTOOLS_LICENSE_FILES = COPYING
IOTOOLS_CONF_OPTS = -DSTATIC=OFF

ifeq ($(BR2_PACKAGE_IOTOOLS_NOLINKS),y)
IOTOOLS_CONF_OPTS += -DINSTALL_LINKS=OFF
endif

$(eval $(cmake-package))

[andy-shev]

Now you can drop the first part of the PR and rebase it against current master/main.

[andy-shev]

Also, please address the comments I have given in #5.

[aaron-sierra]

Now you can drop the first part of the PR and rebase it against current master/main.

I've rebased to account for the merge commit.

[aaron-sierra]

Also, please address the comments I have given in #5.

If you are referring to these discussions, I await further comment from you:

• Default state of INSTALL_LINKS
  • Update documentation and port build system to CMake #5 (comment)
• Availability of readelf:
  • Update documentation and port build system to CMake #5 (comment)

[andy-shev]

Also, please address the comments I have given in #5.

If you are referring to these discussions, I await further comment from you:

• Default state of INSTALL_LINKS

  • Update documentation and port build system to CMake #5 (comment)

• Availability of readelf:

  • Update documentation and port build system to CMake #5 (comment)

Hmm... Maybe I don't understand something. There two comments from me and no replies from you. Can you elaborate what I should do / write?

[aaron-sierra]

Hmm... Maybe I don't understand something. There two comments from me and no replies from you.

I'd left my comments "Pending" 🤦 They should be visible to you, now.

[aaron-sierra]

The current version implements changes discussed here:

• Set minimum CMake version to 3.16
• Omit 2.0.0 bump

[andy-shev]

Some minor comments, but overall looks good to me.

[andy-shev]

@aaron-sierra Sorry, I have screwed up something. Can you create a new PR based on what's now in the master branch?

The problem was that the repository has diverged in a sense of v1.8 tag and what's in the master. I synced it, but it produced two similar merge commits (of v1.8). So I reset the master to v1.8 to start over.

[aaron-sierra]

Hi Andy, I synced my fork to yours right after the merge occurred, so you should be able to pull ddb8a6f (#7 with merge commit) from there.

[andy-shev]

Hi Andy, I synced my fork to yours right after the merge occurred, so you should be able to pull ddb8a6f (#7 with merge commit) from there.

Nope that won't work, your repo has the same issue mine had. You need to be sure that you have commit 3abdb69 in your repo. For this you need to run git pull --rebase origin master and see what will happen.

[aaron-sierra]

I see, you amended 7545cbd with a sign-off, creating 3abdb69:

I tend to work with a fast-forward-only merge policy to avoid automatically generated commit messages, so I'm not entirely clear how my pull request was able to rewrite history in your master. Anyway, look for the new PR shortly.