Great that there are performance benefits, but one thing that I think both meson/ninja and cmake (or cmake/ninja) got wrong, and GNU configure got right is "./configure --help". Why do neither of these tools simply add --help here? Yes, their syntax is different, but the issue is not only about --help. Often I could disable documentation or man page via --disable-man or --disable-doc or something like that. I recently had a discussion with a guy who transitioned into ninja, and I reasoned that there should be an option to skip installing man-pages. He thinks everyone needs manpages. I told him I never look at any local man page ever; I only look online for help. And have been doing so for almost 30 years. I understand the 1970s era of man-pages, but I have no use for them (I do gather local documentation, just not via manpages). He did not want to add a way to install his software without having available xmlto, docbook variants (which are a pain to install, just look at the LFS/BLFS instructions and even then they do not work cleanly). Meson and cmake are better than the GNU autotools, but the few things GNU autotools got right, were lost in both meson and cmake, which is interesting. It seems we can only select between different trade-offs here. New is not automatically better. And being 5% faster really is pointless if functionality is removed or not available anymore.

This is not purely a CMake problem, this is a maintainer problem. It's absolutely possible to not define the man page targets if the related tools are not found, or to add a variable to disable that.

In practice, it does show the limitations of the CMake model though. You should not be paying the price to generate the build files for bits you don't care about. You should be able to just load the CMake files and build a target, ignoring the rest.

This is how tools like Bazel function and that allows them to scale to very large monorepos. Build files are only processed when a target from that folder is built directly or used. CMake has to load everything as it does not know about the build tree and the maintainer has to manage exclusion lists or inclusion lists.

I think what you're talking about is part of a much larger pattern. Autotools imagines that the most important thing is the "user" in a world where people want to download your software and build it on their machine. They don't care (as much) about fast build times, the semantic clarity of the build files themselves, the ease with which one may change or extend the build sequence, test harnesses . . . Those are all developer worries.

They do care that the code can compile on their (possibly old) machine. They do care about the ability to include some features and not others (including their ability to not have to make decisions on this if they don't want to). And they certainly want documentation that helps them to achieve their goal (which, again, is to build the software).

The newer tools make substantial inroads in making the developer's life easier. If your software is not of the sort where people download it and try to build it (because it only ships as a closed binary, is entirely internal, etc.) then maybe that makes sense. But until all these fancy gadgets start seriously caring about the use case above, I'm going to stick with autotools.

> He thinks everyone needs manpages. I told him I never look at any local man page ever; I only look online for help...

Everyone's workflow is different. Your's obviously works for you and many others.

But I like local man pages. They get updated at the same time as the software is updated. Sometimes I have an older version of software and the online version is for a newer version, or vice-versa. Local man pages are always available and generally quicker than accessing them online.

If the machines in question are mult user, you should always install local documentation regardless of your personal preferences.

I wish that when people posted a performance win, they would include a theory explaining it. Unexpected speedups are usually noise, bugs, or harness issues. No program is faster due to the superior manly virtue of its programmer. Every effect has a cause.

When, from first principles, something shouldn't be faster, yet it is, you dig and dig until you understand. You don't just say your shit rules and the other thing sux0rz. If you do, you'll regret it: if you don't understand how you got a win, you don't understand how to keep it.