Figure out if we should remove CMake #164
Comments
|
I've found CMake useful. It is another thing to maintain, but it will help others who aren't using meson. |
|
For my money it should be kept. I'm unconvinced that meson is as popular as it's proponents sometimes claim, and or better or worse cmake is more the entrenched standard. (I just did a quick search for stats but couldn't find any). |
|
Not the most scientific experiment but grepping through bioconda recipes their are 497 occurences of "cmake" and only 31 of "meson". The meson hits appear to be exclusively PacBio related software. |
|
@ekg and @cjw85 thanks for pitching in! |
|
During an average workday, I probably build 2-3 new pieces of C/C++ code.
2/3 of them use Make, and 1/3 uses CMake. I'm surprised to have never run
into meson. I won't doubt that it's better than CMake (it wouldn't be hard)
but it's worth noting that getting the right CMake executable set up is
already a point of difficulty for users.
I have also included edlib in other CMake-based projects. It's easy to do.
Will I be able to use meson with minimal changes and do the same kind of
inclusion?
…On Mon, Oct 12, 2020, 01:27 Martin Šošić ***@***.***> wrote:
@ekg <https://github.com/ekg> and @cjw85 <https://github.com/cjw85>
thanks for pitching in!
What have you found CMake useful regarding, would you mind elaborating in
more details? I am asking because I am wondering if Meson could easily
replace that use case or not. CMake is certainly much more used than Meson,
but Meson seems to be much easier to maintain / read.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#164 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/AABDQEPJN2NMVOIV55JRAWDSKI5PDANCNFSM4SK7F4MQ>
.
|
|
I'd echo Erik's comments, other than not being surprised about not running into meson: I'm not surprised because its simply not widely used (as far as I can tell, the only other time I've come across it was this PR). Maybe Erik's surprise is based on it appearing nicer to use and why aren't we all using it? Because of this unfamiliarity I can't really comment on how meson interacts with other projects using other build systems. I should be clear I'm not saying "don't use meson", I'm all for breaking the mould of entrenched standards; but I do think it would be a mistake to drop CMake, or at least not include in your README how to use edlib within projects using CMake (you already have some great docs there). |
|
@ekg wow that is some interesting kind of workday, when you build couple of new pieces of C/C++ code every day!! @ekg and @cjw85 I get what you are saying. On one hand I would like to drop CMake and go just with Meson because it seems easier to maintain. On the other hand, I don't want to make it harder for people to use Edlib and fact is, CMake is the dominant solution and having CMake in Edlib enables easy integration. Solution could be what @cjw85 mentioned: If I could provide easy way for CMake users to plug in Edlib built with Meson, then they would not loose anything. Good docs might be enough for this, however right now I also don't know enough I would have to research it, so for now I will certainly leave CMake. @ekg when you said you included edlib in other CMake-based projects and it's easy to do -> how do you do it? Is it as described in Edlib docs, or smth different? What about plugging edlib in Cmake project via pkg-config, is that something you would be comfortable with / have ever done? |
With Meson, we don't need CMake. However, it might be useful to have it, so people can more easily use Edlib in their CMake projects. On the other hand, that means we have to maintain it which is not great. So what do we do? I would gladly drop it, but I don't know what effect would it have on Edlib users.
This has been discussed here: #161 (comment) .
The text was updated successfully, but these errors were encountered: