> That's not really true, you can just fine host your GP fork wherever you 
> want. 

With a fork of GP, you carry along all the extra files and git history and 
Github Issues and stuff of plugins you don't care about.

> what's the advantage of own separate build system that we couldn't get in GP

Ability to use own (non-Autotools) build system, or to use non-recursive 
automake, or just to have a much simpler and light weight Autotools build 
system.

> And most people suck at Autotools

Thats why special options like `--disable-all` in combination with 
`--enable-foo=yes` aren't ideal if a plugin wants a standalone distribution.

> But it still duplicates work

It does, but it separates it too. GP requires special autoconf macros and 
boilerplate files and stuff, so it's not a huge deal having a separate non-GP 
specific (potentially Autotools) build system, IMO. The redundant boilerplate 
in GP becomes just like the boilerplate/metadata needed to put packages in a 
package manager, just a way to get an external plugin distributed with 
geany-plugins.

> but currently any external plugin that want to do things right have to 
> re-implement a lot of GP's build system checks and features

Is there really that much it would need to re-implement? Maybe some gettext 
stuff, possibly something like `--with-geany-libdir` and own check like 
`PKG_CHECK_MODULES([GEANY],[geany])`?

> You can probably use ACLOCAL_AMFLAGS = -Itherightdirectory in Makefile.am

IIUC @sardemff7 tried it and it still had the issue I experienced.

> I'm not sure it's possible to have several AC_CONFIG_MACRO_DIRS calls

It takes multiple space-separated directories, presumably equivalent to 
multiple `AC_CONFIG_MACRO_DIR` calls, which is probably not valid. I think it 
requires new-ish Automake though.

---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/geany/geany-plugins/pull/440#issuecomment-225371983

Reply via email to