Do we have ideas for that error?
The error location is fully wrong unfortunately.
And I think that it is not a real bug in my macros or templates, as all other test programs compile and work fine.
That error was in github in 2019 already.
The concrete problem is
https://github.com/StefanSalewski/gintro/issues/99#issuecomment-738700076
With wrong inactive version of seq2GSList() it compiles but of course does not work. But when I try to fix that proc so that it does do some work compiler fails with Error: template instantiation too nested.
The reason for the issue seems to be
l = g_slist_prepend(l, s[i].impl)
when the seq base type does not contain a impl field.
The correct error message is something like
gobject.nim(2657, 34) Error: undeclared field: 'impl' for type gtk.BindingArg
which the compiler gives for a minimal test program, but not for a real larger program.
The funny fact is that I guessed such error already this morning, and I tried to catch it with when compiles("s[i].impl == nil"), But obviously using quotation marks was not such a good idea.
That is an interesting issue. Currently we get for most typos in module gobject.nim the same wrong error message:
gimplgobj.nim(16, 29) Error: template instantiation too nested
I have just installed latest devel compiler, makes no difference.