Actually, a second thought. IIUC the reason for this change is that without
these rules, no indentation is done for variable names in code like
static myttype *
variable_name;
static myttype
variable_name;
This is fine in normal C code, since no indentation is the same as
zero indentation, which is what we want. But for C embedded in Bison,
no indentation is not the same as zero indentation. We want it indent
to the top-level indentation of the embedded C. E.g.,
{
static myttype
variable_name;
}
rather than
{
static myttype
variable_name;
}
Right? In that case, we should really add a fallback indent rule for the
embedded c in bison, such that it indents to the top-level of the embedded c.
This is more robust than trying to cover all cases in the c rules. Does that
make sense?
Basically, add something like
(catch-all bison-ts-mode--bison-parent ,bison-ts-mode-indent-offset)
To the end of C’s indent rules.
Yuan
bison-ts-mode.el