rpcgen: make compiler arglist allocation dynamic

[brooks]

Summary

In CheriBSD we exceed the 19 non-NULL arguments in the static array. Add
a simple size doubling allocator and increase the default to 32.

GC remnants of support for fixed arguments.

Test Plan

Survives a buildworld including some calls to rpcgen.

[arichardson]

I wish the C standard library had some kind of dynamic array so that this is not reimplement in every program...

I wonder if we could use something like the utarray.h macros from uthash?
Could we import the five headers that https://github.com/troydhanson/uthash into contrib and use utarray instead? I know the library is not that efficient since it uses strdup() for all strings that you insert but at least it means not reinventing the wheel.
We already have two copies of uthash.h in the source tree and elftoolchain also includes with utarray.h.

But for now this definitely looks good to me.

[brooks]

In D21971#480075, @arichardson wrote:

I wish the C standard library had some kind of dynamic array so that this is not reimplement in every program...

I wonder if we could use something like the utarray.h macros from uthash?
Could we import the five headers that https://github.com/troydhanson/uthash into contrib and use utarray instead? I know the library is not that efficient since it uses strdup() for all strings that you insert but at least it means not reinventing the wheel.
We already have two copies of uthash.h in the source tree and elftoolchain also includes with utarray.h.

We certainly could use utarray instead, though I'm not super convinced by its readability.

(Importing uthash.h into contrib is on my todo list since I'd like to merge my changes to use it instead of bdb for hashes in init and a couple other places.)

[arichardson]

In D21971#480086, @brooks wrote:

In D21971#480075, @arichardson wrote:

I wish the C standard library had some kind of dynamic array so that this is not reimplement in every program...

I wonder if we could use something like the utarray.h macros from uthash?
Could we import the five headers that https://github.com/troydhanson/uthash into contrib and use utarray instead? I know the library is not that efficient since it uses strdup() for all strings that you insert but at least it means not reinventing the wheel.
We already have two copies of uthash.h in the source tree and elftoolchain also includes with utarray.h.

We certainly could use utarray instead, though I'm not super convinced by its readability.

(Importing uthash.h into contrib is on my todo list since I'd like to merge my changes to use it instead of bdb for hashes in init and a couple other places.)

Sounds good. I agree that utarray is quite ugly (and inefficient) but we could still use it to avoid reimplementing dynamic vectors for non-performance-critical code duplication. I haven't looked if there are any alternatives, I only discovered it because we already use uthash.

[jrtc27]

usr.bin/rpcgen/rpc_main.c

924

This is undefined behaviour the first time round as arglist is NULL, even though argcount is 0. Also, a realloc+memset should be faster than calloc+memcpy, whilst also side-stepping this issue (and without having to introduce a branch on arglist).

[jrtc27]

usr.bin/rpcgen/rpc_main.c

924

In fact we don't even need the memset; addarg((char *)NULL); is always called just before the execvp so the array will always be NULL-terminated.

[brooks]

• Use realloc() rather than calloc()+memcpy().
• Limit argmax to prevent overflow (no possible on FreeBSD due to ARG_MAX).