coimage and isInjective take no options, so kernel's DegreeLimit, SubringLimit and Strategy cannot be reached through them

This issue was triaged from [`bugs/mike/1-pushForward1`](https://github.com/Macaulay2/M2/blob/388c1ff0ce30d83751dea7bc7eac77fdc1305dd7/bugs/mike/1-pushForward1), one of the 857 files removed from the pre-GitHub `bugs/` tree by [`d2c8d27826`](https://github.com/Macaulay2/M2/commit/d2c8d27826) and catalogued in [#36](https://github.com/Macaulay2/M2/issues/36). **The commentary below was written by Claude (Claude Opus 5, via Claude Code)**, not by @d-torrance, whose account posted it -- please weigh it accordingly.

### The original file, verbatim

```text
Mike, pushForward1 has been supplanted by coimage, but pushForward1 had many
options, no longer available. Is that a problem?

i1 : options pushForward1

o1 = OptionTable{BasisElementLimit => infinity }
DegreeLimit => {}
MonomialOrder => EliminationOrder
PairLimit => infinity
StopBeforeComputation => false
StopWithMinimalGenerators => false
Strategy => NonLinear
UseHilbertFunction => true

o1 : OptionTable

i3 : options coimage
```

### Where it stands today

`kernel RingMap` accepts `DegreeLimit`, `SubringLimit` and `Strategy`, and honours them. The two
one-line functions in `ringmap.m2` that exist only to call it accept no options at all, so those
controls cannot be reached through them.

```m2
i1 : R = ZZ/101[a..d]; S = ZZ/101[s,t];

i3 : F = map(S,R,matrix{{s^4, s^3*t, s*t^3, t^4}});

i4 : kernel F

3 2 2 2 2 3 2
o4 = ideal (b*c - a*d, c - b*d , a*c - b d, b - a c)

o4 : Ideal of R

i5 : kernel(F, SubringLimit => 1)

o5 = ideal(b*c - a*d)

o5 : Ideal of R

i6 : coimage(F, SubringLimit => 1)
stdio:6:1:(3): error: no method found for applying coimage to: ...

i7 : isInjective(F, SubringLimit => 1)
stdio:7:1:(3): error: no method found for applying isInjective to: ...
```

`options coimage` and `options isInjective` are both `null`.

### Where it comes from

Both are single expressions over `kernel f`, declared with no option table:

https://github.com/Macaulay2/M2/blob/development/M2/Macaulay2/m2/ringmap.m2#L375-L377

so `coimage f` always runs the full `kernel f`, and there is no spelling of "compute the coimage,
stopping after the first *n* subring generators" or "…up to degree *d*". `kernel` itself is fully
hooked and strategy-aware:

https://github.com/Macaulay2/M2/blob/development/M2/Macaulay2/m2/ringmap.m2#L231-L236

### Why it is worth a line of code

For an elimination that does not finish, a partial answer is the only answer available, and `coimage`
is the documented way to compute the ideal of the image of a ring map. The workaround is to call
`kernel` directly and form the quotient by hand — which is exactly what `coimage` is, so nothing is
lost mathematically; but the convenience wrapper is the thing a caller reaches for, and it silently
has no way to express the bound.

There is no `image RingMap`, so `coimage` and `isInjective` are the whole of the affected set.

### A caution for whoever measures this

`kernel RingMap` caches on the ring map, and the underlying `gb` caches on the graph matrix, so
re-testing option variants against one `F` in one session will show the options being *ignored* when
they are not:

```m2
-- misleading: kernel F has already completed and cached the gb
kernel F -- 4 generators
kernel(F, SubringLimit => 1) -- 4 generators, apparently ignoring the option
```

Build a fresh ring and map per variant and the options behave correctly — `SubringLimit => 1` gives
one generator, `=> 2` gives two. Worth stating because it is the first thing a reader will try.

### Where this came from

`bugs/mike/1-pushForward1`, one of the 857 files removed with the `bugs/` directory in
[`d2c8d27826`](https://github.com/Macaulay2/M2/commit/d2c8d27826) and catalogued in
[#36](https://github.com/Macaulay2/M2/issues/36). Dan Grayson's note asks:

> Mike, pushForward1 has been supplanted by coimage, but pushForward1 had many options, no longer
> available. Is that a problem?

followed by `options pushForward1`, listing eight, and then a bare `options coimage`.

Answering it: mostly not. `pushForward1` was removed in favour of `relations coimage map(M,f)`
(`changes.m2:1827`), and of its eight options, `DegreeLimit` and `Strategy` are on `kernel` by name,
`BasisElementLimit`'s analogue for an elimination is `SubringLimit`, and `UseHilbertFunction` is
better than restored — the `AffineRing` strategy calls `canUseHilbertHint` itself and then asserts the
hint was used, so it is automatic rather than a user's decision. `MonomialOrder`, `PairLimit`,
`StopBeforeComputation` and `StopWithMinimalGenerators` have no successor.

The one gap that is not merely a rename is the one above: the options exist, and the wrappers cannot
pass them on.

Note that `pushForward` is *not* the successor and is not what this is about — it requires the map to
be module-finite, which `pushForward1` never did.

### Searches this rests on

Titles and bodies for `coimage`, `pushForward`, `pushForward1`, `kernel`, `SubringLimit`; comments via
`gh search issues` for `pushForward1`, `coimage`, `minPres`. Pull requests included, not filtered.
[#3284](https://github.com/Macaulay2/M2/issues/3284) is the nearest open neighbour — `kernel`
misbehaving under `pushFwd` on multigraded input — and is about answers rather than options.
`pushForward` is under active development
([#4366](https://github.com/Macaulay2/M2/issues/4366),
[#4435](https://github.com/Macaulay2/M2/issues/4435),
[#4465](https://github.com/Macaulay2/M2/issues/4465),
[#4176](https://github.com/Macaulay2/M2/issues/4176)), all widening which ring maps are accepted,
which is a different axis from this.

_{`open` · disposition `issue` · source of truth: [`bug-triage/catalog.tsv`](https://github.com/d-torrance/M2/blob/bug-triage/bug-triage/catalog.tsv)}

Start in M2/Macaulay2/m2/ringmap.m2 around lines 375-377, then compare the wrappers with the option handling for kernel around lines 231-236. Use fresh rings and maps for each option variant, since kernel and gb cache results. Done means coimage and isInjective accept the relevant options and honor bounds such as SubringLimit.

Written by the indexing model from the issue text.