[common] Add fmt_debug_string polyfill

[jwnimmer-tri]

Towards #22078.

---

Sometimes we want to print or log a string. Imagine this function:

void DoFoo(const std::string& arg);

For logging, we might try:

void DoFoo(const std::string& arg) {
  if (condition) {
    drake::log()->warn("Suspicious arg = {}", arg);
  }
  ...
}

That would print e.g.

WARNING Suspicious arg = Hello, world!

However, we'll usually want to print the quoted string value, so that the user can see e.g. if they have training whitespace:

void DoFoo(const std::string& arg) {
  if (condition) {
    drake::log()->warn("Suspicious arg = \"{}\"", arg);
  }
  ...
}

That would print e.g.

WARNING Suspicious arg = "Hello, world!"

That's great! But what if the string itself has double quotes inside of it? The output would be confusing. Or, some strings might contain non-printable characters like 0x08, or awkwardly-printable characters like a tab (\t).

Format has a solution for that:

void DoFoo(const std::string& arg) {
  if (condition) {
    drake::log()->warn("Suspicious arg = {:?}", arg);
  }
  ...
}

The ? presentation type is the "debug string" format, which not only double-quotes the string but also converts non-printable characters into their \xNN hex code, so that the output is a string literal that can be pasted back into either C++ or Python.

Unfortunately, our Jammy copy of fmt is just slightly too old to contain this feature, so we need to polyfill it until we drop jammy.

void DoFoo(const std::string& arg) {
  if (condition) {
    drake::log()->warn("Suspicious arg = {}", fmt_debug_string(arg));
  }
  ...
}

---

This change is 

[jwnimmer-tri]

+@SeanCurtis-TRI for feature review, please. Low priority.

[SeanCurtis-TRI]

+@xuchenhan-tri for platform review.

Reviewed 4 of 4 files at r1, all commit messages.
Reviewable status: 2 unresolved discussions, LGTM missing from assignee xuchenhan-tri(platform)

---

common/fmt.h line 46 at r1 (raw file):

/** Returns `fmt::("{:?}", x)`, i.e, using fmt's "debug string format"; see
the fmt.dev docs for the '?' presentation type for details. We provide this

BTW Is this a widely used abbreviation of which I'm not familiar? The format implies a file or a URL more generally. Or is it simply a typo?

Code quote:

fmt.dev

---

common/fmt.h line 49 at r1 (raw file):

wrapper because not all of our supported platforms have a new-enough fmt
to rely on it. On platforms with older fmt, we use a Drake re-implementation
of the feature that does NOT handle unicode correctly. */

BTW This gives me a great appreciation of all the work that fmtlib does.

[xuchenhan-tri]

Reviewed 4 of 4 files at r1, all commit messages.
Reviewable status: 2 unresolved discussions

---

common/fmt.h line 46 at r1 (raw file):

Previously, SeanCurtis-TRI (Sean Curtis) wrote…

BTW Is this a widely used abbreviation of which I'm not familiar? The format implies a file or a URL more generally. Or is it simply a typo?

I think this is meant as a URL. https://fmt.dev/11.0/syntax/ has some more details on the use of '?'

[SeanCurtis-TRI]

Reviewable status: 2 unresolved discussions

---

common/fmt.h line 46 at r1 (raw file):

Previously, xuchenhan-tri wrote…

I think this is meant as a URL. https://fmt.dev/11.0/syntax/ has some more details on the use of '?'

Oh, I inferred that as the most logical. But I'd expect either something more explicit (like a link) or more colloquial "the fmt dev docs" -- without the dot. It's the dot that just threw me off as an expression of the concept I've not seen before.

[xuchenhan-tri]

Reviewed 1 of 1 files at r2, all commit messages.
Reviewable status: complete! all discussions resolved, LGTM from assignees SeanCurtis-TRI(platform),xuchenhan-tri(platform)