Add item_name parse callback.

[ekse]

Issue: #428

[emilio]

Looks ok, but needs tests, any chance you can add some basic tests to bindgen-integration?

[jethrogb]

lgtm

[ekse]

I added two tests, let me know if you need more.

[emilio]

Looks fine, thanks! It may also be worth adding tests with types and users of the test, so something like:

struct my_prefix_bar {
    int foo;
};

struct my_prefix_foo {
   my_prefix_bar member;
};

Just knowing it builds is probably fine :)

[ekse]

I added the test you mentioned, I create an instance of the foo struct to validate it works as expected.

[emilio]

Looks great, thanks!

[emilio]

@bors-servo r+

[bors-servo]

📌 Commit 5b741da has been approved by emilio

[bors-servo]

⌛ Testing commit 5b741da with merge 8489d42...

[bors-servo]

☀️ Test successful - status-travis
Approved by: emilio
Pushing 8489d42 to master...

[jethrogb]

This turns out to be not very useful, because whitelist/blacklist checking is done after the mapping in the PR is performed. @emilio do you know what the correct place to put the rename logic would be?

[emilio]

This turns out to be not very useful, because whitelist/blacklist checking is done after the mapping in the PR is performed. @emilio do you know what the correct place to put the rename logic would be?

Well that means that you could whitelist with the stripped prefix, right? I'd rather not introduce two different names... The nice thing of this PR is that the complexity doesn't blow up because everything uses the new name.

But we could somehow add a new option NameOptions to avoid using the callback, and somehow use that for whitelisting / blacklisting.

[jethrogb]

Well that means that you could whitelist with the stripped prefix, right?

That's not possible, because the whitelist regex is matching exactly on the part that's being stripped.

[emilio]

Oh I see, that's indeed unfortunate... Let me think a bit about how to do this.

[emilio]

I opened #1452, could you give that a spin?