cargo: require 'std' when enabling 'unicode'

BurntSushi · BurntSushi · commit 72024a801c26 · 2022-09-07T09:11:10.000-04:00
The comment in the src/lib.rs diff should explain things for the most
part here. The main reason for this commit is to avoid a backcompat
hazard after 1.0 is released when and if we want to drop the 'once_cell'
dependency.
diff --git a/README.md b/README.md
@@ -143,7 +143,9 @@ and Unicode support.
 * `unicode` - **Enabled** by default. This provides APIs that require sizable
   Unicode data compiled into the binary. This includes, but is not limited to,
   grapheme/word/sentence segmenters. When this is disabled, basic support such
-  as UTF-8 decoding is still included.
+  as UTF-8 decoding is still included. Note that currently, enabling this
+  feature also requires enabling the `std` feature. It is expected that this
+  limitation will be lifted at some point.
 * `serde` - Enables implementations of serde traits for `BStr`, and also
   `BString` when `alloc` is enabled.
 
diff --git a/src/lib.rs b/src/lib.rs
@@ -384,14 +384,30 @@ and Unicode support.
 * `unicode` - **Enabled** by default. This provides APIs that require sizable
   Unicode data compiled into the binary. This includes, but is not limited to,
   grapheme/word/sentence segmenters. When this is disabled, basic support such
-  as UTF-8 decoding is still included.
+  as UTF-8 decoding is still included. Note that currently, enabling this
+  feature also requires enabling the `std` feature. It is expected that this
+  limitation will be lifted at some point.
 * `serde` - Enables implementations of serde traits for `BStr`, and also
   `BString` when `alloc` is enabled.
 */
 
 #![cfg_attr(not(any(feature = "std", test)), no_std)]
 #![cfg_attr(docsrs, feature(doc_auto_cfg))]
 
+// Why do we do this? Well, in order for us to use once_cell's 'Lazy' type to
+// load DFAs, it requires enabling its 'std' feature. Yet, there is really
+// nothing about our 'unicode' feature that requires 'std'. We could declare
+// that 'unicode = [std, ...]', which would be fine, but once regex-automata
+// 0.3 is a thing, I believe we can drop once_cell altogether and thus drop
+// the need for 'std' to be enabled when 'unicode' is enabled. But if we make
+// 'unicode' also enable 'std', then it would be a breaking change to remove
+// 'std' from that list.
+//
+// So, for right now, we force folks to explicitly say they want 'std' if they
+// want 'unicode'. In the future, we should be able to relax this.
+#[cfg(all(feature = "unicode", not(feature = "std")))]
+compile_error!("enabling 'unicode' requires enabling 'std'");
+
 #[cfg(feature = "alloc")]
 extern crate alloc;
 
