server : return HTTP 400 on invalid grammar (#24144) (#24154)

Throw on grammar parse failure so the server returns HTTP 400
instead of silently dropping the constraint.
Add a regression test for the invalid-grammar response.

Fixes #24144

common/sampling.cpp

@@ -259,6 +259,9 @@ struct common_sampler * common_sampler_init(const struct llama_model * model, st
              }
         }
     }
+    if (!grmr && !grammar_str.empty()) {
+        throw std::runtime_error("failed to parse grammar");
+    }
 
     // Compute prefill tokens from the generation prompt
     std::vector<llama_token> prefill_tokens;