SL Baur <steve(a)xemacs.org&gt; writes: REPLACE is hosed because insert-file-contents, for reasons unknown, narrows the buffer between (point) and (point) just before calling insert-file-contents-internal. OTOH, the relevant code in Finsert_file_contents_internal() looks like this: if (!NILP (replace)) { buffer_delete_range (buf, BUF_BEG (buf), BUF_Z (buf), !NILP (visit) ? INSDEL_NO_LOCKING : 0); } Obviously, if the buffer is narrowed to zero characters' length, the above code is no-op. If you call the -internal function directly, you'll see that it works 100% right under Mule. The person who introduced the narrowing should explain its purpose. I cannot and I think it should be removed, pronto. Oh, and insert-file-contents also "documents" REPLACE not to work with Mule, but I don't see any reason why it shouldn't. The only thing that really doesn't work with Mule is the FSFMACS_SPEEDY_INSERT stuff, but that's only an internal optimization. -- Sam, is "pronto" a real word? Ceterum censeo Mule esse delendam! --me, after reading the code for insert-file-contents.