The relevant code seems to be at line 33 of https://sourceforge.net/p/maxima/code/ci/master/tree/src/inmis.lisp and line 1396 of https://sourceforge.net/p/maxima/code/ci/master/tree/src/combin.lisp, but I have not understood it.
listofvars([taylor(..),1]) gives an error when listdummyvars=false
Thanks. I just replaced the file /usr/share/maxima/5.43.2/share/stringproc/stringproc.lisp with the one that you sent me; I am not sure if there is supposed to be a compiled version anywhere. This does indeed eliminate the bug for me. It is not very fast, though; 3 or 4 seconds for a string of length 16384.
Output of build_info() is as follows: Maxima version: "5.43.2" Maxima build date: "2020-02-21 05:22:38" Host type: "x86_64-pc-linux-gnu" Lisp implementation type: "GNU Common Lisp (GCL)" Lisp implementation version: "GCL 2.6.12" User dir: "/root/.maxima" Temp dir: "/tmp" Object dir: "/root/.maxima/binary/5_43_2/gcl/GCL_2_6_12" Frontend: false
Actually, here is a much simpler case. If we set s to be a string consisting of the letter "a" repeated 1355 (or more) times, and evaluate ssubst("b","a",s), then we get the same error. So in particular, special properties of the backslash character do not seem to be relevant. I note that ssubst is defined at line 1129 of /usr/share/maxima/5.43.2/share/stringproc/stringproc.lisp, and that the definition is recursive, with one nested function invocation for each occurrence of the string to be substituted....
Mysterious error in ssubst