ezmlm-get: fatal: Program error: bad range in idx_mkthreads

[z-image]

It seems dignum is created on first delivery to the list, after digest sub-list is turned on. But apparently it's also used when getting the archive digest (and maybe other cases?). So, isn't it better to create it always if missing?

Something like this makes the error disappear for me:

--- ezmlm-get.c.original        2020-11-18 15:20:29.084183718 +0200
+++ ezmlm-get.c 2020-11-19 16:11:19.191753412 +0200
@@ -1026,10 +1026,10 @@
        u = max;
     }
     if (u == 0) u = 1;                 /* -get.5 => 1-5 */
-    if (to < u) to = u;                        /* -get23_2 => 23 */
     if (to >= u + MAXGET) to = u + MAXGET - 1;
                                        /* no more than MAXGET at a time */
     if (to > max) to = max;
+    if (to < u) to = u;                        /* -get23_2 => 23 */
     if (flagindexed && (outformat != NATIVE))  /* fake out threading */
       idx_mkthreads(&msgtable,&subtable,&authtable,&datetable,
        u,to,max,0);

[z-image]

Here's an updated patch, that fixes the issue for me:

--- ezmlm-get.c.original        2020-11-18 15:20:29.084183718 +0200
+++ ezmlm-get.c 2020-11-30 16:02:12.002512268 +0200
@@ -1030,6 +1030,7 @@
     if (to >= u + MAXGET) to = u + MAXGET - 1;
                                        /* no more than MAXGET at a time */
     if (to > max) to = max;
+    if (to < u) u = to;                        /* LINUX-2145 */
     if (flagindexed && (outformat != NATIVE))  /* fake out threading */
       idx_mkthreads(&msgtable,&subtable,&authtable,&datetable,
        u,to,max,0);