Neutral header comments in tstfabnd, tstx37, tstnospc (#724)

test/host/tstfabnd.c

@@ -9,13 +9,12 @@
  * allocated for the life of the job.  The data set can then be
  * scratched neither by the program that created it nor by its user.
  *
- * Measured on mvsdev 2026-09-09, mvslovers/ftpd#129:
+ * The failure was seen on a real system.  A program wrote to a data set
+ * until it was full; the system reported IEC031I D37-04, the program's
+ * ESTAE recovered, and the program then closed the FILE.  That close
+ * abended in turn and left the DD allocated.
  *
- *     IEC031I D37-04,IFG0554T,FTPDT,FTPDT,SYS00006,251,WORK00,...
- *     FTPD070E ABEND SD37 RECOVERED CMD=STOR SOCKET=3 TOTAL=1
- *     FTPD076W CLOSE ABENDED AFTER STOR, DD=SYS00006 FREE RC=4
- *
- * ftpd opens with fopen(dsn,"wb") - the buffered __fputc path.  The
+ * The program opens with fopen(dsn,"wb") - the buffered __fputc path.  The
  * D37 fires inside __awrite(); the caller's ESTAE unwinds THROUGH
  * libc370, so @@fflush.c's reset: label never runs and fp->upto still
  * points past the block that failed.  That is the state this test

test/mvs/tstnospc.c

@@ -39,9 +39,10 @@
  * line also goes to the console, because an abend discards whatever SYSOUT
  * is still sitting in the QSAM buffer.
  *
- * SHAPE.  mvslovers/ftpd's (ftpd#129), same as tstx37.c: __dsalcf() creates
- * with SPACE=TRK(1,0) - no secondary, so the first overflow is a D37 and not
- * an extend - __dsfree() catalogs, fopen(dsn,"wb") reopens by name.
+ * SHAPE.  A file-transfer upload into a full data set, same as tstx37.c:
+ * __dsalcf() creates with SPACE=TRK(1,0) - no secondary, so the first
+ * overflow is a D37 and not an extend - __dsfree() catalogs, fopen(dsn,"wb")
+ * reopens by name.
  *
  * CHECKS - these are the ones a green run must satisfy whatever the numbers
  * turn out to be; the numbers themselves are reported, not asserted.
@@ -74,12 +75,12 @@
  * upload probe.xmit to IBMUSER.MBT.XMIT.IN, run jcl/recvnosp.jcl, then
  * jcl/tstnospc.jcl.
  *
- * MEASURED TWICE on mvsdev 2026-09-13, before and after the #149 fix.  The
+ * MEASURED TWICE on the reference system, before and after the #149 fix.  The
  * checks below are written so that BOTH runs are green - they pin the shape
  * (no abend, no drift to EIO, clearerr() works), not the policy.  What the
  * fix changed is the map, and the map is the evidence.
  *
- * AFTER the fix, JOB00254: CC 0000, 9/9, no IEC031I.
+ * AFTER the fix: CC 0000, 9/9, no IEC031I.
  *
  *   map   XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
  *         0 full, 50 short.  Every write after the error is refused at
@@ -92,9 +93,9 @@
  *         the buffer refills, the eleventh flush fails, and the twelfth
  *         is refused.
  *
- * BEFORE the fix, JOB00252: CC 0000, 9/9 PASS, and no
- * IEC031I line in the job log.  TRK(1,0) on WORK00 took the same 200 records
- * of 80 that tstx37 fills it with.  The three answers:
+ * BEFORE the fix: CC 0000, 9/9 PASS, and no IEC031I line in the job log.
+ * TRK(1,0) took the same 200 records of 80 that tstx37 fills it with.  The
+ * three answers:
  *
  *   (a) THE MAP.  ..........X..........X..........X..........X......
  *       46 of 50 retries returned the full 80 bytes.  Not one of those 46
@@ -204,7 +205,7 @@ int main(int argc, char **argv)
         wtof("TSTNOSPC SETUP FAILED __dsalcf rc=%d", rc);
         return 8;
     }
-    __dsfree(dd);                   /* the free catalogs it - ftpd's step 1 */
+    __dsfree(dd);                   /* the free catalogs it */
 
     sprintf(fname, "'%s'", dsn);
     fp = fopen(fname, "wb");

test/mvs/tstx37.c

@@ -7,7 +7,7 @@
  * with an ESTAE could recover, but the abend unwound THROUGH @@fflush.c
  * before its reset: label, so fp->upto still pointed past the block that
  * failed and the next flush re-drove the identical WRITE.  Measured on
- * mvsdev JOB00235/241/245: that second attempt does not fail cleanly, it
+ * a real system: that second attempt does not fail cleanly, it
  * PROGRAM-CHECKS - 0x0C4 on one run and 0x0C6 on another for identical
  * code.  And one path reaches it with no API call at all: @@exit.c walks
  * grt->grtfile at program termination and fclose()s every survivor with no
@@ -29,9 +29,10 @@
  * it is CC 0000.  Every check is also written to the console, because an
  * abend discards whatever SYSOUT is still sitting in the QSAM buffer.
  *
- * SHAPE.  mvslovers/ftpd's (ftpd#129): __dsalcf() creates the data set with
- * SPACE=TRK(1,0) - no secondary, so the first overflow is a D37 and not an
- * extend - __dsfree() catalogs it, and fopen(dsn,"wb") reopens it by name.
+ * SHAPE.  A file-transfer upload into a full data set: __dsalcf() creates the
+ * data set with SPACE=TRK(1,0) - no secondary, so the first overflow is a
+ * D37 and not an extend - __dsfree() catalogs it, and fopen(dsn,"wb")
+ * reopens it by name.
  *
  * CHECKS
  *   (1) fwrite() stops short of the limit         - the write did fail
@@ -68,14 +69,14 @@
  * upload probe.xmit to IBMUSER.MBT.XMIT.IN, run jcl/recvx37.jcl, then
  * jcl/tstx37.jcl.
  *
- * MEASURED 2026-09-13 on mvsdev, JOB00247: CC 0000, 5/5 PASS, and - the
- * check no return code can make - NO IEC031I LINE IN THE JOB LOG AT ALL.
- * TRK(1,0) on WORK00 took the same 200 records of 80 that produced the D37
- * in JOB00245; this time fwrite() stopped at 200 with ferror() set and
+ * MEASURED on the reference system: CC 0000, 5/5 PASS, and - the check no
+ * return code can make - NO IEC031I LINE IN THE JOB LOG AT ALL.  TRK(1,0)
+ * took the same 200 records of 80 that produced the D37 before the exit
+ * was planted; this time fwrite() stopped at 200 with ferror() set and
  * errno 28 (ENOSPC), fclose() returned, and remove() answered 0.
  *
- * Cross-checked the same day by JOB00249: test/mvs/tstfabnd.c, whose three
- * steps each produced a D37 in JOB00245, now report try()=0x00000000 and
+ * Cross-checked: test/mvs/tstfabnd.c, whose three steps each produced a
+ * D37 before the exit was planted, now reports try()=0x00000000 and
  * "nothing to measure" in all three - the same shape, no abend.
  *
  * RC: 0 = every check passed, 8 = at least one did not (it is the COND CODE).
@@ -125,7 +126,7 @@ int main(int argc, char **argv)
         wtof("TSTX37 SETUP FAILED __dsalcf rc=%d", rc);
         return 8;
     }
-    __dsfree(dd);                       /* the free catalogs it - ftpd's step 1 */
+    __dsfree(dd);                       /* the free catalogs it */
 
     sprintf(fname, "'%s'", dsn);
     fp = fopen(fname, "wb");