news@jpusa1.UUCP (usenet) (08/27/87)
An obscure bug surfaced when trying to co a bunch of
files. The variable 'linecorr' is not initialized to
0 for each file. This will cause editstring to break
if linecorr happened to be < 0 at the end of doing the
last file. Following are diffs to rcs.c and rcsgen.c
*** /tmp/,RCSt1a11512 Wed Aug 26 19:06:54 1987
--- rcs.c Wed Aug 26 18:45:44 1987
***************
*** 504,509
/* no revision deleted */
fastcopy(finptr,frewrite);
} else {
if ( cuttail )
buildeltatext(gendeltas);
else
--- 504,510 -----
/* no revision deleted */
fastcopy(finptr,frewrite);
} else {
+ extern int linecorr; linecorr = 0;
if ( cuttail )
buildeltatext(gendeltas);
else
*** /tmp/,RCSt1a11526 Wed Aug 26 19:08:47 1987
--- rcsgen.c Wed Aug 26 18:47:37 1987
***************
*** 94,99
*/
{
int i;
if (deltas[0]==target) {
/* only latest revision to generate */
--- 94,100 -----
*/
{
int i;
+ extern int linecorr; linecorr = 0;
if (deltas[0]==target) {
/* only latest revision to generate */
--
Stu Heiss {gargoyle,ihnp4}!jpusa1!sturoy@felix.UUCP (Roy Morris) (09/03/87)
In article <317@jpusa1.UUCP> stu@jpusa1.UUCP (Stu Heiss) writes: >An obscure bug surfaced when trying to co a bunch of >files. The variable 'linecorr' is not initialized to >0 for each file. This will cause editstring to break >if linecorr happened to be < 0 at the end of doing the >last file. Following are diffs to rcs.c and rcsgen.c A cleaner fix to the problem is to add the following line to the routine initeditfiles() in rcsedit.c. We weren't bitten by the variable 'editline' not being initialized, but it's not worth waiting to see if it happens. 88a89 > editline = linecorr = 0; Roy Morris FileNet Corp ...!felix!roy