For any of you who still have GCG 8.1 around, here are the fixes for two
bugs which I recently uncovered. After making the changes shown below
rebuild the object modules and then the shared libraries after making the
changes. The VMS commands were:
$ gcgsupport
$ cc 'gencflags' formoutput
$ cc 'gencflats' specialfileio
$ lib/object genoblib:genlib.olb formatout.obj,specialfileio.obj
$ makeshare
The GCG programs did not need to be relinked for the changes to take
effect.
1. formoutput.c has a Y2K bug resulting in the year 2000 coming out as
"19100". Change:
sprintf(string, "%s %2d, 19%02d %02d:%02d:%02d.%02d"
months[(int)tmP->tm_mon], (int)tmP->tm_mday, (int)tmP->tm_year,
to
sprintf(string, "%s %2d, %04d %02d:%02d:%02d.%02d",
months[(int)tmP->tm_mon], (int)tmP->tm_mday, 1900+(int)tmP->tm_year,
2. The mark file examples all use "!" as a comment delimiter, but
the application code is actually looking for "*". This results in
the following two lines from a mark file generating marks from 100 to 200
when no marks are expected:
!50 100
!150 200
The GCG supplied examples all worked not because they were syntactically
correct, but because none had integers following the "!". Since it's
possible that some function somewhere depends upon "*" working as a comment
the fix adds "!" as a second comment delimiter. That is, if no comment
delimiter is found the code continues to process tokens and emits integers
when found, ignoring everything else.
Change both instances of the following in specialfileio.c:
if (token[0] != '*') {
to
if (token[0] != '*' && token[0] != '!') {
Regards,
David Mathog
mathog at seqaxp.bio.caltech.edu
Manager, sequence analysis facility, biology division, Caltech
