issue1577: fix broken test by assuming less about CVS output.
Specifically, output of "cvs ci" varies unpredictably across CVS versions,
so any test that includes the output of "cvs ci" is doomed to fail
some of the time. This fixes that by discarding the output of "cvs ci".
% create cvs repository with one project
cvs -f -d *REPO* init
cvs -f -q co proj
% create file1 on the trunk
cvs -f -Q add file1
cvs -f ci -madd file1 on trunk file1
% create two branches
cvs -f -q tag -b v1_0
T file1
cvs -f -q tag -b v1_1
T file1
% create file2 on branch v1_0
cvs -f -q up -rv1_0
cvs -f -Q add file2
cvs -f ci -madd file2 on branch v1_0 file2
% create file3, file4 on branch v1_1
cvs -f -Q up -rv1_1
cvs -f -Q add file3 file4
cvs -f ci -madd file3, file4 on branch v1_1 file3 file4
% merge file2 from v1_0 to v1_1
cvs -f -q up -jv1_0
U file2
cvs -f ci -mmerge file2 from v1_0 to v1_1
% cvs rlog output
RCS file: *REPO*/proj/file1,v
revision 1.1
RCS file: *REPO*/proj/Attic/file2,v
revision 1.1
revision 1.1.4.2
revision 1.1.4.1
revision 1.1.2.1
RCS file: *REPO*/proj/Attic/file3,v
revision 1.1
revision 1.1.2.1
RCS file: *REPO*/proj/Attic/file4,v
revision 1.1
revision 1.1.2.1
% convert to hg
initializing destination proj.hg repository
using builtin cvsps
collecting CVS rlog
9 log entries
creating changesets
4 changeset entries
connecting to *REPO*
scanning source...
sorting...
converting...
3 add file1 on trunk
2 add file2 on branch v1_0
1 add file3, file4 on branch v1_1
0 merge file2 from v1_0 to v1_1
% hg log output
3 merge file2 from v1_0 to v1_1
2 add file3, file4 on branch v1_1
1 add file2 on branch v1_0
0 add file1 on trunk