summaryrefslogtreecommitdiff
path: root/doc/gendiff.1
diff options
context:
space:
mode:
Diffstat (limited to 'doc/gendiff.1')
-rw-r--r--doc/gendiff.144
1 files changed, 44 insertions, 0 deletions
diff --git a/doc/gendiff.1 b/doc/gendiff.1
new file mode 100644
index 000000000..40d100f52
--- /dev/null
+++ b/doc/gendiff.1
@@ -0,0 +1,44 @@
+.TH GENDIFF 1 "Mon Jan 10 2000"
+.UC 4
+.SH NAME
+gendiff \- utility to aid in error-free diff file generation
+.SH SYNOPSIS
+\fBgendiff\fR <directory> <diff-extension>
+.SH DESCRIPTION
+\fBgendiff\fR is a rather simple script which aids in generating a
+diff file from a single directory. It takes a directory name and a
+"diff-extension" as its only arguments. The diff extension should be
+a unique sequence of characters added to the end of all original,
+unmodified files. The output of the program is a diff file which may
+be applied with the \fBpatch\fR program to recreate the changes.
+.PP
+The usual sequence of events for creating a diff is to create two
+identical directories, make changes in one directory, and then use the
+\fBdiff\fR utility to create a list of differences between the two.
+Using gendiff eliminates the need for the extra, original and
+unmodified directory copy. Instead, only the individual files that
+are modified need to be saved.
+.PP
+Before editing a file, copy the file, appending the extension you have chosen
+to the filename. I.e. if you were going to edit somefile.cpp, copy it to
+somefile.cpp.save before editing it. Then edit the first copy (somefile.cpp).
+.PP
+After editing all the files you need to edit in this fashion, enter the
+directory one level above where your source code resides, and then type
+.SP
+.NF
+ $ gendiff somedirectory .orig > mydiff-orig.patch
+.FI
+.PP
+You should redirect the output to a file (as illustrated) unless you want to
+see the results on stdout.
+
+.PD
+.SH "SEE ALSO"
+.BR diff (1),
+.BR patch (1)
+
+.SH AUTHOR
+.nf
+Erik Troan <ewt@redhat.com>
+.fi