One of the most confusing areas in RPM is certainly the %setup macro and
how its options interact with each other. I found myself cut & pasting
existing examples or doing trail & error for every new package. I've now
read the (incomplete) documentation and (more complete ;-) source code
in-depth and tried to summarize this beast for us.
- - -
Synopsis
--------
%setup[<x>]
-q # quiet unpacking operation
-n <dir> # set name of directory
-c # create directory (and change to it) before unpacking
-D # do not delete directory before unpacking sources
-T # do not perform default archive unpacking of %{SOURCE<x>}
-b <n> # also unpack %{SOURCE<n>} before changing to directory
(usually combined with -T, -D)
-a <n> # also unpack %{SOURCE<n>} after changing to directory
(usually combined with -T, -D, -c)
%patch[<x>]
-P <n> # apply %{PATCH<n>} instead of %{PATCH<x>}
-p <#> # strip <#> leading slashes and directories from patch filenames
-b <ext> # set the backup file extension to <ext> instead of ".orig"
-E # remove empty output files
-R # apply patch as it would be reversed
Visualization
-------------
Input:
%setup[<x>] # default for <x> is 0
Output:
rm -rf <name>-<version> # changed by -n <dir>, omitted by -D
mkdir <name>-<version> # changed by -n <dir>, added by -c (else omitted)
cd <name>-<version> # changed by -n <dir>, added by -c (else omitted)
gunzip <%{SOURCE<x>} | tar xvf - # omitted by -T, "xvf" replaced with "xf" by -q
gunzip <%{SOURCE<n>} | tar xvf - # added by -b <n>, "xvf" replaced with "xf" by -q
cd <name>-<version> # changed by -n <dir>
gunzip <%{SOURCE<n>} | tar xvf - # added by -a <n>, "xvf" replaced with "xf" by -q
- - -
Keep especially the above "Output" sequence in mind when fiddling around
with the %setup options. Then %setup is no more mystery and can be used
straight-forward...
Ralf S. Engelschall
[EMAIL PROTECTED]
www.engelschall.com
______________________________________________________________________
The OpenPKG Project www.openpkg.org
Developer Communication List [EMAIL PROTECTED]
