Remove paragraph from file

Actually, sed can also take ranges. This command will delete all lines between Match user foo and the first empty line (inclusive):

$ sed '/Match user foo/,/^\s*$/{d}' file


Match user bar
        ChrootDirectory /NAS/bar.co.uk/
        ForceCommand internal-sftp
        AllowTcpForwarding no
        GatewayPorts no
        X11Forwarding no



Match user baz
        ChrootDirectory /NAS/baz.com/
        ForceCommand internal-sftp
        AllowTcpForwarding no
        GatewayPorts no
        X11Forwarding no

Personally, however, I would do this using perl's paragraph mode (-00) that has the benefit of removing the leading blank lines:

$ perl -00ne 'print unless /Match user foo/' file
Match user bar
        ChrootDirectory /NAS/bar.co.uk/
        ForceCommand internal-sftp
        AllowTcpForwarding no
        GatewayPorts no
        X11Forwarding no

Match user baz
        ChrootDirectory /NAS/baz.com/
        ForceCommand internal-sftp
        AllowTcpForwarding no
        GatewayPorts no
        X11Forwarding no

In both cases, you can use -i to edit the file in place (these will create a backup of the original called file.bak):

sed -i.bak '/Match user foo/,/^\s*$/{d}' file

or

perl -i.bak -00ne 'print unless /Match user foo/' file 

Slightly more complicated than terdon’s sed answer:

awk '/foo/ {suppress=1} /^\s*$/ {suppress=0} !suppress' file.php

produces almost exactly the same result as terdon’s answer:

...



Match user bar
        ChrootDirectory /NAS/bar.co.uk/
        ForceCommand internal-sftp
        AllowTcpForwarding no
        GatewayPorts no
        X11Forwarding no



Match user baz
        ChrootDirectory /NAS/baz.com/
        ForceCommand internal-sftp
        AllowTcpForwarding no
        GatewayPorts no
        X11Forwarding no

i.e., it deletes (suppresses) every line starting with the one that matches foo up to the first subsequent line that contains only whitespace.  Lines 8, 9, and 10, which are blank in the file.php input file (between user foo and user bar) come out in the output.  By contrast, terdon’s answer deletes every line starting with the one that matches foo up through the first subsequent line that contains only whitespace; so line 8 is deleted, but 9 and 10 come through.

These are not exactly what the user asked for.

awk '/foo/ {suppress=1}
     /^\s*$/ && suppress==1 {suppress=2}
     /[^\s]/ && suppress==2 {suppress=0}
     !suppress' file.php

is.  (This is broken into multiple lines for readability; it could be entered all on one line.)  When it sees foo, it goes into suppress mode #1 (suppress=1).  When it sees a blank line while in suppress mode #1, it switches into suppress mode #2.  When it sees a non-blank line while in suppress mode #2, it switches into mode 0.  Finally, it does the obvious — print lines that are not suppressed.