| Stephen Gran on 11 Nov 2003 09:56:03 -0500 |
|
On Tue, Nov 11, 2003 at 12:45:22AM -0500, Michael C. Toren said: > On Mon, Nov 10, 2003 at 10:45:00PM -0500, Stephen Gran wrote: > > > The "[:space:]" needs to occur within a character class, which > > > effectively requires the use of double square brackets. Try: > > > > > > egrep -v '^[[:space:]]*(#|$)' > > > > Ah that explains my problem. Thank you very much. Just to mke sure I > > am parsing your solution correctly, I read this as: if the line begins > > with 0 or more spaces, followed by either # or newline, then ignore it - > > yes? That looks like exactly what I want. > > Yup, you got it. Ofcourse, if you'll be doing even further analysis of > the output of the above grep command, you may want to investigate writing > your application in a language more suited to text processing, such as > perl. I would have loved to, but this may be run before /usr is mounted, so relying on perl may not work. If I could have used perl, I wouldn't have had to write in :) On Tue, Nov 11, 2003 at 01:06:15AM -0500, Paul said: > I know this isn't what you want to do, but I usually just put > variables > directly into the file and include the file in the main program. That > way there is no need to parse anything. Unfortunately, that's not going to work either - this is going in a debian package, and I want to rely on Debian's conffile handling, so I need the main script to remain unmodified, while the conffile that it reads is freely modifiable, without being overridden. Thanks for all the help, -- -------------------------------------------------------------------------- | Stephen Gran | Keep your eyes wide open before | | steve@lobefin.net | marriage, half shut afterwards. -- | | http://www.lobefin.net/~steve | Benjamin Franklin | -------------------------------------------------------------------------- Attachment:
pgpxaX5aQ7PmV.pgp
|
|