Gleaned from Linux Server Hacks
ci -i filename
co -l filename
and edit as you wish.ci -u filename
to check in changes.The initial time you checkout the copy, it will be locked, and this can cause problems if someone else wishes to edit it; you should probably edit it once and put in the version placeholder in comments somewhere at the top or bottom:
$VERSION$
and then check it back in with the -u
flag to unlock it.
I stopped at Borders in downtown Burlington on New Year's Eve day, and found a book called Linux Server Hacks. I loved it immediately, but I wasn't quite willing to shell out $25 for such a slim volume, even if it did have many tidbits I could immediately use.
When I told my co-worker, Rob, about it, it turned out he already had the book, and brought it in to work for me to borrow the next day.
My nose has barely been out of it since. I've done such things as:
/etc/init.d
scripts so that I can start, stop, and reload the firewall at
will.-e
switch; it just seems to cumbersome. However, combine it with the -p
and/or -i
switch, and you can use perl as a filter on globbed files!movein.sh
turned my life around when it came to working on
the servers. I now have a .skel directory on my work machine that contains
links to oft-used configuration files and directories, as well as to my
~/bin
directory; this allows me to then type movein.sh server
and have all
these files uploaded to the server. I can now use vim, screen, and other
programs on any system we have in exactly the manner I expect to.mod_rewrite
hacks,
how to make your directory indexes show full file names, and more; as well as
how to monitor your mysql processes and, if necessary, kill them. I'm also
very interested in how to use MySQL as an authentication backend for an FTP
daemon — it could give us very fine-grained control of our webserver for
editors.And that's just the tip of the iceberg. All in all, I highly recommend the book — though most likely as a book to check out from the library for a few weeks, digest, put into practice, and return. The hacks are so damn useful, I've found that after using one, I don't need to refer to that one ever again. But that's the point.
Tonight was Papa night, which meant that I got to look after Maeve while Jen worked late doing a group at work. Last week, Maeve and I established that Papa Night would always include going to the bookstore, which means Barnes & Noble in South Burlington.
Last week, Maeve was perfectly content to look at books by herself, and didn't want me interfering, so I decided this week to grab a book for myself to peruse while she was busy. It didn't work as I intended — Maeve saw that I wasn't paying full attention to her, and then demanded my attention — but I was able to look through some of the new items in the second edition of The Perl Cookbook.
Among them were:
Setting up both an XML-RPC server and client, using SOAP::Lite
Setting up both a SOAP-RPC server and client, using SOAP::Lite
and other
modules; I could have used this in ROX::Filer
to communicate with ROX
instead of using the filer's RPC call.
Better coverage of DBI (it actually covered it!):
When you expect only a single row, this is a nice way to grab it:
$row = $dbi->selectrow_(array|hash)ref($statement)
This is a great way to grab a bunch of columns from a large resultset:
$results = $dbi->selectall_hashref($sql);
foreach $record (keys(%{$results})) {
print $results->{$record}{fieldname};
}
This one is nice for a large resultset from which you only want one column:
$results = $dbi->selectcol_arrayref($sql);
foreach $result (@{$results}) {
print $result;
}
If you need to quote values before inserting them, try:
$quoted = $dbi->quote($unquoted);
$sql = "UPDATE table SET textfield = $quoted";
If you need to check for errors, don't check with each DBI
call; instead,
wrap all of them in an eval statement:
eval {
$sth = $dbi->prepare($sql);
$sth->do;
while ($row = $sth->fetchrow_hashref) {
...
}
}
if ($@) {
print $DBI::errstr;
}
Coverage of templating, including Text::Template
(very interesting!)
Whole new chapters on mod_perl
and XML (including DOM!) which I didn't really even get to peruse.
autouse pragma: if you use:
use autouse Module::Name;
perl will use the module at runtime instead of compiletime; basically, it
only uses it if it actually needs it (i.e., if it encounters code that
utilizes functionality from that module). It's a good way to keep down on the
bloat — I should use this with librox-perl, and possibly with CGI::App
.