On 11 Apr 1993 21:24:20 -0400 (EDT), "Pres Smith" <cons052@titan.ucs.umass.edu> wrote:
>
> Is there a sub-program for chopping up archives into "byte"-sized
> chunks of about 100K automatically for Anastasios Kotsikonas' unix
> list server program? Any program? It takes too much time for systems
> people to do this by "hand" now.
>
At the current time I do not think this is implemented for archives.
It IS available when archiving using the "farch" command. I will make sure
that Tasos puts it on the "To: do" list.
>
> Also, are such archives searchable by LDbase rules?
>
The "search" command was added in ListServer 6.0, and a copy of the
"help search" is included at the end of this message. Tasos will have
to clear this up, and he's on vacation.
>
> It's taking me too much time to retrieve old files at subscriber
> requests by "memory" and I don't have a large enough hard disk to hold
> my own "saves" of old posts for use of some file search program. If I
> did, what do listowners use--Ask Sam? Zyindex? WordCruncher? Fgrep?
>
> Thanks Prescott Smith, listowner, Ednet pgsmith@educ.umass.edu
>
>
Syntax: search <archive | path-to-archive>] [/password] [-all] <pattern>
Search all files of the specified archive (and all of its subarchives
if -all is specified) for lines that match the pattern. The pattern
can be an egrep(1)-style regular expression with support for the
following additional operators: '~', if leading the regular expression
it reverses its meaning; '|' and '&' separate multiple regular
expressions (logical OR and AND); '<' '>' group regular expressions
(we preserve the meaning of the parentheses from ed(1), and remove the
meaning of < and > from ed(1) since in the ListServer context they are
either the default, or inappropriate). These can be used literally by
escaping them with '\'. In addition, the following characters should
be defined in matched pairs: (), <>, []. The pattern may be enclosed
in single or double quotes.
Pattern matching is case insensitive.
Certain archives may be private, and in this case you have to
specify a password for accessing them. The slash is required.
Different archives may have different passwords.
The archiving system is hierarchical. Therefore subarchives
may have the same names; they can be distinguished by the path
(the branch in the hierarchy) to them. For example, the archives
unix and pub/unix are distinct. 'path-to-archive' is a UNIX style
path (such as pub/unix) -- i.e. a '/' is used to move through the
branches of the hierarchy. An 'index' request always reports paths
to archives for your convinience.
Examples:
search listserver -all "<oranges &~apples>|.*andarin[sz]?"
search iul '\|.*\|.$'
Notes:
- . matches ANY character including newline, so to find all lines
that contain a newline only, one should use '^.$' instead of '^$'
- If you do not quote your pattern, it will include any blanks you
may have appended after it by mistake, and it will not include any
blanks that may start the regular expression.
|
|
