<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
On 11/05/2010 03:48 AM, David Peterson wrote:
<blockquote
cite="mid:OF6D40BA30.8A4E72DF-ON872577D2.000E999B-882577D2.000F62CD@us.ibm.com"
type="cite"><font face="sans-serif" size="2">Looking in rpmdb.h, I
see the following
function:</font>
<br>
<br>
<font face="sans-serif" size="2"> rpmdbMatchIterator
rpmdbInitIterator(rpmdb
db, rpmTag rpmtag, const void* keyp, size_t keylen);</font>
<br>
<br>
<font face="sans-serif" size="2">My question is, what is the set
of valid
inputs that the function is supposed to</font>
<br>
<font face="sans-serif" size="2">handle for the 'keyp' parameter?
For
instance, is a regular expression for</font>
<br>
<font face="sans-serif" size="2">matching package names ok, or is
the
set of valid inputs more constrained</font>
<br>
<font face="sans-serif" size="2">than that?</font><br>
</blockquote>
It only allows either a string that is matched as is or NULL for
all. You can use rpmdbSetIteratorRE to restrict the results
(typically using RPMTAG_PACKAGES, NULL for <font face="sans-serif"
size="2">rpmdbInitIterator</font>). Unfortunately this does not
give you any performance benefit. RPM still loads and inspects all
headers. Sorry, the API (and the implementation) isn't very helpful
in this area...<br>
<br>
Florian<br>
</body>
</html>