<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
<title></title>
</head>
<body text="#000000" bgcolor="#ffffff">
On 07/09/2009 03:44 AM, Alan Pevec wrote:
<blockquote
cite="mid:2be7262f0907090044t6c8fc28cycc2f3208d38c7d21@mail.gmail.com"
type="cite">
<div class="gmail_quote">On Wed, Jul 8, 2009 at 8:40 PM, Arjun Roy <span
dir="ltr"><<a moz-do-not-send="true" href="mailto:arroy@redhat.com"
target="_blank">arroy@redhat.com</a>></span> wrote:<br>
<blockquote class="gmail_quote"
style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">Dependencies
added include:<br>
-libboost_regex library<br>
-libicu<br>
</blockquote>
<div>hmm, that's an unfortunate dependency: complete libicu adds 17MB
and blacklisting unused parts doesn't help much: <br>
</div>
</div>
libboost_regex.so links with libicuuc.so (1.3M), libicui18n.so (1.6M)
and libicudata.so (14M) which makes 99% of libicu :(<br>
<br>
Could we avoid this by using more lightweight regexp library?<br>
<br>
</blockquote>
Ian Main mentioned pcre,
(<a class="moz-txt-link-freetext" href="http://koji.fedoraproject.org/koji/buildinfo?buildID=89460">http://koji.fedoraproject.org/koji/buildinfo?buildID=89460</a>) which
looks like it doesn't ask for as much. So I could convert my package to
use it.<br>
<br>
In the future, the data we are pulling by reading /proc/cpuinfo (and
anything else needing regexes) would hopefully be exposed by libvirt
anyways, which means no regex dependency at all. I'm looking forward to
that.<br>
<br>
-Arjun<br>
</body>
</html>