summaryrefslogtreecommitdiffstats
path: root/zarb-ml/mageia-dev/attachments/20110113/b99ffec2
diff options
context:
space:
mode:
Diffstat (limited to 'zarb-ml/mageia-dev/attachments/20110113/b99ffec2')
-rw-r--r--zarb-ml/mageia-dev/attachments/20110113/b99ffec2/attachment-0001.html40
-rw-r--r--zarb-ml/mageia-dev/attachments/20110113/b99ffec2/attachment.html40
2 files changed, 80 insertions, 0 deletions
diff --git a/zarb-ml/mageia-dev/attachments/20110113/b99ffec2/attachment-0001.html b/zarb-ml/mageia-dev/attachments/20110113/b99ffec2/attachment-0001.html
new file mode 100644
index 000000000..9f4b91d3a
--- /dev/null
+++ b/zarb-ml/mageia-dev/attachments/20110113/b99ffec2/attachment-0001.html
@@ -0,0 +1,40 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
+<html>
+<head>
+ <meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
+</head>
+<body bgcolor="#ffffff" text="#000000">
+Daniel Kreuter wrote:
+<blockquote
+ cite="mid:AANLkTik_goSz3Rybp0M0p5YJZ94phafiKWPhPNjvLRqh@mail.gmail.com"
+ type="cite"><br>
+I don't agree with you at that point. I always take the tarball from <a
+ moz-do-not-send="true" href="http://eclipse.org">eclipse.org</a> (for
+eclipse) or <a moz-do-not-send="true" href="http://netbeans.org">netbeans.org</a>
+(for netbeans) instead of the one's of repository provided by the
+distro.<br>
+The reason is quite simple, the one's mentioned above are newer than
+the one's in the repos (often, not always but everywhere i looked for
+it, it was so)<br>
+ <br>
+But there may be people who will first look in rpmdrake or urpmi that's
+right, but not everybody.<br>
+</blockquote>
+<br>
+Well, I didn't say *everybody*, I was referring to newer users who
+might actually do what we tell them to :-)<br>
+<br>
+The problem is that there is a very high probability that anyone who
+installs from an RPM will then install plugins directly.  Given that, I
+would question even packaging the base product as an RPM.<br>
+<br>
+Then again, there's the concern about whether an RPM-provided package
+may provide something a tarball does not, or vice-versa.  For example,
+the MDV Java RPMs play all sorts of games with /etc/alternatives and
+don't (I think) set environment variables like JAVA_HOME.  This means
+that tarball installs of things like Ant need manual tweaking to work
+correctly with an RPM-based JDK.  The RPM-based Ant has wrapper scripts
+which depend on the /etc/alternatives stuff and determine JAVA_HOME on
+the fly.<br>
+</body>
+</html>
diff --git a/zarb-ml/mageia-dev/attachments/20110113/b99ffec2/attachment.html b/zarb-ml/mageia-dev/attachments/20110113/b99ffec2/attachment.html
new file mode 100644
index 000000000..9f4b91d3a
--- /dev/null
+++ b/zarb-ml/mageia-dev/attachments/20110113/b99ffec2/attachment.html
@@ -0,0 +1,40 @@
+<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
+<html>
+<head>
+ <meta content="text/html; charset=UTF-8" http-equiv="Content-Type">
+</head>
+<body bgcolor="#ffffff" text="#000000">
+Daniel Kreuter wrote:
+<blockquote
+ cite="mid:AANLkTik_goSz3Rybp0M0p5YJZ94phafiKWPhPNjvLRqh@mail.gmail.com"
+ type="cite"><br>
+I don't agree with you at that point. I always take the tarball from <a
+ moz-do-not-send="true" href="http://eclipse.org">eclipse.org</a> (for
+eclipse) or <a moz-do-not-send="true" href="http://netbeans.org">netbeans.org</a>
+(for netbeans) instead of the one's of repository provided by the
+distro.<br>
+The reason is quite simple, the one's mentioned above are newer than
+the one's in the repos (often, not always but everywhere i looked for
+it, it was so)<br>
+ <br>
+But there may be people who will first look in rpmdrake or urpmi that's
+right, but not everybody.<br>
+</blockquote>
+<br>
+Well, I didn't say *everybody*, I was referring to newer users who
+might actually do what we tell them to :-)<br>
+<br>
+The problem is that there is a very high probability that anyone who
+installs from an RPM will then install plugins directly.  Given that, I
+would question even packaging the base product as an RPM.<br>
+<br>
+Then again, there's the concern about whether an RPM-provided package
+may provide something a tarball does not, or vice-versa.  For example,
+the MDV Java RPMs play all sorts of games with /etc/alternatives and
+don't (I think) set environment variables like JAVA_HOME.  This means
+that tarball installs of things like Ant need manual tweaking to work
+correctly with an RPM-based JDK.  The RPM-based Ant has wrapper scripts
+which depend on the /etc/alternatives stuff and determine JAVA_HOME on
+the fly.<br>
+</body>
+</html>