From 3cc59d796433b903dafd5a39809c19f5965366ed Mon Sep 17 00:00:00 2001
From: Evan Battaglia <elb at eecs point berkeley>
Date: Fri, 9 May 2008 11:34:31 -0700
Subject: [PATCH] doc remix

---
 Svelte/doc/COMMON-SLICING-ERRORS                | 9 +++++++++
 Svelte/{ => doc}/HACKING.txt                    | 0
 Svelte/{ => doc}/HOWTOINSTALLSOURCE-GITBASH.txt | 0
 Svelte/{ => doc}/HOWTOINSTALLSOURCE-GITGUI.txt  | 0
 4 files changed, 9 insertions(+)
 create mode 100644 Svelte/doc/COMMON-SLICING-ERRORS
 rename Svelte/{ => doc}/HACKING.txt (100%)
 rename Svelte/{ => doc}/HOWTOINSTALLSOURCE-GITBASH.txt (100%)
 rename Svelte/{ => doc}/HOWTOINSTALLSOURCE-GITGUI.txt (100%)

diff --git a/Svelte/doc/COMMON-SLICING-ERRORS b/Svelte/doc/COMMON-SLICING-ERRORS
new file mode 100644
index 0000000..2598116
--- /dev/null
+++ b/Svelte/doc/COMMON-SLICING-ERRORS
@@ -0,0 +1,9 @@
+===Out of memory errors===
+The WALA call graph algorihm and/or slicing algorithm sometimes scales poorly to large or medium-sized programs. You can increase the amount of memory available to the slicing process with the java VM argument <code>-Xmx</code>. If you are building the slicer from source and running from a child Eclipse window, go to the Debug or Run dialog, select the Eclipse Application launch profile you are using, and in the arguments tab, change the VM Arguments so that they contain something like the following:
+
+<pre>-Xmx600M</pre>
+
+This will set the Java heap size to 600 megabytes.
+
+===Cannot find superclass errors with programs using swing===
+By default some Java 6, including the swing libraries, are excluded from the analysis because they are too big. If you are trying to slice a program that uses Swing and it does not work, you can try changing which classes are excluded by editing the file Java60RegressionExclusions.txt (there are several of these files in both the WALA directories and the eclipsethinslicer directory, and it is difficult to determine which is being used, so it is safest to change them all). Note however, that what will likely happen is that the swing libraries will be to big to analyze, and generating a call graph will take too long or run out of memory.
diff --git a/Svelte/HACKING.txt b/Svelte/doc/HACKING.txt
similarity index 100%
rename from Svelte/HACKING.txt
rename to Svelte/doc/HACKING.txt
diff --git a/Svelte/HOWTOINSTALLSOURCE-GITBASH.txt b/Svelte/doc/HOWTOINSTALLSOURCE-GITBASH.txt
similarity index 100%
rename from Svelte/HOWTOINSTALLSOURCE-GITBASH.txt
rename to Svelte/doc/HOWTOINSTALLSOURCE-GITBASH.txt
diff --git a/Svelte/HOWTOINSTALLSOURCE-GITGUI.txt b/Svelte/doc/HOWTOINSTALLSOURCE-GITGUI.txt
similarity index 100%
rename from Svelte/HOWTOINSTALLSOURCE-GITGUI.txt
rename to Svelte/doc/HOWTOINSTALLSOURCE-GITGUI.txt
-- 
2.11.4.GIT