Skip to content
Switch branches/tags
Go to file
Cannot retrieve contributors at this time
105 lines (76 sloc) 3.11 KB
# warning-lifetime.R, 26 May 20
# Data from:
# The Life and Death of Statically Detected Vulnerabilities: an Empirical Study
# Massimiliano Di Penta and Luigi Cerulo and Lerina Aversano
# Example from:
# Evidence-based Software Engineering: based on the publicly available data
# Derek M. Jones
# TAG static-analysis mistake_lifetime mistake_fixed mistake_code-change
plot_layout(2, 1)
prog_hist=read.csv(paste0(ESEUR_dir, "faults/tool-rep/", prog_str),
# Extract and order by SNAPSHOT so we can lookup TIMESTAMP
prog_hist=subset(prog_hist, TOOL == "Splint")
prog_hist=prog_hist[order(prog_hist$SNAPSHOT), ]
# Zero base and convert from seconds to days
plot_surv=function(df, tool_hist, col_str)
df_surv=Surv(df$END-df$START, df_event)
df_mod=survfit(df_surv ~1)
lines(df_mod, col=col_str)
plot_categ=function(df, tool_hist, prog_str)
splint_mem=subset(df, CATEG == "Memory Problem")
splint_type=subset(df, CATEG == "Type Mismatch")
plot(0, type="n",
xaxs="i", yaxs="i",
xlim=c(0, 1000), ylim=c(0, 1),
xlab="Days since created", ylab="Coding mistake survival\n")
text(250, 0.92, prog_str, cex=1.3)
plot_surv(splint_mem, tool_hist, pal_col[1])
plot_surv(splint_type, tool_hist, pal_col[2])
legend(x="topright", legend=c("Memory problem", "Type mismatch"),
bty="n", fill=pal_col, cex=1.2)
# ID: the unique ID
# FILE: the file where it was detected
# LINE: the source line where it was detected
# CATEG: the vulnerability category according to our classification
# START: snapshot where the vulnerability appeared
# END: snapshot where the vulnerability disappeared or was removed (NA if live at the last snapshot analyzed)
# WASREM: 1 if the vulnerability was removed, 0 otherwise
# WASREMBUG: 1 if the vulnerability was removed because of a bug fixing, 0 otherwise
# DISAP: 1 if the vulnerability disappeared, 0 otherwise
# DISAPBUG: 1 if the vulnerability disappeared because of a bug fixing, 0 otherwise
# HOWDISAP: NA if the vulnerability did not disappear, CHG if it disappeared because of a change in the source code line, COCHG if it disappeared because of a co-change
# LONGDESCR: long vulnerability description as printed by the detection tool
# SYSTEM: analyzed system
splint_rep=read.csv(paste0(ESEUR_dir, "faults/tool-rep/splint.csv.xz"),
# table(splint_rep$WASREMBUG)
# Remove cases where a mistake was fixed
splint_rep=subset(splint_rep, !WASREMBUG)
splint_samba=subset(splint_rep, SYSTEM == "samba")
splint_squid=subset(splint_rep, SYSTEM == "squid")
# plot_surv(splint_samba, samba_hist, pal_col[1])
# par(new=TRUE)
# plot_surv(splint_squid, squid_hist, pal_col[2])
plot_categ(splint_samba, samba_hist, "Samba\n")
plot_categ(splint_squid, squid_hist, "Squid\n")