Do set-diff for bash state in C (#346)

Signed-off-by: Konstantinos Kallas <[email protected]>

Author: angelhof

compiler/pash_init_setup.sh

@@ -11,6 +11,8 @@ export distro=$(printf '%s\n' "$distro" | LC_ALL=C tr '[:upper:]' '[:lower:]')
 
 ## File directory
 export RUNTIME_DIR=$(dirname "${BASH_SOURCE[0]}")
+## TODO: Is there a better way to do this?
+export RUNTIME_LIBRARY_DIR="$RUNTIME_DIR/../runtime/"
 export PASH_REDIR="&2"
 export PASH_DEBUG_LEVEL=0
 

compiler/pash_set_from_to.sh

@@ -4,10 +4,9 @@ from_set=${1?From set not given}
 to_set=${2?To set not given}
 
 ## Finds the difference of set variables (removing the c, s one since it cannot be actually set and unset)
-##
-## TODO: This is very inefficient! It can probably happen in a single C command.
-pash_set_to_add="$(comm -23 <(echo $to_set | sed -E 's/(.)/\1\n/g' | sort) <(echo $from_set | sed -E 's/(.)/\1\n/g' | sort) | grep -v 'c' | grep -v 's' || true )"
-pash_set_to_remove="$(comm -13 <(echo $to_set | sed -E 's/(.)/\1\n/g' | sort) <(echo $from_set | sed -E 's/(.)/\1\n/g' | sort) | grep -v 'c' | grep -v 's' || true )"
+pash_redir_output echo "From set: $from_set"
+pash_redir_output echo "To set: $to_set"
+IFS=',' read -r pash_set_to_remove pash_set_to_add <<< "$($RUNTIME_LIBRARY_DIR/set-diff $from_set $to_set)"
 pash_redir_output echo "To add: $pash_set_to_add"
 pash_redir_output echo "To remove: $pash_set_to_remove"
 pash_redir_all_output_always_execute set "-$pash_set_to_add"

runtime/.gitignore

@@ -4,5 +4,6 @@ r_split
 r_merge
 r_unwrap
 r_wrap
+set-diff
 tests/perf*
-tests/*out
+tests/*out

runtime/Makefile

@@ -1,6 +1,6 @@
 .PHONY: all eager-debug split-debug clean agg
 
-all: eager split r-merge r-wrap r-split r-unwrap dgsh-tee agg
+all: eager split r-merge r-wrap r-split r-unwrap dgsh-tee agg set-diff
 
 CFLAGS=-Wall
 
@@ -50,6 +50,9 @@ r-wrap: r_wrap.c r_split.h
 r-unwrap: r_unwrap.c r_split.h
 	gcc ${CFLAGS} r_unwrap.c -o r_unwrap
 
+set-diff: set-diff.c
+	gcc ${CFLAGS} set-diff.c -o set-diff
+
 
 libdgsh_a_SOURCES = negotiate.c $(DGSH_ASSEMBLY_FILE)
 dgsh_tee_SOURCES = dgsh-tee.c

runtime/set-diff.c

@@ -0,0 +1,104 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+int comp (const void * elem1, const void * elem2) 
+{
+    char f = *((char*)elem1);
+    char s = *((char*)elem2);
+    if (f > s) return  1;
+    if (f < s) return -1;
+    return 0;
+}
+
+int main(int argc, char* argv[]) {
+    // arg#1 -> old set
+    // arg#2 -> new set
+    if (argc != 3) {
+        printf("ERROR: wrong input!\n");
+        exit(1);
+    }
+
+    // Returns the - and + difference of the two sets of sorted characters.
+    //
+    // First ensures that characters are already sorted
+    //
+    // It requires that there are no duplicates
+    int fromLen = strlen(argv[1]);
+    int toLen = strlen(argv[2]);
+
+    qsort(argv[1], fromLen, sizeof(char), comp);
+    qsort(argv[2], toLen, sizeof(char), comp);
+
+    int maxLen = (fromLen < toLen) ? toLen : fromLen;
+
+    char* toRemove = malloc((maxLen + 1) * sizeof(char));
+    char* toAdd = malloc((maxLen + 1) * sizeof(char));
+
+    int iFrom = 0; 
+    int iTo = 0;
+    int iRemove = 0; 
+    int iAdd = 0;
+
+    while (iFrom < fromLen && iTo < toLen )
+    {
+        // We do not add or remove -s and -c
+        if (argv[1][iFrom] == 'c' || argv[1][iFrom] == 's')
+        {
+            iFrom++;
+        }
+        else if (argv[2][iTo] == 'c' || argv[2][iTo] == 's')
+        {
+            iTo++;
+        }
+        else if (argv[1][iFrom] < argv[2][iTo])
+        {
+            // Only exists in iFrom, therefore we need to remove
+            toRemove[iRemove] = argv[1][iFrom];
+            iRemove ++;
+            iFrom ++;
+        }
+        else if (argv[1][iFrom] == argv[2][iTo])
+        {
+            // Don't add or remove anything!
+            iFrom ++;
+            iTo ++;
+        }
+        else
+        {
+            // Only exists in iTo, therefore we need to add
+            toAdd[iAdd] = argv[2][iTo];
+            iAdd ++;
+            iTo ++;
+        }
+    }
+
+    // If there is something left in iFrom
+    while (iFrom < fromLen)
+    {
+        if (argv[1][iFrom] != 'c' && argv[1][iFrom] != 's')
+        {
+            toRemove[iRemove] = argv[1][iFrom];
+            iRemove ++;
+        }
+        iFrom ++;
+    }
+
+    // If there is something left in iTo
+    while (iTo < toLen)
+    {
+        if (argv[2][iTo] != 'c' && argv[2][iTo] != 's')
+        {
+            toAdd[iAdd] = argv[2][iTo];
+            iAdd ++;
+        }
+        iTo ++;
+    }
+
+    toRemove[iRemove] = '\0';
+    toAdd[iAdd] = '\0';
+
+    printf("%s,%s\n", toRemove, toAdd);
+
+    return 0;
+}