Add Shell dsl type

7 files changed, 101 insertions, 1 deletions

diff --git a/Makefile b/Makefile
index 5f0f2be8..876ac3be 100644
--- a/Makefile
+++ b/Makefile
@@ -26,7 +26,7 @@ CFLAGS=$(CCONFIG) $(EXTRA) $(CWARN) $(G) $(O) $(OSFLAGS)
 LDLIBS=-lgc -lcord -lm -lunistring -lgmp -ldl
 BUILTIN_OBJS=builtins/siphash.o builtins/array.o builtins/bool.o builtins/channel.o builtins/nums.o builtins/functions.o builtins/integers.o \
 						 builtins/pointer.o builtins/memory.o builtins/text.o builtins/thread.o builtins/c_string.o builtins/table.o \
-						 builtins/types.o builtins/util.o builtins/files.o builtins/range.o
+						 builtins/types.o builtins/util.o builtins/files.o builtins/range.o builtins/shell.o
 TESTS=$(patsubst %.tm,%.tm.testresult,$(wildcard test/*.tm))
 
 all: libtomo.so tomo
diff --git a/builtins/shell.c b/builtins/shell.c
new file mode 100644
index 00000000..da6cbe5b
--- /dev/null
+++ b/builtins/shell.c
@@ -0,0 +1,67 @@
+// Boolean methods/type info
+#include <stdbool.h>
+#include <stdint.h>
+
+#include "array.h"
+#include "functions.h"
+#include "integers.h"
+#include "shell.h"
+#include "text.h"
+#include "types.h"
+#include "util.h"
+
+public Pattern_t Shell$escape_text(Text_t text)
+{
+    // TODO: optimize for ASCII and short strings
+    Array_t shell_graphemes = {.atomic=1};
+#define add_char(c) Array$insert(&shell_graphemes, (uint32_t[1]){c}, I_small(0), sizeof(uint32_t))
+    add_char('\'');
+    const char *text_utf8 = Text$as_c_string(text);
+    for (const char *p = text_utf8; *p; p++) {
+        if (*p == '\'') {
+            add_char('\'');
+            add_char('"');
+            add_char('\'');
+            add_char('"');
+            add_char('\'');
+        } else
+            add_char(*p);
+    }
+    add_char('\'');
+#undef add_char
+    return (Text_t){.length=shell_graphemes.length, .tag=TEXT_GRAPHEMES, .graphemes=shell_graphemes.data};
+}
+
+public Text_t Shell$run(Shell_t command, int32_t *status)
+{
+    const char *cmd_str = Text$as_c_string(command);
+    FILE *prog = popen(cmd_str, "r");
+
+    const int chunk_size = 256;
+    char *buf = GC_MALLOC_ATOMIC(chunk_size);
+    Text_t output = Text("");
+    size_t just_read;
+    do {
+        just_read = fread(buf, sizeof(char), chunk_size, prog);
+        if (just_read > 0) {
+            output = Texts(output, Text$from_strn(buf, just_read));
+            buf = GC_MALLOC_ATOMIC(chunk_size);
+        }
+    } while (just_read > 0);
+
+    if (status)
+        *status = WEXITSTATUS(pclose(prog));
+    else
+        pclose(prog);
+
+    return Text$trim(output, Pattern("{1 nl}"), false, true);
+}
+
+public const TypeInfo Shell$info = {
+    .size=sizeof(Shell_t),
+    .align=__alignof__(Shell_t),
+    .tag=TextInfo,
+    .TextInfo={.lang="Shell"},
+};
+
+// vim: ts=4 sw=0 et cino=L2,l1,(0,W4,m1,\:0
diff --git a/builtins/shell.h b/builtins/shell.h
new file mode 100644
index 00000000..0429d496
--- /dev/null
+++ b/builtins/shell.h
@@ -0,0 +1,22 @@
+#pragma once
+
+// A lang for Shell Command Language
+
+#include <gc/cord.h>
+#include <stdbool.h>
+#include <stdint.h>
+
+#include "types.h"
+#include "datatypes.h"
+
+#define Shell_t Text_t
+#define Shell(text) ((Shell_t)Text(text))
+#define Shells(...) ((Shell_t)Texts(__VA_ARGS__))
+
+Text_t Shell$run(Shell_t command, int32_t *status);
+Shell_t Shell$escape_text(Text_t text);
+
+extern const TypeInfo Shell$info;
+
+// vim: ts=4 sw=0 et cino=L2,l1,(0,W4,m1,\:0
+
diff --git a/builtins/tomo.h b/builtins/tomo.h
index 8d6661ca..68e115bb 100644
--- a/builtins/tomo.h
+++ b/builtins/tomo.h
@@ -22,6 +22,7 @@
 #include "nums.h"
 #include "pointer.h"
 #include "range.h"
+#include "shell.h"
 #include "siphash.h"
 #include "table.h"
 #include "text.h"
diff --git a/compile.c b/compile.c
index e53b43bb..0be3aa89 100644
--- a/compile.c
+++ b/compile.c
@@ -2994,6 +2994,8 @@ CORD compile_type_info(env_t *env, type_t *t)
             return "&Text$info";
         else if (streq(text->lang, "Pattern"))
             return "&Pattern$info";
+        else if (streq(text->lang, "Shell"))
+            return "&Shell$info";
         return CORD_all("(&", namespace_prefix(text->env->libname, text->env->namespace->parent), text->lang, ")");
     }
     case StructType: {
diff --git a/environment.c b/environment.c
index 666149cc..39a3e24a 100644
--- a/environment.c
+++ b/environment.c
@@ -243,6 +243,10 @@ env_t *new_compilation_unit(CORD *libname)
         {"Pattern", Type(TextType, .lang="Pattern", .env=namespace_env(env, "Pattern")), "Pattern_t", "Pattern$info", TypedArray(ns_entry_t,
             {"escape_text", "Pattern$escape_text", "func(text:Text)->Pattern"},
         )},
+        {"Shell", Type(TextType, .lang="Shell", .env=namespace_env(env, "Shell")), "Shell_t", "Shell$info", TypedArray(ns_entry_t,
+            {"escape_text", "Shell$escape_text", "func(text:Text)->Shell"},
+            {"run", "Shell$run", "func(command:Shell, status=!&Int32?)->Text"},
+        )},
         {"Text", TEXT_TYPE, "Text_t", "Text$info", TypedArray(ns_entry_t,
             {"find", "Text$find", "func(text:Text, pattern:Pattern, start=1, length=!&Int64)->Int"},
             {"find_all", "Text$find_all", "func(text:Text, pattern:Pattern)->[Text]"},
diff --git a/parse.c b/parse.c
index 2f3a8c3d..1e421085 100644
--- a/parse.c
+++ b/parse.c
@@ -1185,6 +1185,8 @@ PARSER(parse_text) {
         } else if (strchr(interp_chars, *pos)) {
             open_interp = *pos;
             ++pos;
+        } else if (*pos == '(') {
+            open_interp = '@'; // For shell commands
         }
         static const char *quote_chars = "\"'`|/;([{<";
         if (!strchr(quote_chars, *pos))
@@ -1195,6 +1197,8 @@ PARSER(parse_text) {
 
         if (!lang && open_quote == '/')
             lang = "Pattern";
+        else if (!lang && open_quote == '(')
+            lang = "Shell";
     } else {
         return NULL;
     }