summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorPatrick Steinhardt <ps@pks.im>2017-07-14 10:57:49 +0200
committerPatrick Steinhardt <ps@pks.im>2017-08-25 16:42:11 +0200
commit57bc9daba35dcabdacc40fd8bab0577fa7e86367 (patch)
tree5f58c817914270de0e0070894bf4ef07b1c4758c
parent0a93ded1b931c0cd3476300a0794dc2f3e946356 (diff)
downloadlibgit2-57bc9daba35dcabdacc40fd8bab0577fa7e86367.tar.gz
patch_parse: implement state machine for parsing patch headers
Our code parsing Git patch headers is rather lax in parsing headers of a Git-style patch. Most notably, we do not care for the exact order in which header lines appear and as such, we may parse patch files which are not really valid after all. Furthermore, the state transitions inside of the parser are not as obvious as they could be, making it harder than required to follow its logic. To improve upon this situation, this patch introduces a real state machine to parse the patches. Instead of simply parsing each line without caring for previous state and the exact ordering, we define a set of states with their allowed transitions. This makes the patch parser more strict in only allowing valid successions of header lines. As the transition table is defined inside of a single structure with the expected line, required state as well as the state that we end up in, all state transitions are immediately obvious from just having a look at this structure. This improves both maintainability and eases reasoning about the patch parser.
-rw-r--r--src/patch_parse.c127
1 files changed, 81 insertions, 46 deletions
diff --git a/src/patch_parse.c b/src/patch_parse.c
index a531eec4d..2f0b2570d 100644
--- a/src/patch_parse.c
+++ b/src/patch_parse.c
@@ -372,31 +372,73 @@ static int parse_header_dissimilarity(
return 0;
}
+static int parse_header_start(git_patch_parsed *patch, git_patch_parse_ctx *ctx)
+{
+ if (parse_header_path(&patch->header_old_path, ctx) < 0)
+ return parse_err("corrupt old path in git diff header at line %"PRIuZ,
+ ctx->line_num);
+
+ if (parse_advance_ws(ctx) < 0 ||
+ parse_header_path(&patch->header_new_path, ctx) < 0)
+ return parse_err("corrupt new path in git diff header at line %"PRIuZ,
+ ctx->line_num);
+
+ return 0;
+}
+
+typedef enum {
+ STATE_START,
+
+ STATE_DIFF,
+ STATE_FILEMODE,
+ STATE_MODE,
+ STATE_INDEX,
+ STATE_PATH,
+
+ STATE_SIMILARITY,
+ STATE_RENAME,
+ STATE_COPY,
+
+ STATE_END,
+} parse_header_state;
+
typedef struct {
const char *str;
+ parse_header_state expected_state;
+ parse_header_state next_state;
int(*fn)(git_patch_parsed *, git_patch_parse_ctx *);
-} header_git_op;
-
-static const header_git_op header_git_ops[] = {
- { "diff --git ", NULL },
- { "@@ -", NULL },
- { "GIT binary patch", NULL },
- { "Binary files ", NULL },
- { "--- ", parse_header_git_oldpath },
- { "+++ ", parse_header_git_newpath },
- { "index ", parse_header_git_index },
- { "old mode ", parse_header_git_oldmode },
- { "new mode ", parse_header_git_newmode },
- { "deleted file mode ", parse_header_git_deletedfilemode },
- { "new file mode ", parse_header_git_newfilemode },
- { "rename from ", parse_header_renamefrom },
- { "rename to ", parse_header_renameto },
- { "rename old ", parse_header_renamefrom },
- { "rename new ", parse_header_renameto },
- { "copy from ", parse_header_copyfrom },
- { "copy to ", parse_header_copyto },
- { "similarity index ", parse_header_similarity },
- { "dissimilarity index ", parse_header_dissimilarity },
+} parse_header_transition;
+
+static const parse_header_transition transitions[] = {
+ /* Start */
+ { "diff --git " , STATE_START, STATE_DIFF, parse_header_start },
+
+ { "deleted file mode " , STATE_DIFF, STATE_FILEMODE, parse_header_git_deletedfilemode },
+ { "new file mode " , STATE_DIFF, STATE_FILEMODE, parse_header_git_newfilemode },
+ { "old mode " , STATE_DIFF, STATE_MODE, parse_header_git_oldmode },
+ { "new mode " , STATE_MODE, STATE_END, parse_header_git_newmode },
+
+ { "index " , STATE_FILEMODE, STATE_INDEX, parse_header_git_index },
+ { "index " , STATE_DIFF, STATE_INDEX, parse_header_git_index },
+ { "index " , STATE_END, STATE_INDEX, parse_header_git_index },
+
+ { "--- " , STATE_INDEX, STATE_PATH, parse_header_git_oldpath },
+ { "+++ " , STATE_PATH, STATE_END, parse_header_git_newpath },
+ { "GIT binary patch" , STATE_INDEX, STATE_END, NULL },
+ { "Binary files " , STATE_INDEX, STATE_END, NULL },
+
+ { "similarity index " , STATE_DIFF, STATE_SIMILARITY, parse_header_similarity },
+ { "dissimilarity index ", STATE_DIFF, STATE_SIMILARITY, parse_header_dissimilarity },
+ { "rename from " , STATE_SIMILARITY, STATE_RENAME, parse_header_renamefrom },
+ { "rename old " , STATE_SIMILARITY, STATE_RENAME, parse_header_renamefrom },
+ { "copy from " , STATE_SIMILARITY, STATE_COPY, parse_header_copyfrom },
+ { "rename to " , STATE_RENAME, STATE_END, parse_header_renameto },
+ { "rename new " , STATE_RENAME, STATE_END, parse_header_renameto },
+ { "copy to " , STATE_COPY, STATE_END, parse_header_copyto },
+
+ /* Next patch */
+ { "diff --git " , STATE_END, 0, NULL },
+ { "@@ -" , STATE_END, 0, NULL },
};
static int parse_header_git(
@@ -405,44 +447,32 @@ static int parse_header_git(
{
size_t i;
int error = 0;
-
- /* Parse the diff --git line */
- if (parse_advance_expected_str(ctx, "diff --git ") < 0)
- return parse_err("corrupt git diff header at line %"PRIuZ, ctx->line_num);
-
- if (parse_header_path(&patch->header_old_path, ctx) < 0)
- return parse_err("corrupt old path in git diff header at line %"PRIuZ,
- ctx->line_num);
-
- if (parse_advance_ws(ctx) < 0 ||
- parse_header_path(&patch->header_new_path, ctx) < 0)
- return parse_err("corrupt new path in git diff header at line %"PRIuZ,
- ctx->line_num);
+ parse_header_state state = STATE_START;
/* Parse remaining header lines */
- for (parse_advance_line(ctx);
- ctx->remain_len > 0;
- parse_advance_line(ctx)) {
-
+ for (; ctx->remain_len > 0; parse_advance_line(ctx)) {
bool found = false;
if (ctx->line_len == 0 || ctx->line[ctx->line_len - 1] != '\n')
break;
- for (i = 0; i < ARRAY_SIZE(header_git_ops); i++) {
- const header_git_op *op = &header_git_ops[i];
- size_t op_len = strlen(op->str);
+ for (i = 0; i < ARRAY_SIZE(transitions); i++) {
+ const parse_header_transition *transition = &transitions[i];
+ size_t op_len = strlen(transition->str);
- if (memcmp(ctx->line, op->str, min(op_len, ctx->line_len)) != 0)
+ if (transition->expected_state != state ||
+ memcmp(ctx->line, transition->str, min(op_len, ctx->line_len)) != 0)
continue;
+ state = transition->next_state;
+
/* Do not advance if this is the patch separator */
- if (op->fn == NULL)
+ if (transition->fn == NULL)
goto done;
parse_advance_chars(ctx, op_len);
- if ((error = op->fn(patch, ctx)) < 0)
+ if ((error = transition->fn(patch, ctx)) < 0)
goto done;
parse_advance_ws(ctx);
@@ -456,7 +486,7 @@ static int parse_header_git(
found = true;
break;
}
-
+
if (!found) {
error = parse_err("invalid patch header at line %"PRIuZ,
ctx->line_num);
@@ -464,6 +494,11 @@ static int parse_header_git(
}
}
+ if (state != STATE_END) {
+ error = parse_err("unexpected header line %"PRIuZ, ctx->line_num);
+ goto done;
+ }
+
done:
return error;
}