From 4b8574aa878f0f21ff5ffdcda27ad510f8a2f911 Mon Sep 17 00:00:00 2001
From: Erica Fischer <enf@felt.com>
Date: Fri, 2 Feb 2024 14:12:34 -0800
Subject: [PATCH] A few more cases to populate the hash in addpool

---
 main.cpp    |  3 ++-
 memfile.cpp |  4 +++-
 memfile.hpp |  2 +-
 pool.cpp    | 25 ++++++++++++++++++-------
 4 files changed, 24 insertions(+), 10 deletions(-)

diff --git a/main.cpp b/main.cpp
index 8748a552..82621b59 100644
--- a/main.cpp
+++ b/main.cpp
@@ -1336,7 +1336,8 @@ std::pair<int, metadata> read_input(std::vector<source> &sources, char *fname, i
 		// To distinguish a null value
 		{
 			struct stringpool p;
-			memfile_write(r->treefile, &p, sizeof(struct stringpool));
+			bool in_memory;
+			memfile_write(r->treefile, &p, sizeof(struct stringpool), in_memory);
 		}
 
 		r->file_bbox[0] = r->file_bbox[1] = UINT_MAX;
diff --git a/memfile.cpp b/memfile.cpp
index a363b0b9..ec2e8adc 100644
--- a/memfile.cpp
+++ b/memfile.cpp
@@ -49,10 +49,11 @@ int memfile_close(struct memfile *file) {
 	return 0;
 }
 
-int memfile_write(struct memfile *file, void *s, long long len) {
+int memfile_write(struct memfile *file, void *s, long long len, bool &in_memory) {
 	// If it is full, append to the file.
 	// If it is not full yet, append to the string in memory.
 
+	in_memory = false;
 	if (file->fp != NULL) {
 		if (fwrite(s, sizeof(char), len, file->fp) != (size_t) len) {
 			return 0;
@@ -61,6 +62,7 @@ int memfile_write(struct memfile *file, void *s, long long len) {
 	} else {
 		file->map.append(std::string((char *) s, len));
 		file->off += len;
+		in_memory = true;
 	}
 
 	return len;
diff --git a/memfile.hpp b/memfile.hpp
index e8ab8c6d..e339120d 100644
--- a/memfile.hpp
+++ b/memfile.hpp
@@ -15,7 +15,7 @@ struct memfile {
 
 struct memfile *memfile_open(int fd);
 int memfile_close(struct memfile *file);
-int memfile_write(struct memfile *file, void *s, long long len);
+int memfile_write(struct memfile *file, void *s, long long len, bool &in_memory);
 void memfile_full(struct memfile *file);
 
 #endif
diff --git a/pool.cpp b/pool.cpp
index bf3059c3..16fe4d49 100644
--- a/pool.cpp
+++ b/pool.cpp
@@ -64,14 +64,20 @@ long long addpool(struct memfile *poolfile, struct memfile *treefile, const char
 			// the pool is full yet.
 
 			long long off = poolfile->off;
-			if (memfile_write(poolfile, &type, 1) < 0) {
+			bool in_memory = false;
+
+			if (memfile_write(poolfile, &type, 1, in_memory) < 0) {
 				perror("memfile write");
 				exit(EXIT_WRITE);
 			}
-			if (memfile_write(poolfile, (void *) s, strlen(s) + 1) < 0) {
+			if (memfile_write(poolfile, (void *) s, strlen(s) + 1, in_memory) < 0) {
 				perror("memfile write");
 				exit(EXIT_WRITE);
 			}
+
+			if (in_memory) {
+				dedup[hash_off] = off;
+			}
 			return off;
 		}
 	}
@@ -94,11 +100,12 @@ long long addpool(struct memfile *poolfile, struct memfile *treefile, const char
 		// to the newly-added strings.
 
 		long long off = poolfile->off;
-		if (memfile_write(poolfile, &type, 1) < 0) {
+		bool in_memory;
+		if (memfile_write(poolfile, &type, 1, in_memory) < 0) {
 			perror("memfile write");
 			exit(EXIT_WRITE);
 		}
-		if (memfile_write(poolfile, (void *) s, strlen(s) + 1) < 0) {
+		if (memfile_write(poolfile, (void *) s, strlen(s) + 1, in_memory) < 0) {
 			perror("memfile write");
 			exit(EXIT_WRITE);
 		}
@@ -114,14 +121,18 @@ long long addpool(struct memfile *poolfile, struct memfile *treefile, const char
 	}
 
 	long long off = poolfile->off;
-	if (memfile_write(poolfile, &type, 1) < 0) {
+	bool in_memory = false;
+	if (memfile_write(poolfile, &type, 1, in_memory) < 0) {
 		perror("memfile write");
 		exit(EXIT_WRITE);
 	}
-	if (memfile_write(poolfile, (void *) s, strlen(s) + 1) < 0) {
+	if (memfile_write(poolfile, (void *) s, strlen(s) + 1, in_memory) < 0) {
 		perror("memfile write");
 		exit(EXIT_WRITE);
 	}
+	if (in_memory) {
+		dedup[hash_off] = off;
+	}
 
 	if (off >= LONG_MAX || treefile->off >= LONG_MAX) {
 		// Tree or pool is bigger than 2GB
@@ -139,7 +150,7 @@ long long addpool(struct memfile *poolfile, struct memfile *treefile, const char
 	tsp.off = off;
 
 	long long p = treefile->off;
-	if (memfile_write(treefile, &tsp, sizeof(struct stringpool)) < 0) {
+	if (memfile_write(treefile, &tsp, sizeof(struct stringpool), in_memory) < 0) {
 		perror("memfile write");
 		exit(EXIT_WRITE);
 	}