change: support version = 4 (noenc) format

laumann

Oct 6, 2023, 9:09 AM

ZS3UJEZU4PZEYOXQNZWZGDNXTP5ESNPRAEHPJQ24O3TWZH746IBQC

Dependencies

[2] XRGUHSC4 change: fix some formatting issues
[3] 7NQTS36D bincode: change signature of getstr to receive str destination
[4] QSQNGA5K changestore: load changes on demand
[5] JAGXXSR7 change: initial work on decoding hunks
[6] YFBKBUKB change: rename binstat -> bc
[7] VXQYIOBX change: avoid crash on different format versions
[8] 7WA2F6RS change: decode resurrectzombies, addroot, delroot
[9] 5D2IYPL7 change: introduce changestore, print deleted lines
[10] XTKRT6OQ format the codebase
[11] 2U7P5SFQ Change struct names "struct foo -> typedef struct Foo"
[12] VKLGQREY change: add base32 decode, initial deconstruction of hashed
[13] ZKAOPMCH change: Implement decoding of edits, atoms, positions, among others
[14] NEORNIZE change: decode and print file moves
[15] LPGHALLK change: adjust printing of file adds
[16] OBKF6SII change: decompress the hashed section too
[17] TGT4VSME change: decode FILE_ADD
[18] Q7TKZCJP Add initial support for reading the offsets from a (fixed) change
[19] 2YF4J5SJ change: decode and print file undel
[20] QYRJIOYP change: separate decoding and printing of hashed struct
[21] JVU3TTT5 all: switch away from typedefing anonymous structs
[22] NZNIG2UL Fix lots of warnings, adjust build flags
[23] YG4DZB3A add representation of hash, decode dependencies
[24] FMYCPGKD change: extract hash list decoding into reused function
[25] LCEKN25G change: parse authors
[26] O7XAUFM6 change: decode and print {,un}solve order conflicts
[*] RIWSVVAS change: decompress the 'contents' with zstd_seekable
[*] XTVLIC24 change: refactor to print_change()

Change contents

edit in change.c at line 289

[4.3757]

[4.3757]

static void
change_decode_encoding_noenc(struct bincode *bc, char **encoding)
{
(void)bc;
*encoding = xstrdup("UTF-8");
}
edit in change.c at line 304

[4.489]

[4.3878]

void (*fn_change_decode_encoding)(struct bincode *, char **);

fn_change_decode_encoding = &change_decode_encoding;
if (hashed->version == VERSION_NOENC)
fn_change_decode_encoding = change_decode_encoding_noenc;
replacement in change.c at line 324

[4.3181]→[4.3181:3235](∅→∅)

change_decode_encoding(bc, &bhunk->edit.encoding);

[4.3181]

[4.4305]

fn_change_decode_encoding(bc, &bhunk->edit.encoding);
replacement in change.c at line 330

[4.3406]→[4.3406:3467](∅→∅)

change_decode_encoding(bc, &bhunk->replacement.encoding);

[4.3406]

[4.0]

fn_change_decode_encoding(
bc, &bhunk->replacement.encoding
);
replacement in change.c at line 342

[3.238]→[4.3734:3791](∅→∅),[4.3734]→[4.3734:3791](∅→∅)

change_decode_encoding(bc, &bhunk->fileadd.encoding);

[3.238]

[4.583]

fn_change_decode_encoding(bc, &bhunk->fileadd.encoding);
replacement in change.c at line 359

[3.430]→[4.4185:4242](∅→∅),[4.4185]→[4.4185:4242](∅→∅)

change_decode_encoding(bc, &bhunk->filedel.encoding);

[3.430]

[4.58]

fn_change_decode_encoding(bc, &bhunk->filedel.encoding);
replacement in change.c at line 370

[3.529]→[4.4442:4501](∅→∅),[4.4442]→[4.4442:4501](∅→∅)

change_decode_encoding(bc, &bhunk->fileundel.encoding);

[3.529]

[4.258]

fn_change_decode_encoding(
bc, &bhunk->fileundel.encoding
);
replacement in change.c at line 407

[4.5040]→[4.490:517](∅→∅),[4.490]→[4.490:517](∅→∅)

change_decode_encoding(

[4.5040]

[4.5041]

fn_change_decode_encoding(
edit in change.c at line 423

[4.620]

[4.620]

}
}

void
change_decode_author(struct bincode *bc, struct author *author)
{
size_t len, i;

len = bincode_getu64(bc);

author->len = len;
author->entries = xmalloc(sizeof(struct authorentry) * len);

for (i = 0; i < author->len; i++) {
len = bincode_getu64(bc);
author->entries[i].key = xmalloc(len + 1);
bincode_getstr(bc, author->entries[i].key, len);

len = bincode_getu64(bc);
author->entries[i].value = xmalloc(len + 1);
bincode_getstr(bc, author->entries[i].value, len);
}
}

void
change_decode_author_noenc(struct bincode *bc, struct author *author)
{
size_t len, i;
char *name, *full_name, *email;

i = 1;
full_name = email = NULL;

len = bincode_getu64(bc);
name = xmalloc(len + 1);
bincode_getstr(bc, name, len);

if (bincode_getu8(bc)) {
i++;
len = bincode_getu64(bc);
full_name = xmalloc(len + 1);
bincode_getstr(bc, full_name, len);
}

if (bincode_getu8(bc)) {
i++;
len = bincode_getu64(bc);
email = xmalloc(len + 1);
bincode_getstr(bc, email, len);
}

author->len = i;
author->entries = xmalloc(sizeof(struct authorentry) * i);
i = 0;

author->entries[i].key = xstrdup("name");
author->entries[i].value = name;

if (full_name) {
i++;
author->entries[i].key = xstrdup("full_name");
author->entries[i].value = full_name;
edit in change.c at line 486

[4.623]

[4.623]

if (email) {
i++;
author->entries[i].key = xstrdup("email");
author->entries[i].value = email;
}
edit in change.c at line 508

[4.5304]

[28.2591]

void (*fn_change_decode_author)(struct bincode *, struct author *);
replacement in change.c at line 566

[4.514]→[4.427:479](∅→∅),[4.479]→[4.490:502](∅→∅),[4.502]→[4.5718:5747](∅→∅),[4.5747]→[4.480:522](∅→∅),[4.634]→[4.480:522](∅→∅),[4.522]→[4.3897:3938](∅→∅),[4.3169]→[4.523:626](∅→∅),[4.3938]→[4.523:626](∅→∅),[4.1331]→[4.523:626](∅→∅),[4.626]→[4.43:176](∅→∅)

for (i = 0; i < hashed->header.authors.len; i++) {
size_t j;
len = bincode_getu64(&bc);
hashed->header.authors.map[i].entries =
xmalloc(sizeof(struct author) * len);
hashed->header.authors.map[i].len = len;
for (j = 0; j < hashed->header.authors.map[i].len; j++) {
// In VERSION_NOENC, the authors are a struct { name: str, full_name: maybe str, email: maybe str } instead of a more generic map

[4.514]

[4.176]

fn_change_decode_author = change_decode_author;
if (hashed->version == VERSION_NOENC)
fn_change_decode_author = change_decode_author_noenc;
replacement in change.c at line 570

[4.177]→[4.177:265](∅→∅),[4.265]→[4.5748:5778](∅→∅),[4.5778]→[4.627:677](∅→∅),[4.849]→[4.627:677](∅→∅),[4.677]→[3.1075:1189](∅→∅),[3.1189]→[4.5809:5839](∅→∅),[4.5809]→[4.5809:5839](∅→∅),[4.5839]→[4.678:730](∅→∅),[4.964]→[4.678:730](∅→∅),[4.730]→[3.1190:1306](∅→∅),[3.1306]→[4.1172:1176](∅→∅),[4.1617]→[4.1172:1176](∅→∅),[4.5870]→[4.1172:1176](∅→∅),[4.1172]→[4.1172:1176](∅→∅)

// TODO switch on offsets.version to figure
// out how to decode the authors here
len = bincode_getu64(&bc);
hashed->header.authors.map[i].entries[j].key =
xmalloc(len + 1);
bincode_getstr(
&bc,
hashed->header.authors.map[i].entries[j].key,
len
);
len = bincode_getu64(&bc);
hashed->header.authors.map[i].entries[j].value =
xmalloc(len + 1);
bincode_getstr(
&bc,
hashed->header.authors.map[i].entries[j].value,
len
);
}

[4.177]

[4.838]

for (i = 0; i < hashed->header.authors.len; i++) {
fn_change_decode_author(&bc, &hashed->header.authors.map[i]);
edit in change.c at line 807

[4.765]

[4.765]
edit in change.c at line 815

[4.979]

[29.253]

/**
* In the VERSION_NOENC file metadata (for file additions), the
* filemetadata contains *only* the inodemetadata directly followed by
* the basename. The basename is not bincode-encoded, so we assume the
* length of the string is the given content length minus 2 (for the
* first two bytes used by the inodemetadata).
*/
static void
read_filemetadata_noenc(
struct filemetadata *m, uint8_t *contents, size_t contents_len
)
{
size_t len;
struct bincode bc = { .avail = contents_len, .buf = contents };

m->inodemetadata = bincode_getu16(&bc);
len = contents_len - 2;
m->basename = xmalloc(len + 1);
memcpy(m->basename, &contents[2], len);
m->basename[len] = '\0';
}
edit in change.c at line 844

[4.605]

[29.316]

void (*fn_read_filemetadata)(struct filemetadata *, uint8_t *, size_t);
edit in change.c at line 846

[29.317]

[29.317]

fn_read_filemetadata = read_filemetadata;
if (hashed->version == VERSION_NOENC)
fn_read_filemetadata = read_filemetadata_noenc;
replacement in change.c at line 917

[4.197]→[2.0:23](∅→∅)

read_filemetadata(

[4.197]

[2.23]

fn_read_filemetadata(
replacement in change.c at line 944

[4.1214]→[4.1214:1236](∅→∅)

read_filemetadata(

[4.1214]

[4.1236]

fn_read_filemetadata(
replacement in change.c at line 1074

[4.1874]→[4.7277:7348](∅→∅),[4.7348]→[4.335:421](∅→∅),[4.335]→[4.335:421](∅→∅),[4.421]→[4.7349:7374](∅→∅),[4.7374]→[4.445:512](∅→∅),[4.445]→[4.445:512](∅→∅),[4.512]→[4.7375:7400](∅→∅),[4.7400]→[4.536:540](∅→∅),[4.536]→[4.536:540](∅→∅)

if (off->version != VERSION) {
if (off->version == VERSION_NOENC) {
printf("The NOENC version (%lu) is not yet supported, sorry (but it's planned)\n",
off->version);
} else {
printf("Version %lu is not yet supported, sorry.\n",
off->version);
}

[4.1874]

[4.540]

if (off->version != VERSION && off->version != VERSION_NOENC) {
printf("Version %lu is not yet supported, sorry.\n",
off->version);
err = off->version;
replacement in change.c at line 1232

[4.8278]→[4.8278:8292](∅→∅)

size_t i, x;

[4.8278]

[4.8292]

size_t x;