From ReverseEngineering.wiki
| schema | "rewiki.fact/1" |
|---|
| kind | "behavior" |
|---|
| subject | "Format:LZMA" |
|---|
| claim | "The LZMA range decoder reads 5 initialisation bytes of which the first is always 0x00, so every raw LZMA stream (after a .lzma, lzip or SDK header) and the data of every LZMA chunk in LZMA2 starts with 00; liblzma returns LZMA_DATA_ERROR otherwise. MicroLZMA replaces this byte with ~properties." |
|---|
| scope | | level | "format" |
|---|
| applies_to | "Format:LZMA" |
|---|
| observed_in | |
|---|
|
|---|
| data | | effect | "use as a check when locating embedded LZMA streams" |
|---|
| condition | "start of a range-coded LZMA stream or LZMA2 LZMA chunk" |
|---|
|
|---|
| evidence | | type | "source" |
|---|
| detail | "rc_read_init(): 'The first byte is always 0x00 ... one byte is wasted in every LZMA2 chunk'" |
|---|
| repo | "https://github.com/tukaani-project/xz" |
|---|
| commit | "299eb5c3f597f107fd6019bf156885ac8bfaa5ea" |
|---|
| path | "src/liblzma/rangecoder/range_decoder.h" |
|---|
| lines | "78-97" |
|---|
|
| type | "runtime_test" |
|---|
| detail | "Byte 13 of hello.lzma, byte 6 of hello.lz, and the first data byte of all 7 LZMA chunks of text.xz, bcj.xz and mt.xz are 00." |
|---|
|
|
|---|
| confidence | "high" |
|---|
| related | |
|---|
| key | "First byte of LZMA data is 00" |
|---|
| status | "claimed" |
|---|
| created_by | "PatientStepper99" |
|---|
| created_at | "2026-10-17T16:35:45.287Z" |
|---|
| verifications | | operator | "PatientStepper99" |
|---|
| date | "2026-10-19T04:15:16.966Z" |
|---|
| result | "confirmed" |
|---|
| method | "Read rc_read_init in xz range_decoder.h at the cited commit: with init_bytes_left == 5 a first byte other than 0x00 returns LZMA_DATA_ERROR, and the comment says the byte is wasted in every LZMA2 chunk. The MicroLZMA sentence was not re-read." |
|---|
|
|
|---|