From 121035c83e18b70d3128e9ac966109ebedb7e516 Mon Sep 17 00:00:00 2001 From: Grzegorz Antoniak Date: Sat, 15 Dec 2018 08:38:26 +0100 Subject: [PATCH] ZIP reader: added support for XZ, LZMA, PPMD8 and BZIP2 decompression This commit adds some support for extraction of '.zipx' files. Those files are standard ZIP files that can contain files compressed with different set of algorithms that standard '.zip' files use. Support is still missing for Deflate64, JPEG and WavPack algorithms. --- libarchive/CMakeLists.txt | 2 + libarchive/archive_ppmd8.c | 1287 +++++++++++++++++ libarchive/archive_ppmd8_private.h | 148 ++ libarchive/archive_read_support_format_zip.c | 824 ++++++++++- libarchive/test/test_read_format_zip.c | 447 ++++++ .../test/test_read_format_zip_bzip2.zipx.uu | 19 + .../test_read_format_zip_bzip2_multi.zipx.uu | 96 ++ .../test/test_read_format_zip_lzma.zipx.uu | 19 + .../test_read_format_zip_lzma_multi.zipx.uu | 95 ++ .../test/test_read_format_zip_ppmd8.zipx.uu | 17 + .../test_read_format_zip_ppmd8_multi.zipx.uu | 84 ++ .../test_read_format_zip_xz_multi.zipx.uu | 125 ++ 12 files changed, 3123 insertions(+), 40 deletions(-) create mode 100644 libarchive/archive_ppmd8.c create mode 100644 libarchive/archive_ppmd8_private.h create mode 100644 libarchive/test/test_read_format_zip_bzip2.zipx.uu create mode 100644 libarchive/test/test_read_format_zip_bzip2_multi.zipx.uu create mode 100644 libarchive/test/test_read_format_zip_lzma.zipx.uu create mode 100644 libarchive/test/test_read_format_zip_lzma_multi.zipx.uu create mode 100644 libarchive/test/test_read_format_zip_ppmd8.zipx.uu create mode 100644 libarchive/test/test_read_format_zip_ppmd8_multi.zipx.uu create mode 100644 libarchive/test/test_read_format_zip_xz_multi.zipx.uu diff --git a/libarchive/CMakeLists.txt b/libarchive/CMakeLists.txt index 79719773c..8e86aade2 100644 --- a/libarchive/CMakeLists.txt +++ b/libarchive/CMakeLists.txt @@ -51,6 +51,8 @@ SET(libarchive_SOURCES archive_platform_acl.h archive_platform_xattr.h archive_ppmd_private.h + archive_ppmd8.c + archive_ppmd8_private.h archive_ppmd7.c archive_ppmd7_private.h archive_private.h diff --git a/libarchive/archive_ppmd8.c b/libarchive/archive_ppmd8.c new file mode 100644 index 000000000..d1779395d --- /dev/null +++ b/libarchive/archive_ppmd8.c @@ -0,0 +1,1287 @@ +/* Ppmd8.c -- PPMdI codec +2016-05-21 : Igor Pavlov : Public domain +This code is based on PPMd var.I (2002): Dmitry Shkarin : Public domain */ + +#include "archive_platform.h" + +#include + +#include "archive_ppmd8_private.h" + +const Byte PPMD8_kExpEscape[16] = { 25, 14, 9, 7, 5, 5, 4, 4, 4, 3, 3, 3, 2, 2, 2, 2 }; +static const UInt16 kInitBinEsc[] = { 0x3CDD, 0x1F3F, 0x59BF, 0x48F3, 0x64A1, 0x5ABC, 0x6632, 0x6051}; + +#define MAX_FREQ 124 +#define UNIT_SIZE 12 + +#define U2B(nu) ((UInt32)(nu) * UNIT_SIZE) +#define U2I(nu) (p->Units2Indx[(nu) - 1]) +#define I2U(indx) (p->Indx2Units[indx]) + +#ifdef PPMD_32BIT + #define REF(ptr) (ptr) +#else + #define REF(ptr) ((UInt32)((Byte *)(ptr) - (p)->Base)) +#endif + +#define STATS_REF(ptr) ((CPpmd_State_Ref)REF(ptr)) + +#define CTX(ref) ((CPpmd8_Context *)Ppmd8_GetContext(p, ref)) +#define STATS(ctx) Ppmd8_GetStats(p, ctx) +#define ONE_STATE(ctx) Ppmd8Context_OneState(ctx) +#define SUFFIX(ctx) CTX((ctx)->Suffix) + +#define kTop (1 << 24) +#define kBot (1 << 15) + +typedef CPpmd8_Context * CTX_PTR; + +struct CPpmd8_Node_; + +typedef + #ifdef PPMD_32BIT + struct CPpmd8_Node_ * + #else + UInt32 + #endif + CPpmd8_Node_Ref; + +typedef struct CPpmd8_Node_ +{ + UInt32 Stamp; + CPpmd8_Node_Ref Next; + UInt32 NU; +} CPpmd8_Node; + +#ifdef PPMD_32BIT + #define NODE(ptr) (ptr) +#else + #define NODE(offs) ((CPpmd8_Node *)(p->Base + (offs))) +#endif + +#define EMPTY_NODE 0xFFFFFFFF + +void Ppmd8_Construct(CPpmd8 *p) +{ + unsigned i, k, m; + + p->Base = 0; + + for (i = 0, k = 0; i < PPMD_NUM_INDEXES; i++) + { + unsigned step = (i >= 12 ? 4 : (i >> 2) + 1); + do { p->Units2Indx[k++] = (Byte)i; } while (--step); + p->Indx2Units[i] = (Byte)k; + } + + p->NS2BSIndx[0] = (0 << 1); + p->NS2BSIndx[1] = (1 << 1); + memset(p->NS2BSIndx + 2, (2 << 1), 9); + memset(p->NS2BSIndx + 11, (3 << 1), 256 - 11); + + for (i = 0; i < 5; i++) + p->NS2Indx[i] = (Byte)i; + for (m = i, k = 1; i < 260; i++) + { + p->NS2Indx[i] = (Byte)m; + if (--k == 0) + k = (++m) - 4; + } +} + +void Ppmd8_Free(CPpmd8 *p) +{ + free(p->Base); + p->Size = 0; + p->Base = 0; +} + +Bool Ppmd8_Alloc(CPpmd8 *p, UInt32 size) +{ + if (p->Base == 0 || p->Size != size) + { + Ppmd8_Free(p); + p->AlignOffset = + #ifdef PPMD_32BIT + (4 - size) & 3; + #else + 4 - (size & 3); + #endif + if ((p->Base = (Byte *)malloc(p->AlignOffset + size)) == 0) + return False; + p->Size = size; + } + return True; +} + +static void InsertNode(CPpmd8 *p, void *node, unsigned indx) +{ + ((CPpmd8_Node *)node)->Stamp = EMPTY_NODE; + ((CPpmd8_Node *)node)->Next = (CPpmd8_Node_Ref)p->FreeList[indx]; + ((CPpmd8_Node *)node)->NU = I2U(indx); + p->FreeList[indx] = REF(node); + p->Stamps[indx]++; +} + +static void *RemoveNode(CPpmd8 *p, unsigned indx) +{ + CPpmd8_Node *node = NODE((CPpmd8_Node_Ref)p->FreeList[indx]); + p->FreeList[indx] = node->Next; + p->Stamps[indx]--; + return node; +} + +static void SplitBlock(CPpmd8 *p, void *ptr, unsigned oldIndx, unsigned newIndx) +{ + unsigned i, nu = I2U(oldIndx) - I2U(newIndx); + ptr = (Byte *)ptr + U2B(I2U(newIndx)); + if (I2U(i = U2I(nu)) != nu) + { + unsigned k = I2U(--i); + InsertNode(p, ((Byte *)ptr) + U2B(k), nu - k - 1); + } + InsertNode(p, ptr, i); +} + +static void GlueFreeBlocks(CPpmd8 *p) +{ + CPpmd8_Node_Ref head = 0; + CPpmd8_Node_Ref *prev = &head; + unsigned i; + + p->GlueCount = 1 << 13; + memset(p->Stamps, 0, sizeof(p->Stamps)); + + /* Order-0 context is always at top UNIT, so we don't need guard NODE at the end. + All blocks up to p->LoUnit can be free, so we need guard NODE at LoUnit. */ + if (p->LoUnit != p->HiUnit) + ((CPpmd8_Node *)p->LoUnit)->Stamp = 0; + + /* Glue free blocks */ + for (i = 0; i < PPMD_NUM_INDEXES; i++) + { + CPpmd8_Node_Ref next = (CPpmd8_Node_Ref)p->FreeList[i]; + p->FreeList[i] = 0; + while (next != 0) + { + CPpmd8_Node *node = NODE(next); + if (node->NU != 0) + { + CPpmd8_Node *node2; + *prev = next; + prev = &(node->Next); + while ((node2 = node + node->NU)->Stamp == EMPTY_NODE) + { + node->NU += node2->NU; + node2->NU = 0; + } + } + next = node->Next; + } + } + *prev = 0; + + /* Fill lists of free blocks */ + while (head != 0) + { + CPpmd8_Node *node = NODE(head); + unsigned nu; + head = node->Next; + nu = node->NU; + if (nu == 0) + continue; + for (; nu > 128; nu -= 128, node += 128) + InsertNode(p, node, PPMD_NUM_INDEXES - 1); + if (I2U(i = U2I(nu)) != nu) + { + unsigned k = I2U(--i); + InsertNode(p, node + k, nu - k - 1); + } + InsertNode(p, node, i); + } +} + +static void *AllocUnitsRare(CPpmd8 *p, unsigned indx) +{ + unsigned i; + void *retVal; + if (p->GlueCount == 0) + { + GlueFreeBlocks(p); + if (p->FreeList[indx] != 0) + return RemoveNode(p, indx); + } + i = indx; + do + { + if (++i == PPMD_NUM_INDEXES) + { + UInt32 numBytes = U2B(I2U(indx)); + p->GlueCount--; + return ((UInt32)(p->UnitsStart - p->Text) > numBytes) ? (p->UnitsStart -= numBytes) : (NULL); + } + } + while (p->FreeList[i] == 0); + retVal = RemoveNode(p, i); + SplitBlock(p, retVal, i, indx); + return retVal; +} + +static void *AllocUnits(CPpmd8 *p, unsigned indx) +{ + UInt32 numBytes; + if (p->FreeList[indx] != 0) + return RemoveNode(p, indx); + numBytes = U2B(I2U(indx)); + if (numBytes <= (UInt32)(p->HiUnit - p->LoUnit)) + { + void *retVal = p->LoUnit; + p->LoUnit += numBytes; + return retVal; + } + return AllocUnitsRare(p, indx); +} + +#define MyMem12Cpy(dest, src, num) \ + { UInt32 *d = (UInt32 *)dest; const UInt32 *z = (const UInt32 *)src; UInt32 n = num; \ + do { d[0] = z[0]; d[1] = z[1]; d[2] = z[2]; z += 3; d += 3; } while (--n); } + +static void *ShrinkUnits(CPpmd8 *p, void *oldPtr, unsigned oldNU, unsigned newNU) +{ + unsigned i0 = U2I(oldNU); + unsigned i1 = U2I(newNU); + if (i0 == i1) + return oldPtr; + if (p->FreeList[i1] != 0) + { + void *ptr = RemoveNode(p, i1); + MyMem12Cpy(ptr, oldPtr, newNU); + InsertNode(p, oldPtr, i0); + return ptr; + } + SplitBlock(p, oldPtr, i0, i1); + return oldPtr; +} + +static void FreeUnits(CPpmd8 *p, void *ptr, unsigned nu) +{ + InsertNode(p, ptr, U2I(nu)); +} + +static void SpecialFreeUnit(CPpmd8 *p, void *ptr) +{ + if ((Byte *)ptr != p->UnitsStart) + InsertNode(p, ptr, 0); + else + { + #ifdef PPMD8_FREEZE_SUPPORT + *(UInt32 *)ptr = EMPTY_NODE; /* it's used for (Flags == 0xFF) check in RemoveBinContexts */ + #endif + p->UnitsStart += UNIT_SIZE; + } +} + +static void *MoveUnitsUp(CPpmd8 *p, void *oldPtr, unsigned nu) +{ + unsigned indx = U2I(nu); + void *ptr; + if ((Byte *)oldPtr > p->UnitsStart + 16 * 1024 || REF(oldPtr) > p->FreeList[indx]) + return oldPtr; + ptr = RemoveNode(p, indx); + MyMem12Cpy(ptr, oldPtr, nu); + if ((Byte*)oldPtr != p->UnitsStart) + InsertNode(p, oldPtr, indx); + else + p->UnitsStart += U2B(I2U(indx)); + return ptr; +} + +static void ExpandTextArea(CPpmd8 *p) +{ + UInt32 count[PPMD_NUM_INDEXES]; + unsigned i; + memset(count, 0, sizeof(count)); + if (p->LoUnit != p->HiUnit) + ((CPpmd8_Node *)p->LoUnit)->Stamp = 0; + + { + CPpmd8_Node *node = (CPpmd8_Node *)p->UnitsStart; + for (; node->Stamp == EMPTY_NODE; node += node->NU) + { + node->Stamp = 0; + count[U2I(node->NU)]++; + } + p->UnitsStart = (Byte *)node; + } + + for (i = 0; i < PPMD_NUM_INDEXES; i++) + { + CPpmd8_Node_Ref *next = (CPpmd8_Node_Ref *)&p->FreeList[i]; + while (count[i] != 0) + { + CPpmd8_Node *node = NODE(*next); + while (node->Stamp == 0) + { + *next = node->Next; + node = NODE(*next); + p->Stamps[i]--; + if (--count[i] == 0) + break; + } + next = &node->Next; + } + } +} + +#define SUCCESSOR(p) ((CPpmd_Void_Ref)((p)->SuccessorLow | ((UInt32)(p)->SuccessorHigh << 16))) + +static void SetSuccessor(CPpmd_State *p, CPpmd_Void_Ref v) +{ + (p)->SuccessorLow = (UInt16)((UInt32)(v) & 0xFFFF); + (p)->SuccessorHigh = (UInt16)(((UInt32)(v) >> 16) & 0xFFFF); +} + +#define RESET_TEXT(offs) { p->Text = p->Base + p->AlignOffset + (offs); } + +static void RestartModel(CPpmd8 *p) +{ + unsigned i, k, m, r; + + memset(p->FreeList, 0, sizeof(p->FreeList)); + memset(p->Stamps, 0, sizeof(p->Stamps)); + RESET_TEXT(0); + p->HiUnit = p->Text + p->Size; + p->LoUnit = p->UnitsStart = p->HiUnit - p->Size / 8 / UNIT_SIZE * 7 * UNIT_SIZE; + p->GlueCount = 0; + + p->OrderFall = p->MaxOrder; + p->RunLength = p->InitRL = -(Int32)((p->MaxOrder < 12) ? p->MaxOrder : 12) - 1; + p->PrevSuccess = 0; + + p->MinContext = p->MaxContext = (CTX_PTR)(p->HiUnit -= UNIT_SIZE); /* AllocContext(p); */ + p->MinContext->Suffix = 0; + p->MinContext->NumStats = 255; + p->MinContext->Flags = 0; + p->MinContext->SummFreq = 256 + 1; + p->FoundState = (CPpmd_State *)p->LoUnit; /* AllocUnits(p, PPMD_NUM_INDEXES - 1); */ + p->LoUnit += U2B(256 / 2); + p->MinContext->Stats = REF(p->FoundState); + for (i = 0; i < 256; i++) + { + CPpmd_State *s = &p->FoundState[i]; + s->Symbol = (Byte)i; + s->Freq = 1; + SetSuccessor(s, 0); + } + + for (i = m = 0; m < 25; m++) + { + while (p->NS2Indx[i] == m) + i++; + for (k = 0; k < 8; k++) + { + UInt16 val = (UInt16)(PPMD_BIN_SCALE - kInitBinEsc[k] / (i + 1)); + UInt16 *dest = p->BinSumm[m] + k; + for (r = 0; r < 64; r += 8) + dest[r] = val; + } + } + + for (i = m = 0; m < 24; m++) + { + while (p->NS2Indx[i + 3] == m + 3) + i++; + for (k = 0; k < 32; k++) + { + CPpmd_See *s = &p->See[m][k]; + s->Summ = (UInt16)((2 * i + 5) << (s->Shift = PPMD_PERIOD_BITS - 4)); + s->Count = 7; + } + } +} + +void Ppmd8_Init(CPpmd8 *p, unsigned maxOrder, unsigned restoreMethod) +{ + p->MaxOrder = maxOrder; + p->RestoreMethod = restoreMethod; + RestartModel(p); + p->DummySee.Shift = PPMD_PERIOD_BITS; + p->DummySee.Summ = 0; /* unused */ + p->DummySee.Count = 64; /* unused */ +} + +static void Refresh(CPpmd8 *p, CTX_PTR ctx, unsigned oldNU, unsigned scale) +{ + unsigned i = ctx->NumStats, escFreq, sumFreq, flags; + CPpmd_State *s = (CPpmd_State *)ShrinkUnits(p, STATS(ctx), oldNU, (i + 2) >> 1); + ctx->Stats = REF(s); + #ifdef PPMD8_FREEZE_SUPPORT + /* fixed over Shkarin's code. Fixed code is not compatible with original code for some files in FREEZE mode. */ + scale |= (ctx->SummFreq >= ((UInt32)1 << 15)); + #endif + flags = (ctx->Flags & (0x10 + 0x04 * scale)) + 0x08 * (s->Symbol >= 0x40); + escFreq = ctx->SummFreq - s->Freq; + sumFreq = (s->Freq = (Byte)((s->Freq + scale) >> scale)); + do + { + escFreq -= (++s)->Freq; + sumFreq += (s->Freq = (Byte)((s->Freq + scale) >> scale)); + flags |= 0x08 * (s->Symbol >= 0x40); + } + while (--i); + ctx->SummFreq = (UInt16)(sumFreq + ((escFreq + scale) >> scale)); + ctx->Flags = (Byte)flags; +} + +static void SwapStates(CPpmd_State *t1, CPpmd_State *t2) +{ + CPpmd_State tmp = *t1; + *t1 = *t2; + *t2 = tmp; +} + +static CPpmd_Void_Ref CutOff(CPpmd8 *p, CTX_PTR ctx, unsigned order) +{ + int i; + unsigned tmp; + CPpmd_State *s; + + if (!ctx->NumStats) + { + s = ONE_STATE(ctx); + if ((Byte *)Ppmd8_GetPtr(p, SUCCESSOR(s)) >= p->UnitsStart) + { + if (order < p->MaxOrder) + SetSuccessor(s, CutOff(p, CTX(SUCCESSOR(s)), order + 1)); + else + SetSuccessor(s, 0); + if (SUCCESSOR(s) || order <= 9) /* O_BOUND */ + return REF(ctx); + } + SpecialFreeUnit(p, ctx); + return 0; + } + + ctx->Stats = STATS_REF(MoveUnitsUp(p, STATS(ctx), tmp = ((unsigned)ctx->NumStats + 2) >> 1)); + + for (s = STATS(ctx) + (i = ctx->NumStats); s >= STATS(ctx); s--) + if ((Byte *)Ppmd8_GetPtr(p, SUCCESSOR(s)) < p->UnitsStart) + { + CPpmd_State *s2 = STATS(ctx) + (i--); + SetSuccessor(s, 0); + SwapStates(s, s2); + } + else if (order < p->MaxOrder) + SetSuccessor(s, CutOff(p, CTX(SUCCESSOR(s)), order + 1)); + else + SetSuccessor(s, 0); + + if (i != ctx->NumStats && order) + { + ctx->NumStats = (Byte)i; + s = STATS(ctx); + if (i < 0) + { + FreeUnits(p, s, tmp); + SpecialFreeUnit(p, ctx); + return 0; + } + if (i == 0) + { + ctx->Flags = (Byte)((ctx->Flags & 0x10) + 0x08 * (s->Symbol >= 0x40)); + *ONE_STATE(ctx) = *s; + FreeUnits(p, s, tmp); + /* 9.31: the code was fixed. It's was not BUG, if Freq <= MAX_FREQ = 124 */ + ONE_STATE(ctx)->Freq = (Byte)(((unsigned)ONE_STATE(ctx)->Freq + 11) >> 3); + } + else + Refresh(p, ctx, tmp, ctx->SummFreq > 16 * i); + } + return REF(ctx); +} + +#ifdef PPMD8_FREEZE_SUPPORT +static CPpmd_Void_Ref RemoveBinContexts(CPpmd8 *p, CTX_PTR ctx, unsigned order) +{ + CPpmd_State *s; + if (!ctx->NumStats) + { + s = ONE_STATE(ctx); + if ((Byte *)Ppmd8_GetPtr(p, SUCCESSOR(s)) >= p->UnitsStart && order < p->MaxOrder) + SetSuccessor(s, RemoveBinContexts(p, CTX(SUCCESSOR(s)), order + 1)); + else + SetSuccessor(s, 0); + /* Suffix context can be removed already, since different (high-order) + Successors may refer to same context. So we check Flags == 0xFF (Stamp == EMPTY_NODE) */ + if (!SUCCESSOR(s) && (!SUFFIX(ctx)->NumStats || SUFFIX(ctx)->Flags == 0xFF)) + { + FreeUnits(p, ctx, 1); + return 0; + } + else + return REF(ctx); + } + + for (s = STATS(ctx) + ctx->NumStats; s >= STATS(ctx); s--) + if ((Byte *)Ppmd8_GetPtr(p, SUCCESSOR(s)) >= p->UnitsStart && order < p->MaxOrder) + SetSuccessor(s, RemoveBinContexts(p, CTX(SUCCESSOR(s)), order + 1)); + else + SetSuccessor(s, 0); + + return REF(ctx); +} +#endif + +static UInt32 GetUsedMemory(const CPpmd8 *p) +{ + UInt32 v = 0; + unsigned i; + for (i = 0; i < PPMD_NUM_INDEXES; i++) + v += p->Stamps[i] * I2U(i); + return p->Size - (UInt32)(p->HiUnit - p->LoUnit) - (UInt32)(p->UnitsStart - p->Text) - U2B(v); +} + +#ifdef PPMD8_FREEZE_SUPPORT + #define RESTORE_MODEL(c1, fSuccessor) RestoreModel(p, c1, fSuccessor) +#else + #define RESTORE_MODEL(c1, fSuccessor) RestoreModel(p, c1) +#endif + +static void RestoreModel(CPpmd8 *p, CTX_PTR c1 + #ifdef PPMD8_FREEZE_SUPPORT + , CTX_PTR fSuccessor + #endif + ) +{ + CTX_PTR c; + CPpmd_State *s; + RESET_TEXT(0); + for (c = p->MaxContext; c != c1; c = SUFFIX(c)) + if (--(c->NumStats) == 0) + { + s = STATS(c); + c->Flags = (Byte)((c->Flags & 0x10) + 0x08 * (s->Symbol >= 0x40)); + *ONE_STATE(c) = *s; + SpecialFreeUnit(p, s); + ONE_STATE(c)->Freq = (Byte)(((unsigned)ONE_STATE(c)->Freq + 11) >> 3); + } + else + Refresh(p, c, (c->NumStats+3) >> 1, 0); + + for (; c != p->MinContext; c = SUFFIX(c)) + if (!c->NumStats) + ONE_STATE(c)->Freq = (Byte)(ONE_STATE(c)->Freq - (ONE_STATE(c)->Freq >> 1)); + else if ((c->SummFreq += 4) > 128 + 4 * c->NumStats) + Refresh(p, c, (c->NumStats + 2) >> 1, 1); + + #ifdef PPMD8_FREEZE_SUPPORT + if (p->RestoreMethod > PPMD8_RESTORE_METHOD_FREEZE) + { + p->MaxContext = fSuccessor; + p->GlueCount += !(p->Stamps[1] & 1); + } + else if (p->RestoreMethod == PPMD8_RESTORE_METHOD_FREEZE) + { + while (p->MaxContext->Suffix) + p->MaxContext = SUFFIX(p->MaxContext); + RemoveBinContexts(p, p->MaxContext, 0); + p->RestoreMethod++; + p->GlueCount = 0; + p->OrderFall = p->MaxOrder; + } + else + #endif + if (p->RestoreMethod == PPMD8_RESTORE_METHOD_RESTART || GetUsedMemory(p) < (p->Size >> 1)) + RestartModel(p); + else + { + while (p->MaxContext->Suffix) + p->MaxContext = SUFFIX(p->MaxContext); + do + { + CutOff(p, p->MaxContext, 0); + ExpandTextArea(p); + } + while (GetUsedMemory(p) > 3 * (p->Size >> 2)); + p->GlueCount = 0; + p->OrderFall = p->MaxOrder; + } +} + +static CTX_PTR CreateSuccessors(CPpmd8 *p, Bool skip, CPpmd_State *s1, CTX_PTR c) +{ + CPpmd_State upState; + Byte flags; + CPpmd_Byte_Ref upBranch = (CPpmd_Byte_Ref)SUCCESSOR(p->FoundState); + /* fixed over Shkarin's code. Maybe it could work without + 1 too. */ + CPpmd_State *ps[PPMD8_MAX_ORDER + 1]; + unsigned numPs = 0; + + if (!skip) + ps[numPs++] = p->FoundState; + + while (c->Suffix) + { + CPpmd_Void_Ref successor; + CPpmd_State *s; + c = SUFFIX(c); + if (s1) + { + s = s1; + s1 = NULL; + } + else if (c->NumStats != 0) + { + for (s = STATS(c); s->Symbol != p->FoundState->Symbol; s++); + if (s->Freq < MAX_FREQ - 9) + { + s->Freq++; + c->SummFreq++; + } + } + else + { + s = ONE_STATE(c); + s->Freq = (Byte)(s->Freq + (!SUFFIX(c)->NumStats & (s->Freq < 24))); + } + successor = SUCCESSOR(s); + if (successor != upBranch) + { + c = CTX(successor); + if (numPs == 0) + return c; + break; + } + ps[numPs++] = s; + } + + upState.Symbol = *(const Byte *)Ppmd8_GetPtr(p, upBranch); + SetSuccessor(&upState, upBranch + 1); + flags = (Byte)(0x10 * (p->FoundState->Symbol >= 0x40) + 0x08 * (upState.Symbol >= 0x40)); + + if (c->NumStats == 0) + upState.Freq = ONE_STATE(c)->Freq; + else + { + UInt32 cf, s0; + CPpmd_State *s; + for (s = STATS(c); s->Symbol != upState.Symbol; s++); + cf = s->Freq - 1; + s0 = c->SummFreq - c->NumStats - cf; + upState.Freq = (Byte)(1 + ((2 * cf <= s0) ? (5 * cf > s0) : ((cf + 2 * s0 - 3) / s0))); + } + + do + { + /* Create Child */ + CTX_PTR c1; /* = AllocContext(p); */ + if (p->HiUnit != p->LoUnit) + c1 = (CTX_PTR)(p->HiUnit -= UNIT_SIZE); + else if (p->FreeList[0] != 0) + c1 = (CTX_PTR)RemoveNode(p, 0); + else + { + c1 = (CTX_PTR)AllocUnitsRare(p, 0); + if (!c1) + return NULL; + } + c1->NumStats = 0; + c1->Flags = flags; + *ONE_STATE(c1) = upState; + c1->Suffix = REF(c); + SetSuccessor(ps[--numPs], REF(c1)); + c = c1; + } + while (numPs != 0); + + return c; +} + +static CTX_PTR ReduceOrder(CPpmd8 *p, CPpmd_State *s1, CTX_PTR c) +{ + CPpmd_State *s = NULL; + CTX_PTR c1 = c; + CPpmd_Void_Ref upBranch = REF(p->Text); + + #ifdef PPMD8_FREEZE_SUPPORT + /* The BUG in Shkarin's code was fixed: ps could overflow in CUT_OFF mode. */ + CPpmd_State *ps[PPMD8_MAX_ORDER + 1]; + unsigned numPs = 0; + ps[numPs++] = p->FoundState; + #endif + + SetSuccessor(p->FoundState, upBranch); + p->OrderFall++; + + for (;;) + { + if (s1) + { + c = SUFFIX(c); + s = s1; + s1 = NULL; + } + else + { + if (!c->Suffix) + { + #ifdef PPMD8_FREEZE_SUPPORT + if (p->RestoreMethod > PPMD8_RESTORE_METHOD_FREEZE) + { + do { SetSuccessor(ps[--numPs], REF(c)); } while (numPs); + RESET_TEXT(1); + p->OrderFall = 1; + } + #endif + return c; + } + c = SUFFIX(c); + if (c->NumStats) + { + if ((s = STATS(c))->Symbol != p->FoundState->Symbol) + do { s++; } while (s->Symbol != p->FoundState->Symbol); + if (s->Freq < MAX_FREQ - 9) + { + s->Freq += 2; + c->SummFreq += 2; + } + } + else + { + s = ONE_STATE(c); + s->Freq = (Byte)(s->Freq + (s->Freq < 32)); + } + } + if (SUCCESSOR(s)) + break; + #ifdef PPMD8_FREEZE_SUPPORT + ps[numPs++] = s; + #endif + SetSuccessor(s, upBranch); + p->OrderFall++; + } + + #ifdef PPMD8_FREEZE_SUPPORT + if (p->RestoreMethod > PPMD8_RESTORE_METHOD_FREEZE) + { + c = CTX(SUCCESSOR(s)); + do { SetSuccessor(ps[--numPs], REF(c)); } while (numPs); + RESET_TEXT(1); + p->OrderFall = 1; + return c; + } + else + #endif + if (SUCCESSOR(s) <= upBranch) + { + CTX_PTR successor; + CPpmd_State *s2 = p->FoundState; + p->FoundState = s; + + successor = CreateSuccessors(p, False, NULL, c); + if (successor == NULL) + SetSuccessor(s, 0); + else + SetSuccessor(s, REF(successor)); + p->FoundState = s2; + } + + if (p->OrderFall == 1 && c1 == p->MaxContext) + { + SetSuccessor(p->FoundState, SUCCESSOR(s)); + p->Text--; + } + if (SUCCESSOR(s) == 0) + return NULL; + return CTX(SUCCESSOR(s)); +} + +static void UpdateModel(CPpmd8 *p) +{ + CPpmd_Void_Ref successor, fSuccessor = SUCCESSOR(p->FoundState); + CTX_PTR c; + unsigned s0, ns, fFreq = p->FoundState->Freq; + Byte flag, fSymbol = p->FoundState->Symbol; + CPpmd_State *s = NULL; + + if (p->FoundState->Freq < MAX_FREQ / 4 && p->MinContext->Suffix != 0) + { + c = SUFFIX(p->MinContext); + + if (c->NumStats == 0) + { + s = ONE_STATE(c); + if (s->Freq < 32) + s->Freq++; + } + else + { + s = STATS(c); + if (s->Symbol != p->FoundState->Symbol) + { + do { s++; } while (s->Symbol != p->FoundState->Symbol); + if (s[0].Freq >= s[-1].Freq) + { + SwapStates(&s[0], &s[-1]); + s--; + } + } + if (s->Freq < MAX_FREQ - 9) + { + s->Freq += 2; + c->SummFreq += 2; + } + } + } + + c = p->MaxContext; + if (p->OrderFall == 0 && fSuccessor) + { + CTX_PTR cs = CreateSuccessors(p, True, s, p->MinContext); + if (cs == 0) + { + SetSuccessor(p->FoundState, 0); + RESTORE_MODEL(c, CTX(fSuccessor)); + } + else + { + SetSuccessor(p->FoundState, REF(cs)); + p->MaxContext = cs; + } + return; + } + + *p->Text++ = p->FoundState->Symbol; + successor = REF(p->Text); + if (p->Text >= p->UnitsStart) + { + RESTORE_MODEL(c, CTX(fSuccessor)); /* check it */ + return; + } + + if (!fSuccessor) + { + CTX_PTR cs = ReduceOrder(p, s, p->MinContext); + if (cs == NULL) + { + RESTORE_MODEL(c, 0); + return; + } + fSuccessor = REF(cs); + } + else if ((Byte *)Ppmd8_GetPtr(p, fSuccessor) < p->UnitsStart) + { + CTX_PTR cs = CreateSuccessors(p, False, s, p->MinContext); + if (cs == NULL) + { + RESTORE_MODEL(c, 0); + return; + } + fSuccessor = REF(cs); + } + + if (--p->OrderFall == 0) + { + successor = fSuccessor; + p->Text -= (p->MaxContext != p->MinContext); + } + #ifdef PPMD8_FREEZE_SUPPORT + else if (p->RestoreMethod > PPMD8_RESTORE_METHOD_FREEZE) + { + successor = fSuccessor; + RESET_TEXT(0); + p->OrderFall = 0; + } + #endif + + s0 = p->MinContext->SummFreq - (ns = p->MinContext->NumStats) - fFreq; + flag = (Byte)(0x08 * (fSymbol >= 0x40)); + + for (; c != p->MinContext; c = SUFFIX(c)) + { + unsigned ns1; + UInt32 cf, sf; + if ((ns1 = c->NumStats) != 0) + { + if ((ns1 & 1) != 0) + { + /* Expand for one UNIT */ + unsigned oldNU = (ns1 + 1) >> 1; + unsigned i = U2I(oldNU); + if (i != U2I(oldNU + 1)) + { + void *ptr = AllocUnits(p, i + 1); + void *oldPtr; + if (!ptr) + { + RESTORE_MODEL(c, CTX(fSuccessor)); + return; + } + oldPtr = STATS(c); + MyMem12Cpy(ptr, oldPtr, oldNU); + InsertNode(p, oldPtr, i); + c->Stats = STATS_REF(ptr); + } + } + c->SummFreq = (UInt16)(c->SummFreq + (3 * ns1 + 1 < ns)); + } + else + { + CPpmd_State *s2 = (CPpmd_State*)AllocUnits(p, 0); + if (!s2) + { + RESTORE_MODEL(c, CTX(fSuccessor)); + return; + } + *s2 = *ONE_STATE(c); + c->Stats = REF(s2); + if (s2->Freq < MAX_FREQ / 4 - 1) + s2->Freq <<= 1; + else + s2->Freq = MAX_FREQ - 4; + c->SummFreq = (UInt16)(s2->Freq + p->InitEsc + (ns > 2)); + } + cf = 2 * fFreq * (c->SummFreq + 6); + sf = (UInt32)s0 + c->SummFreq; + if (cf < 6 * sf) + { + cf = 1 + (cf > sf) + (cf >= 4 * sf); + c->SummFreq += 4; + } + else + { + cf = 4 + (cf > 9 * sf) + (cf > 12 * sf) + (cf > 15 * sf); + c->SummFreq = (UInt16)(c->SummFreq + cf); + } + { + CPpmd_State *s2 = STATS(c) + ns1 + 1; + SetSuccessor(s2, successor); + s2->Symbol = fSymbol; + s2->Freq = (Byte)cf; + c->Flags |= flag; + c->NumStats = (Byte)(ns1 + 1); + } + } + p->MaxContext = p->MinContext = CTX(fSuccessor); +} + +static void Rescale(CPpmd8 *p) +{ + unsigned i, adder, sumFreq, escFreq; + CPpmd_State *stats = STATS(p->MinContext); + CPpmd_State *s = p->FoundState; + { + CPpmd_State tmp = *s; + for (; s != stats; s--) + s[0] = s[-1]; + *s = tmp; + } + escFreq = p->MinContext->SummFreq - s->Freq; + s->Freq += 4; + adder = (p->OrderFall != 0 + #ifdef PPMD8_FREEZE_SUPPORT + || p->RestoreMethod > PPMD8_RESTORE_METHOD_FREEZE + #endif + ); + s->Freq = (Byte)((s->Freq + adder) >> 1); + sumFreq = s->Freq; + + i = p->MinContext->NumStats; + do + { + escFreq -= (++s)->Freq; + s->Freq = (Byte)((s->Freq + adder) >> 1); + sumFreq += s->Freq; + if (s[0].Freq > s[-1].Freq) + { + CPpmd_State *s1 = s; + CPpmd_State tmp = *s1; + do + s1[0] = s1[-1]; + while (--s1 != stats && tmp.Freq > s1[-1].Freq); + *s1 = tmp; + } + } + while (--i); + + if (s->Freq == 0) + { + unsigned numStats = p->MinContext->NumStats; + unsigned n0, n1; + do { i++; } while ((--s)->Freq == 0); + escFreq += i; + p->MinContext->NumStats = (Byte)(p->MinContext->NumStats - i); + if (p->MinContext->NumStats == 0) + { + CPpmd_State tmp = *stats; + tmp.Freq = (Byte)((2 * tmp.Freq + escFreq - 1) / escFreq); + if (tmp.Freq > MAX_FREQ / 3) + tmp.Freq = MAX_FREQ / 3; + InsertNode(p, stats, U2I((numStats + 2) >> 1)); + p->MinContext->Flags = (Byte)((p->MinContext->Flags & 0x10) + 0x08 * (tmp.Symbol >= 0x40)); + *(p->FoundState = ONE_STATE(p->MinContext)) = tmp; + return; + } + n0 = (numStats + 2) >> 1; + n1 = (p->MinContext->NumStats + 2) >> 1; + if (n0 != n1) + p->MinContext->Stats = STATS_REF(ShrinkUnits(p, stats, n0, n1)); + p->MinContext->Flags &= ~0x08; + p->MinContext->Flags |= 0x08 * ((s = STATS(p->MinContext))->Symbol >= 0x40); + i = p->MinContext->NumStats; + do { p->MinContext->Flags |= 0x08*((++s)->Symbol >= 0x40); } while (--i); + } + p->MinContext->SummFreq = (UInt16)(sumFreq + escFreq - (escFreq >> 1)); + p->MinContext->Flags |= 0x4; + p->FoundState = STATS(p->MinContext); +} + +CPpmd_See *Ppmd8_MakeEscFreq(CPpmd8 *p, unsigned numMasked1, UInt32 *escFreq) +{ + CPpmd_See *see; + if (p->MinContext->NumStats != 0xFF) + { + see = p->See[(unsigned)p->NS2Indx[(unsigned)p->MinContext->NumStats + 2] - 3] + + (p->MinContext->SummFreq > 11 * ((unsigned)p->MinContext->NumStats + 1)) + + 2 * (unsigned)(2 * (unsigned)p->MinContext->NumStats < + ((unsigned)SUFFIX(p->MinContext)->NumStats + numMasked1)) + + p->MinContext->Flags; + { + unsigned r = (see->Summ >> see->Shift); + see->Summ = (UInt16)(see->Summ - r); + *escFreq = r + (r == 0); + } + } + else + { + see = &p->DummySee; + *escFreq = 1; + } + return see; +} + +static void NextContext(CPpmd8 *p) +{ + CTX_PTR c = CTX(SUCCESSOR(p->FoundState)); + if (p->OrderFall == 0 && (Byte *)c >= p->UnitsStart) + p->MinContext = p->MaxContext = c; + else + { + UpdateModel(p); + p->MinContext = p->MaxContext; + } +} + +void Ppmd8_Update1(CPpmd8 *p) +{ + CPpmd_State *s = p->FoundState; + s->Freq += 4; + p->MinContext->SummFreq += 4; + if (s[0].Freq > s[-1].Freq) + { + SwapStates(&s[0], &s[-1]); + p->FoundState = --s; + if (s->Freq > MAX_FREQ) + Rescale(p); + } + NextContext(p); +} + +void Ppmd8_Update1_0(CPpmd8 *p) +{ + p->PrevSuccess = (2 * p->FoundState->Freq >= p->MinContext->SummFreq); + p->RunLength += p->PrevSuccess; + p->MinContext->SummFreq += 4; + if ((p->FoundState->Freq += 4) > MAX_FREQ) + Rescale(p); + NextContext(p); +} + +void Ppmd8_UpdateBin(CPpmd8 *p) +{ + p->FoundState->Freq = (Byte)(p->FoundState->Freq + (p->FoundState->Freq < 196)); + p->PrevSuccess = 1; + p->RunLength++; + NextContext(p); +} + +void Ppmd8_Update2(CPpmd8 *p) +{ + p->MinContext->SummFreq += 4; + if ((p->FoundState->Freq += 4) > MAX_FREQ) + Rescale(p); + p->RunLength = p->InitRL; + UpdateModel(p); + p->MinContext = p->MaxContext; +} + +/* Ppmd8Dec.c -- PPMdI Decoder +2010-04-16 : Igor Pavlov : Public domain +This code is based on: + PPMd var.I (2002): Dmitry Shkarin : Public domain + Carryless rangecoder (1999): Dmitry Subbotin : Public domain */ + +Bool Ppmd8_RangeDec_Init(CPpmd8 *p) +{ + unsigned i; + p->Low = 0; + p->Range = 0xFFFFFFFF; + p->Code = 0; + for (i = 0; i < 4; i++) + p->Code = (p->Code << 8) | p->Stream.In->Read(p->Stream.In); + return (p->Code < 0xFFFFFFFF); +} + +static UInt32 RangeDec_GetThreshold(CPpmd8 *p, UInt32 total) +{ + return p->Code / (p->Range /= total); +} + +static void RangeDec_Decode(CPpmd8 *p, UInt32 start, UInt32 size) +{ + start *= p->Range; + p->Low += start; + p->Code -= start; + p->Range *= size; + + while ((p->Low ^ (p->Low + p->Range)) < kTop || + (p->Range < kBot && ((p->Range = (0 - p->Low) & (kBot - 1)), 1))) + { + p->Code = (p->Code << 8) | p->Stream.In->Read(p->Stream.In); + p->Range <<= 8; + p->Low <<= 8; + } +} + +#define MASK(sym) ((signed char *)charMask)[sym] + +int Ppmd8_DecodeSymbol(CPpmd8 *p) +{ + size_t charMask[256 / sizeof(size_t)]; + if (p->MinContext->NumStats != 0) + { + CPpmd_State *s = Ppmd8_GetStats(p, p->MinContext); + unsigned i; + UInt32 count, hiCnt; + if ((count = RangeDec_GetThreshold(p, p->MinContext->SummFreq)) < (hiCnt = s->Freq)) + { + Byte symbol; + RangeDec_Decode(p, 0, s->Freq); + p->FoundState = s; + symbol = s->Symbol; + Ppmd8_Update1_0(p); + return symbol; + } + p->PrevSuccess = 0; + i = p->MinContext->NumStats; + do + { + if ((hiCnt += (++s)->Freq) > count) + { + Byte symbol; + RangeDec_Decode(p, hiCnt - s->Freq, s->Freq); + p->FoundState = s; + symbol = s->Symbol; + Ppmd8_Update1(p); + return symbol; + } + } + while (--i); + if (count >= p->MinContext->SummFreq) + return -2; + RangeDec_Decode(p, hiCnt, p->MinContext->SummFreq - hiCnt); + PPMD_SetAllBitsIn256Bytes(charMask); + MASK(s->Symbol) = 0; + i = p->MinContext->NumStats; + do { MASK((--s)->Symbol) = 0; } while (--i); + } + else + { + UInt16 *prob = Ppmd8_GetBinSumm(p); + if (((p->Code / (p->Range >>= 14)) < *prob)) + { + Byte symbol; + RangeDec_Decode(p, 0, *prob); + *prob = (UInt16)PPMD_UPDATE_PROB_0(*prob); + symbol = (p->FoundState = Ppmd8Context_OneState(p->MinContext))->Symbol; + Ppmd8_UpdateBin(p); + return symbol; + } + RangeDec_Decode(p, *prob, (1 << 14) - *prob); + *prob = (UInt16)PPMD_UPDATE_PROB_1(*prob); + p->InitEsc = PPMD8_kExpEscape[*prob >> 10]; + PPMD_SetAllBitsIn256Bytes(charMask); + MASK(Ppmd8Context_OneState(p->MinContext)->Symbol) = 0; + p->PrevSuccess = 0; + } + for (;;) + { + CPpmd_State *ps[256], *s; + UInt32 freqSum, count, hiCnt; + CPpmd_See *see; + unsigned i, num, numMasked = p->MinContext->NumStats; + do + { + p->OrderFall++; + if (!p->MinContext->Suffix) + return -1; + p->MinContext = Ppmd8_GetContext(p, p->MinContext->Suffix); + } + while (p->MinContext->NumStats == numMasked); + hiCnt = 0; + s = Ppmd8_GetStats(p, p->MinContext); + i = 0; + num = p->MinContext->NumStats - numMasked; + do + { + int k = (int)(MASK(s->Symbol)); + hiCnt += (s->Freq & k); + ps[i] = s++; + i -= k; + } + while (i != num); + + see = Ppmd8_MakeEscFreq(p, numMasked, &freqSum); + freqSum += hiCnt; + count = RangeDec_GetThreshold(p, freqSum); + + if (count < hiCnt) + { + Byte symbol; + CPpmd_State **pps = ps; + for (hiCnt = 0; (hiCnt += (*pps)->Freq) <= count; pps++); + s = *pps; + RangeDec_Decode(p, hiCnt - s->Freq, s->Freq); + Ppmd_See_Update(see); + p->FoundState = s; + symbol = s->Symbol; + Ppmd8_Update2(p); + return symbol; + } + if (count >= freqSum) + return -2; + RangeDec_Decode(p, hiCnt, freqSum - hiCnt); + see->Summ = (UInt16)(see->Summ + freqSum); + do { MASK(ps[--i]->Symbol) = 0; } while (i != 0); + } +} + +/* H->I changes: + NS2Indx + GlewCount, and Glue method + BinSum + See / EscFreq + CreateSuccessors updates more suffix contexts + UpdateModel consts. + PrevSuccess Update +*/ + +const IPpmd8 __archive_ppmd8_functions = +{ + &Ppmd8_Construct, + &Ppmd8_Alloc, + &Ppmd8_Free, + &Ppmd8_Init, + &Ppmd8_RangeDec_Init, + &Ppmd8_DecodeSymbol, +}; diff --git a/libarchive/archive_ppmd8_private.h b/libarchive/archive_ppmd8_private.h new file mode 100644 index 000000000..534927860 --- /dev/null +++ b/libarchive/archive_ppmd8_private.h @@ -0,0 +1,148 @@ +/* Ppmd8.h -- PPMdI codec +2011-01-27 : Igor Pavlov : Public domain +This code is based on: + PPMd var.I (2002): Dmitry Shkarin : Public domain + Carryless rangecoder (1999): Dmitry Subbotin : Public domain */ + +#ifndef __PPMD8_H +#define __PPMD8_H + +#include "archive_ppmd_private.h" + +#define PPMD8_MIN_ORDER 2 +#define PPMD8_MAX_ORDER 16 + +struct CPpmd8_Context_; + +typedef + #ifdef PPMD_32BIT + struct CPpmd8_Context_ * + #else + UInt32 + #endif + CPpmd8_Context_Ref; + +#pragma pack(push, 1) + +typedef struct CPpmd8_Context_ +{ + Byte NumStats; + Byte Flags; + UInt16 SummFreq; + CPpmd_State_Ref Stats; + CPpmd8_Context_Ref Suffix; +} CPpmd8_Context; + +#pragma pack(pop) + +#define Ppmd8Context_OneState(p) ((CPpmd_State *)&(p)->SummFreq) + +/* The BUG in Shkarin's code for FREEZE mode was fixed, but that fixed + code is not compatible with original code for some files compressed + in FREEZE mode. So we disable FREEZE mode support. */ + +enum +{ + PPMD8_RESTORE_METHOD_RESTART, + PPMD8_RESTORE_METHOD_CUT_OFF + #ifdef PPMD8_FREEZE_SUPPORT + , PPMD8_RESTORE_METHOD_FREEZE + #endif +}; + +typedef struct +{ + CPpmd8_Context *MinContext, *MaxContext; + CPpmd_State *FoundState; + unsigned OrderFall, InitEsc, PrevSuccess, MaxOrder; + Int32 RunLength, InitRL; /* must be 32-bit at least */ + + UInt32 Size; + UInt32 GlueCount; + Byte *Base, *LoUnit, *HiUnit, *Text, *UnitsStart; + UInt32 AlignOffset; + unsigned RestoreMethod; + + /* Range Coder */ + UInt32 Range; + UInt32 Code; + UInt32 Low; + union + { + IByteIn *In; + IByteOut *Out; + } Stream; + + Byte Indx2Units[PPMD_NUM_INDEXES]; + Byte Units2Indx[128]; + CPpmd_Void_Ref FreeList[PPMD_NUM_INDEXES]; + UInt32 Stamps[PPMD_NUM_INDEXES]; + + Byte NS2BSIndx[256], NS2Indx[260]; + CPpmd_See DummySee, See[24][32]; + UInt16 BinSumm[25][64]; +} CPpmd8; + +void Ppmd8_Construct(CPpmd8 *p); +Bool Ppmd8_Alloc(CPpmd8 *p, UInt32 size); +void Ppmd8_Free(CPpmd8 *p); +void Ppmd8_Init(CPpmd8 *p, unsigned maxOrder, unsigned restoreMethod); +#define Ppmd8_WasAllocated(p) ((p)->Base != NULL) + + +/* ---------- Internal Functions ---------- */ + +extern const Byte PPMD8_kExpEscape[16]; + +#ifdef PPMD_32BIT + #define Ppmd8_GetPtr(p, ptr) (ptr) + #define Ppmd8_GetContext(p, ptr) (ptr) + #define Ppmd8_GetStats(p, ctx) ((ctx)->Stats) +#else + #define Ppmd8_GetPtr(p, offs) ((void *)((p)->Base + (offs))) + #define Ppmd8_GetContext(p, offs) ((CPpmd8_Context *)Ppmd8_GetPtr((p), (offs))) + #define Ppmd8_GetStats(p, ctx) ((CPpmd_State *)Ppmd8_GetPtr((p), ((ctx)->Stats))) +#endif + +void Ppmd8_Update1(CPpmd8 *p); +void Ppmd8_Update1_0(CPpmd8 *p); +void Ppmd8_Update2(CPpmd8 *p); +void Ppmd8_UpdateBin(CPpmd8 *p); + +#define Ppmd8_GetBinSumm(p) \ + &p->BinSumm[p->NS2Indx[Ppmd8Context_OneState(p->MinContext)->Freq - 1]][ \ + p->NS2BSIndx[Ppmd8_GetContext(p, p->MinContext->Suffix)->NumStats] + \ + p->PrevSuccess + p->MinContext->Flags + ((p->RunLength >> 26) & 0x20)] + +CPpmd_See *Ppmd8_MakeEscFreq(CPpmd8 *p, unsigned numMasked, UInt32 *scale); + + +/* ---------- Decode ---------- */ + +Bool Ppmd8_RangeDec_Init(CPpmd8 *p); +#define Ppmd8_RangeDec_IsFinishedOK(p) ((p)->Code == 0) +int Ppmd8_DecodeSymbol(CPpmd8 *p); /* returns: -1 as EndMarker, -2 as DataError */ + +/* ---------- Encode ---------- */ + +#define Ppmd8_RangeEnc_Init(p) { (p)->Low = 0; (p)->Range = 0xFFFFFFFF; } +void Ppmd8_RangeEnc_FlushData(CPpmd8 *p); +void Ppmd8_EncodeSymbol(CPpmd8 *p, int symbol); /* symbol = -1 means EndMarker */ + +typedef struct +{ + /* Base Functions */ + void (*Ppmd8_Construct)(CPpmd8 *p); + Bool (*Ppmd8_Alloc)(CPpmd8 *p, UInt32 size); + void (*Ppmd8_Free)(CPpmd8 *p); + void (*Ppmd8_Init)(CPpmd8 *p, unsigned max_order, unsigned restore_method); + #define Ppmd7_WasAllocated(p) ((p)->Base != NULL) + + /* Decode Functions */ + int (*Ppmd8_RangeDec_Init)(CPpmd8 *p); + int (*Ppmd8_DecodeSymbol)(CPpmd8 *p); +} IPpmd8; + +extern const IPpmd8 __archive_ppmd8_functions; + +#endif diff --git a/libarchive/archive_read_support_format_zip.c b/libarchive/archive_read_support_format_zip.c index 420004dba..6cd5570c6 100644 --- a/libarchive/archive_read_support_format_zip.c +++ b/libarchive/archive_read_support_format_zip.c @@ -52,6 +52,12 @@ __FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_format_zip.c 201102 #ifdef HAVE_ZLIB_H #include #endif +#ifdef HAVE_BZLIB_H +#include +#endif +#ifdef HAVE_LZMA_H +#include +#endif #include "archive.h" #include "archive_digest_private.h" @@ -63,6 +69,7 @@ __FBSDID("$FreeBSD: head/lib/libarchive/archive_read_support_format_zip.c 201102 #include "archive_private.h" #include "archive_rb.h" #include "archive_read_private.h" +#include "archive_ppmd8_private.h" #ifndef HAVE_ZLIB_H #include "archive_crc32.h" @@ -165,13 +172,29 @@ struct zip { char decompress_init; char end_of_entry; -#ifdef HAVE_ZLIB_H unsigned char *uncompressed_buffer; size_t uncompressed_buffer_size; + +#ifdef HAVE_ZLIB_H z_stream stream; char stream_valid; #endif +#ifdef HAVE_LZMA_H + lzma_stream zipx_lzma_stream; + char zipx_lzma_valid; +#endif + +#ifdef HAVE_BZLIB_H + bz_stream bzstream; + char bzstream_valid; +#endif + + IByteIn zipx_ppmd_stream; + ssize_t zipx_ppmd_read_compressed; + CPpmd8 ppmd8; + char ppmd8_valid; + struct archive_string_conv *sconv; struct archive_string_conv *sconv_default; struct archive_string_conv *sconv_utf8; @@ -222,6 +245,27 @@ struct zip { /* Many systems define min or MIN, but not all. */ #define zipmin(a,b) ((a) < (b) ? (a) : (b)) +/* This function is used by Ppmd8_DecodeSymbol during decompression of Ppmd8 + * streams inside ZIP files. It has 2 purposes: one is to fetch the next + * compressed byte from the stream, second one is to increase the counter how + * many compressed bytes were read. */ +static Byte +ppmd_read(void* p) { + /* Get the handle to current decompression context. */ + struct archive_read *a = ((IByteIn*)p)->a; + struct zip *zip = (struct zip*) a->format->data; + + /* Fetch next byte. */ + const uint8_t* data = __archive_read_ahead(a, 1, NULL); + __archive_read_consume(a, 1); + + /* Increment the counter. */ + ++zip->zipx_ppmd_read_compressed; + + /* Return the next compressed byte. */ + return data[0]; +} + /* ------------------------------------------------------------------------ */ /* @@ -372,6 +416,8 @@ static const struct { {17, "reserved"}, /* Reserved by PKWARE */ {18, "ibm-terse-new"}, /* File is compressed using IBM TERSE (new) */ {19, "ibm-lz777"},/* IBM LZ77 z Architecture (PFS) */ + {95, "xz"}, /* XZ compressed data */ + {96, "jpeg"}, /* JPEG compressed data */ {97, "wav-pack"}, /* WavPack compressed data */ {98, "ppmd-1"}, /* PPMd version I, Rev 1 */ {99, "aes"} /* WinZip AES encryption */ @@ -1296,6 +1342,698 @@ zip_read_data_none(struct archive_read *a, const void **_buff, return (ARCHIVE_OK); } +static int +consume_optional_marker(struct archive_read *a, struct zip *zip) +{ + if (zip->end_of_entry && (zip->entry->zip_flags & ZIP_LENGTH_AT_END)) { + const char *p; + + if (NULL == (p = __archive_read_ahead(a, 24, NULL))) { + archive_set_error(&a->archive, + ARCHIVE_ERRNO_FILE_FORMAT, + "Truncated ZIP end-of-file record"); + return (ARCHIVE_FATAL); + } + /* Consume the optional PK\007\010 marker. */ + if (p[0] == 'P' && p[1] == 'K' && + p[2] == '\007' && p[3] == '\010') { + p += 4; + zip->unconsumed = 4; + } + if (zip->entry->flags & LA_USED_ZIP64) { + uint64_t compressed, uncompressed; + zip->entry->crc32 = archive_le32dec(p); + compressed = archive_le64dec(p + 4); + uncompressed = archive_le64dec(p + 12); + if (compressed > INT64_MAX || uncompressed > INT64_MAX) { + archive_set_error(&a->archive, + ARCHIVE_ERRNO_FILE_FORMAT, + "Overflow of 64-bit file sizes"); + return ARCHIVE_FAILED; + } + zip->entry->compressed_size = compressed; + zip->entry->uncompressed_size = uncompressed; + zip->unconsumed += 20; + } else { + zip->entry->crc32 = archive_le32dec(p); + zip->entry->compressed_size = archive_le32dec(p + 4); + zip->entry->uncompressed_size = archive_le32dec(p + 8); + zip->unconsumed += 12; + } + } + + return (ARCHIVE_OK); +} + +#ifdef HAVE_LZMA_H +static int +zipx_xz_init(struct archive_read *a, struct zip *zip) +{ + lzma_ret r; + + if(zip->zipx_lzma_valid) { + lzma_end(&zip->zipx_lzma_stream); + zip->zipx_lzma_valid = 0; + } + + memset(&zip->zipx_lzma_stream, 0, sizeof(zip->zipx_lzma_stream)); + r = lzma_stream_decoder(&zip->zipx_lzma_stream, UINT64_MAX, 0); + if (r != LZMA_OK) { + archive_set_error(&(a->archive), ARCHIVE_ERRNO_MISC, + "xz initialization failed(%d)", + r); + + return (ARCHIVE_FAILED); + } + + zip->zipx_lzma_valid = 1; + + if(zip->uncompressed_buffer) + free(zip->uncompressed_buffer); + + zip->uncompressed_buffer_size = 256 * 1024; + zip->uncompressed_buffer = + (uint8_t*) malloc(zip->uncompressed_buffer_size); + if (zip->uncompressed_buffer == NULL) { + archive_set_error(&a->archive, ENOMEM, + "No memory for xz decompression"); + return (ARCHIVE_FATAL); + } + + zip->decompress_init = 1; + return (ARCHIVE_OK); +} + +static int +zipx_lzma_alone_init(struct archive_read *a, struct zip *zip) +{ + lzma_ret r; + const uint8_t* p; + +#pragma pack(push) +#pragma pack(1) + struct _alone_header { + uint8_t bytes[5]; + uint64_t uncompressed_size; + } alone_header; +#pragma pack(pop) + + /* To unpack ZIPX's "LZMA" (id 14) stream we can use standard liblzma that + * is a part of XZ Utils. The stream format stored inside ZIPX file is a + * modified "lzma alone" file format, that was used by the `lzma` utility + * which was later deprecated in favour of `xz` utility. Since those + * formats are nearly the same, we can use a standard "lzma alone" decoder + * from XZ Utils. */ + + memset(&zip->zipx_lzma_stream, 0, sizeof(zip->zipx_lzma_stream)); + r = lzma_alone_decoder(&zip->zipx_lzma_stream, UINT64_MAX); + if (r != LZMA_OK) { + archive_set_error(&(a->archive), ARCHIVE_ERRNO_MISC, + "lzma initialization failed(%d)", r); + + return (ARCHIVE_FAILED); + } + + /* Flag the cleanup function that we want our lzma-related structures + * to be freed later. */ + zip->zipx_lzma_valid = 1; + + /* The "lzma alone" file format and the stream format inside ZIPx are + * almost the same. Here's an example of a structure of "lzma alone" + * format: + * + * $ cat /bin/ls | lzma | xxd | head -n 1 + * 00000000: 5d00 0080 00ff ffff ffff ffff ff00 2814 + * + * 5 bytes 8 bytes n bytes + * + * + * lzma_params is a 5-byte blob that has to be decoded to extract + * parameters of this LZMA stream. The uncompressed_size field is an + * uint64_t value that contains information about the size of the + * uncompressed file, or UINT64_MAX if this value is unknown. The + * part is the actual lzma-compressed data stream. + * + * Now here's the structure of the stream inside the ZIPX file: + * + * $ cat stream_inside_zipx | xxd | head -n 1 + * 00000000: 0914 0500 5d00 8000 0000 2814 .... .... + * + * 2byte 2byte 5 bytes n bytes + * + * + * This means that the ZIPX file contains an additional magic1 and magic2 + * headers, the lzma_params field contains the same parameter set as in the + * "lzma alone" format, and the field is the same as in the "lzma + * alone" format as well. Note that also the zipx format is missing the + * uncompressed_size field. + * + * So, in order to use the "lzma alone" decoder for the zipx lzma stream, + * we simply need to shuffle around some fields, prepare a new lzma alone + * header, feed it into lzma alone decoder so it will initialize itself + * properly, and then we can start feeding normal zipx lzma stream into the + * decoder. + */ + + /* Read magic1,magic2,lzma_params from the ZIPX stream. */ + if((p = __archive_read_ahead(a, 9, NULL)) == NULL) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, + "Truncated lzma data"); + return (ARCHIVE_FATAL); + } + + if(p[2] != 0x05 || p[3] != 0x00) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, + "Invalid lzma data"); + return (ARCHIVE_FATAL); + } + + /* Prepare an lzma alone header: copy the lzma_params blob into a proper + * place into the lzma alone header. */ + memcpy(&alone_header.bytes[0], p + 4, 5); + + /* Initialize the 'uncompressed size' field to unknown; we'll manually + * monitor how many bytes there are still to be uncompressed. */ + alone_header.uncompressed_size = UINT64_MAX; + + if(!zip->uncompressed_buffer) { + zip->uncompressed_buffer_size = 256 * 1024; + zip->uncompressed_buffer = + (uint8_t*) malloc(zip->uncompressed_buffer_size); + + if (zip->uncompressed_buffer == NULL) { + archive_set_error(&a->archive, ENOMEM, + "No memory for lzma decompression"); + return (ARCHIVE_FATAL); + } + } + + zip->zipx_lzma_stream.next_in = (void*) &alone_header; + zip->zipx_lzma_stream.avail_in = sizeof(alone_header); + zip->zipx_lzma_stream.total_in = 0; + zip->zipx_lzma_stream.next_out = zip->uncompressed_buffer; + zip->zipx_lzma_stream.avail_out = zip->uncompressed_buffer_size; + zip->zipx_lzma_stream.total_out = 0; + + /* Feed only the header into the lzma alone decoder. This will effectively + * initialize the decoder, and will not produce any output bytes yet. */ + r = lzma_code(&zip->zipx_lzma_stream, LZMA_RUN); + if (r != LZMA_OK) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_PROGRAMMER, + "lzma stream initialization error"); + return ARCHIVE_FATAL; + } + + /* We've already consumed some bytes, so take this into account. */ + __archive_read_consume(a, 9); + zip->entry_bytes_remaining -= 9; + zip->entry_compressed_bytes_read += 9; + + zip->decompress_init = 1; + return (ARCHIVE_OK); +} + +static int +zip_read_data_zipx_xz(struct archive_read *a, const void **buff, + size_t *size, int64_t *offset) +{ + struct zip* zip = (struct zip *)(a->format->data); + int ret; + lzma_ret lz_ret; + const void* compressed_buf; + ssize_t bytes_avail, in_bytes, to_consume = 0; + + (void) offset; /* UNUSED */ + + /* Initialize decompressor if not yet initialized. */ + if (!zip->decompress_init) { + ret = zipx_xz_init(a, zip); + if (ret != ARCHIVE_OK) + return (ret); + } + + compressed_buf = __archive_read_ahead(a, 1, &bytes_avail); + if (bytes_avail < 0) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, + "Truncated xz file body"); + return (ARCHIVE_FATAL); + } + + in_bytes = zipmin(zip->entry_bytes_remaining, bytes_avail); + zip->zipx_lzma_stream.next_in = compressed_buf; + zip->zipx_lzma_stream.avail_in = in_bytes; + zip->zipx_lzma_stream.total_in = 0; + zip->zipx_lzma_stream.next_out = zip->uncompressed_buffer; + zip->zipx_lzma_stream.avail_out = zip->uncompressed_buffer_size; + zip->zipx_lzma_stream.total_out = 0; + + /* Perform the decompression. */ + lz_ret = lzma_code(&zip->zipx_lzma_stream, LZMA_RUN); + switch(lz_ret) { + case LZMA_DATA_ERROR: + archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, + "xz data error (error %d)", (int) lz_ret); + return (ARCHIVE_FATAL); + + case LZMA_NO_CHECK: + case LZMA_OK: + break; + + default: + archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, + "xz unknown error %d", (int) lz_ret); + return (ARCHIVE_FATAL); + + case LZMA_STREAM_END: + lzma_end(&zip->zipx_lzma_stream); + zip->zipx_lzma_valid = 0; + + if((int64_t) zip->zipx_lzma_stream.total_in != + zip->entry_bytes_remaining) + { + archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, + "xz premature end of stream"); + return (ARCHIVE_FATAL); + } + + zip->end_of_entry = 1; + break; + } + + to_consume = zip->zipx_lzma_stream.total_in; + + __archive_read_consume(a, to_consume); + zip->entry_bytes_remaining -= to_consume; + zip->entry_compressed_bytes_read += to_consume; + zip->entry_uncompressed_bytes_read += zip->zipx_lzma_stream.total_out; + + *size = zip->zipx_lzma_stream.total_out; + *buff = zip->uncompressed_buffer; + + ret = consume_optional_marker(a, zip); + if (ret != ARCHIVE_OK) + return (ret); + + return (ARCHIVE_OK); +} + +static int +zip_read_data_zipx_lzma_alone(struct archive_read *a, const void **buff, + size_t *size, int64_t *offset) +{ + struct zip* zip = (struct zip *)(a->format->data); + int ret; + lzma_ret lz_ret; + const void* compressed_buf; + ssize_t bytes_avail, in_bytes, to_consume; + + (void) offset; /* UNUSED */ + + /* Initialize decompressor if not yet initialized. */ + if (!zip->decompress_init) { + ret = zipx_lzma_alone_init(a, zip); + if (ret != ARCHIVE_OK) + return (ret); + } + + /* Fetch more compressed data. The same note as in deflate handler applies + * here as well: + * + * Note: '1' here is a performance optimization. Recall that the + * decompression layer returns a count of available bytes; asking for more + * than that forces the decompressor to combine reads by copying data. + */ + compressed_buf = __archive_read_ahead(a, 1, &bytes_avail); + if (bytes_avail < 0) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, + "Truncated lzma file body"); + return (ARCHIVE_FATAL); + } + + /* Set decompressor parameters. */ + in_bytes = zipmin(zip->entry_bytes_remaining, bytes_avail); + + zip->zipx_lzma_stream.next_in = compressed_buf; + zip->zipx_lzma_stream.avail_in = in_bytes; + zip->zipx_lzma_stream.total_in = 0; + zip->zipx_lzma_stream.next_out = zip->uncompressed_buffer; + zip->zipx_lzma_stream.avail_out = + /* These lzma_alone streams lack end of stream marker, so let's make + * sure the unpacker won't try to unpack more than it's supposed to. */ + zipmin((int64_t) zip->uncompressed_buffer_size, + zip->entry->uncompressed_size - + zip->entry_uncompressed_bytes_read); + zip->zipx_lzma_stream.total_out = 0; + + /* Perform the decompression. */ + lz_ret = lzma_code(&zip->zipx_lzma_stream, LZMA_RUN); + switch(lz_ret) { + case LZMA_DATA_ERROR: + archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, + "lzma data error (error %d)", (int) lz_ret); + return (ARCHIVE_FATAL); + + case LZMA_OK: + break; + + default: + archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, + "lzma unknown error %d", (int) lz_ret); + return (ARCHIVE_FATAL); + } + + to_consume = zip->zipx_lzma_stream.total_in; + + /* Update pointers. */ + __archive_read_consume(a, to_consume); + zip->entry_bytes_remaining -= to_consume; + zip->entry_compressed_bytes_read += to_consume; + zip->entry_uncompressed_bytes_read += zip->zipx_lzma_stream.total_out; + + if(zip->entry_bytes_remaining == 0) { + zip->end_of_entry = 1; + } + + /* Return values. */ + *size = zip->zipx_lzma_stream.total_out; + *buff = zip->uncompressed_buffer; + + /* Behave the same way as during deflate decompression. */ + ret = consume_optional_marker(a, zip); + if (ret != ARCHIVE_OK) + return (ret); + + /* Free lzma decoder handle because we'll no longer need it. */ + if(zip->end_of_entry) { + lzma_end(&zip->zipx_lzma_stream); + zip->zipx_lzma_valid = 0; + } + + /* If we're here, then we're good! */ + return (ARCHIVE_OK); +} +#endif /* HAVE_LZMA_H */ + +static int +zipx_ppmd8_init(struct archive_read *a, struct zip *zip) +{ + const void* p; + uint32_t val; + uint32_t order; + uint32_t mem; + uint32_t restore_method; + + /* Remove previous decompression context if it exists. */ + if(zip->ppmd8_valid) { + __archive_ppmd8_functions.Ppmd8_Free(&zip->ppmd8); + zip->ppmd8_valid = 0; + } + + /* Create a new decompression context. */ + __archive_ppmd8_functions.Ppmd8_Construct(&zip->ppmd8); + + /* Setup function pointers required by Ppmd8 decompressor. The + * 'ppmd_read' function will feed new bytes to the decompressor, + * and will increment the 'zip->zipx_ppmd_read_compressed' counter. */ + zip->ppmd8.Stream.In = &zip->zipx_ppmd_stream; + zip->zipx_ppmd_stream.a = a; + zip->zipx_ppmd_stream.Read = &ppmd_read; + + /* Reset number of read bytes to 0. */ + zip->zipx_ppmd_read_compressed = 0; + + /* Read Ppmd8 header (2 bytes). */ + p = __archive_read_ahead(a, 2, NULL); + if(!p) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, + "Truncated file data in PPMd8 stream"); + return (ARCHIVE_FATAL); + } + __archive_read_consume(a, 2); + + /* Decode the stream's compression parameters. */ + val = archive_le16dec(p); + order = (val & 15) + 1; + mem = ((val >> 4) & 0xff) + 1; + restore_method = (val >> 12); + + if(order < 2 || restore_method > 2) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, + "Invalid parameter set in PPMd8 stream (order=%d, " + "restore=%d)", order, restore_method); + return (ARCHIVE_FAILED); + } + + /* Allocate the memory needed to properly decompress the file. */ + if(!__archive_ppmd8_functions.Ppmd8_Alloc(&zip->ppmd8, mem << 20)) { + archive_set_error(&a->archive, ENOMEM, + "Unable to allocate memory for PPMd8 stream: %d bytes", + mem << 20); + return (ARCHIVE_FATAL); + } + + /* Signal the cleanup function to release Ppmd8 context in the + * cleanup phase. */ + zip->ppmd8_valid = 1; + + /* Perform further Ppmd8 initialization. */ + if(!__archive_ppmd8_functions.Ppmd8_RangeDec_Init(&zip->ppmd8)) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_PROGRAMMER, + "PPMd8 stream range decoder initialization error"); + return (ARCHIVE_FATAL); + } + + __archive_ppmd8_functions.Ppmd8_Init(&zip->ppmd8, order, restore_method); + + /* Allocate the buffer that will hold uncompressed data. */ + if(zip->uncompressed_buffer) + free(zip->uncompressed_buffer); + + zip->uncompressed_buffer_size = 256 * 1024; + zip->uncompressed_buffer = + (uint8_t*) malloc(zip->uncompressed_buffer_size); + + if(zip->uncompressed_buffer == NULL) { + archive_set_error(&a->archive, ENOMEM, + "No memory for PPMd8 decompression"); + return ARCHIVE_FATAL; + } + + /* Ppmd8 initialization is done. */ + zip->decompress_init = 1; + + /* We've already read 2 bytes in the output stream. Additionally, + * Ppmd8 initialization code could read some data as well. So we + * are advancing the stream by 2 bytes plus whatever number of + * bytes Ppmd8 init function used. */ + zip->entry_compressed_bytes_read += 2 + zip->zipx_ppmd_read_compressed; + + return ARCHIVE_OK; +} + +static int +zip_read_data_zipx_ppmd(struct archive_read *a, const void **buff, + size_t *size, int64_t *offset) +{ + struct zip* zip = (struct zip *)(a->format->data); + int ret; + size_t consumed_bytes = 0; + ssize_t bytes_avail = 0; + + (void) offset; /* UNUSED */ + + /* If we're here for the first time, initialize Ppmd8 decompression + * context first. */ + if(!zip->decompress_init) { + ret = zipx_ppmd8_init(a, zip); + if(ret != ARCHIVE_OK) + return ret; + } + + /* Fetch for more data. We're reading 1 byte here, but libarchive should + * prefetch more bytes. */ + (void) __archive_read_ahead(a, 1, &bytes_avail); + if(bytes_avail < 0) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, + "Truncated PPMd8 file body"); + return (ARCHIVE_FATAL); + } + + /* This counter will be updated inside ppmd_read(), which at one + * point will be called by Ppmd8_DecodeSymbol. */ + zip->zipx_ppmd_read_compressed = 0; + + /* Decompression loop. */ + do { + int sym = __archive_ppmd8_functions.Ppmd8_DecodeSymbol(&zip->ppmd8); + if(sym < 0) { + zip->end_of_entry = 1; + break; + } + + zip->uncompressed_buffer[consumed_bytes] = (uint8_t) sym; + ++consumed_bytes; + } while(consumed_bytes < zip->uncompressed_buffer_size); + + /* Update pointers for libarchive. */ + *buff = zip->uncompressed_buffer; + *size = consumed_bytes; + + /* Update pointers so we can continue decompression in another call. */ + zip->entry_bytes_remaining -= zip->zipx_ppmd_read_compressed; + zip->entry_compressed_bytes_read += zip->zipx_ppmd_read_compressed; + zip->entry_uncompressed_bytes_read += consumed_bytes; + + /* If we're at the end of stream, deinitialize Ppmd8 context. */ + if(zip->end_of_entry) { + __archive_ppmd8_functions.Ppmd8_Free(&zip->ppmd8); + zip->ppmd8_valid = 0; + } + + /* Seek for optional marker, same way as in each zip entry. */ + ret = consume_optional_marker(a, zip); + if (ret != ARCHIVE_OK) + return ret; + + return ARCHIVE_OK; +} + +#ifdef HAVE_BZLIB_H +static int +zipx_bzip2_init(struct archive_read *a, struct zip *zip) +{ + int r; + + /* Deallocate already existing BZ2 decompression context if it + * exists. */ + if(zip->bzstream_valid) { + BZ2_bzDecompressEnd(&zip->bzstream); + zip->bzstream_valid = 0; + } + + /* Allocate a new BZ2 decompression context. */ + memset(&zip->bzstream, 0, sizeof(bz_stream)); + r = BZ2_bzDecompressInit(&zip->bzstream, 0, 1); + if(r != BZ_OK) { + archive_set_error(&(a->archive), ARCHIVE_ERRNO_MISC, + "bzip2 initialization failed(%d)", + r); + + return ARCHIVE_FAILED; + } + + /* Mark the bzstream field to be released in cleanup phase. */ + zip->bzstream_valid = 1; + + /* (Re)allocate the buffer that will contain decompressed bytes. */ + if(zip->uncompressed_buffer) + free(zip->uncompressed_buffer); + + zip->uncompressed_buffer_size = 256 * 1024; + zip->uncompressed_buffer = + (uint8_t*) malloc(zip->uncompressed_buffer_size); + if (zip->uncompressed_buffer == NULL) { + archive_set_error(&a->archive, ENOMEM, + "No memory for bzip2 decompression"); + return ARCHIVE_FATAL; + } + + /* Initialization done. */ + zip->decompress_init = 1; + return ARCHIVE_OK; +} + +static int +zip_read_data_zipx_bzip2(struct archive_read *a, const void **buff, + size_t *size, int64_t *offset) +{ + struct zip *zip = (struct zip *)(a->format->data); + ssize_t bytes_avail = 0, in_bytes, to_consume; + const void *compressed_buff; + int r; + uint64_t total_out; + + (void) offset; /* UNUSED */ + + /* Initialize decompression context if we're here for the first time. */ + if(!zip->decompress_init) { + r = zipx_bzip2_init(a, zip); + if(r != ARCHIVE_OK) + return r; + } + + /* Fetch more compressed bytes. */ + compressed_buff = __archive_read_ahead(a, 1, &bytes_avail); + if(bytes_avail < 0) { + archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, + "Truncated bzip2 file body"); + return (ARCHIVE_FATAL); + } + + in_bytes = zipmin(zip->entry_bytes_remaining, bytes_avail); + + /* Setup buffer boundaries. */ + zip->bzstream.next_in = (char*)(uintptr_t) compressed_buff; + zip->bzstream.avail_in = in_bytes; + zip->bzstream.total_in_hi32 = 0; + zip->bzstream.total_in_lo32 = 0; + zip->bzstream.next_out = (char*) zip->uncompressed_buffer; + zip->bzstream.avail_out = zip->uncompressed_buffer_size; + zip->bzstream.total_out_hi32 = 0; + zip->bzstream.total_out_lo32 = 0; + + /* Perform the decompression. */ + r = BZ2_bzDecompress(&zip->bzstream); + switch(r) { + case BZ_STREAM_END: + /* If we're at the end of the stream, deinitialize the + * decompression context now. */ + switch(BZ2_bzDecompressEnd(&zip->bzstream)) { + case BZ_OK: + break; + default: + archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, + "Failed to clean up bzip2 decompressor"); + return ARCHIVE_FATAL; + } + + zip->end_of_entry = 1; + break; + case BZ_OK: + /* The decompressor has successfully decoded this chunk of + * data, but more data is still in queue. */ + break; + default: + archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC, + "bzip2 decompression failed"); + return ARCHIVE_FATAL; + } + + /* Update the pointers so decompressor can continue decoding. */ + to_consume = zip->bzstream.total_in_lo32; + __archive_read_consume(a, to_consume); + + total_out = ((uint64_t) zip->bzstream.total_out_hi32 << 32) + + zip->bzstream.total_out_lo32; + + zip->entry_bytes_remaining -= to_consume; + zip->entry_compressed_bytes_read += to_consume; + zip->entry_uncompressed_bytes_read += total_out; + + /* Give libarchive its due. */ + *size = total_out; + *buff = zip->uncompressed_buffer; + + /* Seek for optional marker, like in other entries. */ + r = consume_optional_marker(a, zip); + if(r != ARCHIVE_OK) + return r; + + return ARCHIVE_OK; +} + +#endif + #ifdef HAVE_ZLIB_H static int zip_deflate_init(struct archive_read *a, struct zip *zip) @@ -1470,42 +2208,9 @@ zip_read_data_deflate(struct archive_read *a, const void **buff, return (r); } - if (zip->end_of_entry && (zip->entry->zip_flags & ZIP_LENGTH_AT_END)) { - const char *p; - - if (NULL == (p = __archive_read_ahead(a, 24, NULL))) { - archive_set_error(&a->archive, - ARCHIVE_ERRNO_FILE_FORMAT, - "Truncated ZIP end-of-file record"); - return (ARCHIVE_FATAL); - } - /* Consume the optional PK\007\010 marker. */ - if (p[0] == 'P' && p[1] == 'K' && - p[2] == '\007' && p[3] == '\010') { - p += 4; - zip->unconsumed = 4; - } - if (zip->entry->flags & LA_USED_ZIP64) { - uint64_t compressed, uncompressed; - zip->entry->crc32 = archive_le32dec(p); - compressed = archive_le64dec(p + 4); - uncompressed = archive_le64dec(p + 12); - if (compressed > INT64_MAX || uncompressed > INT64_MAX) { - archive_set_error(&a->archive, - ARCHIVE_ERRNO_FILE_FORMAT, - "Overflow of 64-bit file sizes"); - return ARCHIVE_FAILED; - } - zip->entry->compressed_size = compressed; - zip->entry->uncompressed_size = uncompressed; - zip->unconsumed += 20; - } else { - zip->entry->crc32 = archive_le32dec(p); - zip->entry->compressed_size = archive_le32dec(p + 4); - zip->entry->uncompressed_size = archive_le32dec(p + 8); - zip->unconsumed += 12; - } - } + r = consume_optional_marker(a, zip); + if (r != ARCHIVE_OK) + return (r); return (ARCHIVE_OK); } @@ -1933,6 +2638,24 @@ archive_read_format_zip_read_data(struct archive_read *a, case 0: /* No compression. */ r = zip_read_data_none(a, buff, size, offset); break; +#ifdef HAVE_BZLIB_H + case 12: /* ZIPx bzip2 compression. */ + r = zip_read_data_zipx_bzip2(a, buff, size, offset); + break; +#endif +#ifdef HAVE_LZMA_H + case 14: /* ZIPx LZMA compression. */ + r = zip_read_data_zipx_lzma_alone(a, buff, size, offset); + break; + case 95: /* ZIPx XZ compression. */ + r = zip_read_data_zipx_xz(a, buff, size, offset); + break; +#endif + /* PPMd support is built-in, so we don't need any #if guards. */ + case 98: /* ZIPx PPMd compression. */ + r = zip_read_data_zipx_ppmd(a, buff, size, offset); + break; + #ifdef HAVE_ZLIB_H case 8: /* Deflate compression. */ r = zip_read_data_deflate(a, buff, size, offset); @@ -1941,8 +2664,8 @@ archive_read_format_zip_read_data(struct archive_read *a, default: /* Unsupported compression. */ /* Return a warning. */ archive_set_error(&a->archive, ARCHIVE_ERRNO_FILE_FORMAT, - "Unsupported ZIP compression method (%s)", - compression_name(zip->entry->compression)); + "Unsupported ZIP compression method (%d: %s)", + zip->entry->compression, compression_name(zip->entry->compression)); /* We can't decompress this entry, but we will * be able to skip() it and try the next entry. */ return (ARCHIVE_FAILED); @@ -2000,11 +2723,30 @@ archive_read_format_zip_cleanup(struct archive_read *a) struct zip_entry *zip_entry, *next_zip_entry; zip = (struct zip *)(a->format->data); + #ifdef HAVE_ZLIB_H if (zip->stream_valid) inflateEnd(&zip->stream); - free(zip->uncompressed_buffer); #endif + +#ifdef HAVA_LZMA_H + if (zip->zipx_lzma_valid) { + lzma_end(&zip->zipx_lzma_stream); + } +#endif + +#ifdef HAVE_BZLIB_H + if (zip->bzstream_valid) { + BZ2_bzDecompressEnd(&zip->bzstream); + } +#endif + + if (zip->uncompressed_buffer) + free(zip->uncompressed_buffer); + + if (zip->ppmd8_valid) + __archive_ppmd8_functions.Ppmd8_Free(&zip->ppmd8); + if (zip->zip_entries) { zip_entry = zip->zip_entries; while (zip_entry != NULL) { @@ -3146,3 +3888,5 @@ archive_read_support_format_zip_seekable(struct archive *_a) free(zip); return (ARCHIVE_OK); } + +/*# vim:set noet:*/ diff --git a/libarchive/test/test_read_format_zip.c b/libarchive/test/test_read_format_zip.c index 6cc25a667..b965299a4 100644 --- a/libarchive/test/test_read_format_zip.c +++ b/libarchive/test/test_read_format_zip.c @@ -26,6 +26,76 @@ #include "test.h" __FBSDID("$FreeBSD: head/lib/libarchive/test/test_read_format_zip.c 189482 2009-03-07 03:30:35Z kientzle $"); +#define __LIBARCHIVE_BUILD +#include + +static +int extract_one(struct archive* a, struct archive_entry* ae, uint32_t crc) +{ + la_ssize_t fsize, bytes_read; + uint8_t* buf; + int ret = 1; + uint32_t computed_crc; + + fsize = archive_entry_size(ae); + buf = malloc(fsize); + if(buf == NULL) + return 1; + + bytes_read = archive_read_data(a, buf, fsize); + if(bytes_read != fsize) { + assertEqualInt(bytes_read, fsize); + goto fn_exit; + } + + computed_crc = crc32(0, buf, fsize); + assertEqualInt(computed_crc, crc); + ret = 0; + +fn_exit: + free(buf); + return ret; +} + +static +int extract_one_using_blocks(struct archive* a, int block_size, uint32_t crc) +{ + uint8_t* buf; + int ret = 1; + uint32_t computed_crc = 0; + la_ssize_t bytes_read; + + buf = malloc(block_size); + if(buf == NULL) + return 1; + + while(1) { + bytes_read = archive_read_data(a, buf, block_size); + if(bytes_read == ARCHIVE_RETRY) + continue; + else if(bytes_read == 0) + break; + else if(bytes_read < 0) { + /* If we're here, it means the decompressor has failed + * to properly decode test file. */ + assertA(0); + ret = 1; + goto fn_exit; + } else { + /* ok */ + } + + computed_crc = crc32(computed_crc, buf, bytes_read); + } + + assertEqualInt(computed_crc, crc); + ret = 0; + +fn_exit: + free(buf); + return ret; +} + /* * The reference file for this has been manually tweaked so that: * * file2 has length-at-end but file1 does not @@ -312,3 +382,380 @@ DEFINE_TEST(test_read_format_zip) test_extract_length_at_end(); test_symlink(); } + +DEFINE_TEST(test_read_format_zip_ppmd_one_file) +{ + const char *refname = "test_read_format_zip_ppmd8.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (ppmd-1)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0xBA8E3BAA)); + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_ppmd_one_file_blockread) +{ + const char *refname = "test_read_format_zip_ppmd8.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (ppmd-1)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 13, 0xBA8E3BAA)); + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_ppmd_multi) +{ + const char *refname = "test_read_format_zip_ppmd8_multi.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (ppmd-1)", archive_format_name(a)); + assertEqualString("smartd.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0x8DD7379E)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (ppmd-1)", archive_format_name(a)); + assertEqualString("ts.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0x7AE59B31)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (ppmd-1)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0xBA8E3BAA)); + + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_ppmd_multi_blockread) +{ + const char *refname = "test_read_format_zip_ppmd8_multi.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (ppmd-1)", archive_format_name(a)); + assertEqualString("smartd.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 12, 0x8DD7379E)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (ppmd-1)", archive_format_name(a)); + assertEqualString("ts.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 13, 0x7AE59B31)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (ppmd-1)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 14, 0xBA8E3BAA)); + + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_lzma_one_file) +{ + const char *refname = "test_read_format_zip_lzma.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (lzma)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0xBA8E3BAA)); + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_lzma_one_file_blockread) +{ + const char *refname = "test_read_format_zip_lzma.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (lzma)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 13, 0xBA8E3BAA)); + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_lzma_multi) +{ + const char *refname = "test_read_format_zip_lzma_multi.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (lzma)", archive_format_name(a)); + assertEqualString("smartd.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0x8DD7379E)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (lzma)", archive_format_name(a)); + assertEqualString("ts.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0x7AE59B31)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (lzma)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0xBA8E3BAA)); + + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_lzma_multi_blockread) +{ + const char *refname = "test_read_format_zip_lzma_multi.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (lzma)", archive_format_name(a)); + assertEqualString("smartd.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 12, 0x8DD7379E)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (lzma)", archive_format_name(a)); + assertEqualString("ts.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 13, 0x7AE59B31)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 6.3 (lzma)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 14, 0xBA8E3BAA)); + + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + + +DEFINE_TEST(test_read_format_zip_bzip2_one_file) +{ + const char *refname = "test_read_format_zip_bzip2.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 4.6 (bzip)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0xBA8E3BAA)); + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_bzip2_one_file_blockread) +{ + const char *refname = "test_read_format_zip_bzip2.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 4.6 (bzip)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 13, 0xBA8E3BAA)); + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_bzip2_multi) +{ + const char *refname = "test_read_format_zip_bzip2_multi.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 4.6 (bzip)", archive_format_name(a)); + assertEqualString("smartd.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0x8DD7379E)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 4.6 (bzip)", archive_format_name(a)); + assertEqualString("ts.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0x7AE59B31)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 4.6 (bzip)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0xBA8E3BAA)); + + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_bzip2_multi_blockread) +{ + const char *refname = "test_read_format_zip_bzip2_multi.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 4.6 (bzip)", archive_format_name(a)); + assertEqualString("smartd.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 12, 0x8DD7379E)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 4.6 (bzip)", archive_format_name(a)); + assertEqualString("ts.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 13, 0x7AE59B31)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 4.6 (bzip)", archive_format_name(a)); + assertEqualString("vimrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 14, 0xBA8E3BAA)); + + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_xz_multi) +{ + const char *refname = "test_read_format_zip_xz_multi.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 2.0 (xz)", archive_format_name(a)); + assertEqualString("bash.bashrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0xF751B8C9)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 2.0 (xz)", archive_format_name(a)); + assertEqualString("pacman.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0xB20B7F88)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 2.0 (xz)", archive_format_name(a)); + assertEqualString("profile", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one(a, ae, 0x2329F054)); + + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} + +DEFINE_TEST(test_read_format_zip_xz_multi_blockread) +{ + const char *refname = "test_read_format_zip_xz_multi.zipx"; + struct archive *a; + struct archive_entry *ae; + + extract_reference_file(refname); + + assert((a = archive_read_new()) != NULL); + assertEqualIntA(a, ARCHIVE_OK, archive_read_support_format_zip(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_open_filename(a, refname, 37)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 2.0 (xz)", archive_format_name(a)); + assertEqualString("bash.bashrc", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 12, 0xF751B8C9)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 2.0 (xz)", archive_format_name(a)); + assertEqualString("pacman.conf", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 13, 0xB20B7F88)); + + assertEqualIntA(a, ARCHIVE_OK, archive_read_next_header(a, &ae)); + assertEqualString("ZIP 2.0 (xz)", archive_format_name(a)); + assertEqualString("profile", archive_entry_pathname(ae)); + assertEqualIntA(a, 0, extract_one_using_blocks(a, 14, 0x2329F054)); + + assertEqualIntA(a, ARCHIVE_EOF, archive_read_next_header(a, &ae)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_close(a)); + assertEqualIntA(a, ARCHIVE_OK, archive_read_free(a)); +} diff --git a/libarchive/test/test_read_format_zip_bzip2.zipx.uu b/libarchive/test/test_read_format_zip_bzip2.zipx.uu new file mode 100644 index 000000000..6d9884aed --- /dev/null +++ b/libarchive/test/test_read_format_zip_bzip2.zipx.uu @@ -0,0 +1,19 @@ +begin 644 bzip2.zipx +M4$L#!"X#```,`#TQD4VJ.XZZ-`(``)`#```%````=FEME,ID:&U-'E`T&C1IIIZ +M@8&J>$9":`IZ@``'J:/2&#```````!@`2FB$T:31D9"F@8FC0`\H^:+Y;81F +M6OH?UN^\U&IFQOP9(.Z29 +MPY-_:/D\-$@RAM/2I>(FNV$2K!!&6I"Q+LXG0-YLG/4-JQE +M=%^\,6JL1H*""1!1W]4\/.E))/*(3J)$L1_D*SPD +MAZY)I5-#7@)=-5H06@2PP-)YSB5D!.3J63\EJ*\X%RMDE]>>,6XG5O@<^@:T +MT671Y,.:/=ICJ(=@\88>!>M^W)$;,QY(*`F*%Q2YZ^%TF$V^/85:V;L\W53/ +M?9[-IX&M.O38C\;C@Q';$E9[,C'?<#304$$9O&4U<2"$^;O(0X:)KR2UT?=.E*K!>,09(4\>BYUO&*8;D4 +MY=51BK5L/;$+WB"S8Z-?)M+GT/]^;,7$S'N0)))KF]$Z(GL[.L=F*G#!4NC) +M4)I$P%W)%.%"0YPC!DA02P$"/P,N`P``#``],9%-JCN.NC0"``"0`P``!0`D +M`````````""`I($`````=FEM-]>-Q`H``"L:```+````H'J-##(C0U&1)Y*>F +MF4:'E-&0:!H``-``)30B!0]3TFDVD])ZF(::`!Z@`T9#30,)"0@FBGD*9-IZ +MIZGE,F@9``-``&0'`````#````````)$00FC1,(TE/T:,HTVJ?J3/5-#,4]3 +MU#TADVAYI!A\^>RTQ(BD3$3I$[^,/; +MLS`?JD6R.G5Z9H!*U7PV]92S^P:OR\HL0A0_3>$BI8)(/R)[BOOLC2%8YJD( +MA.HIB.`H&#CG`1Q7XZ?+[_/^7_[H%_86(A9>\@A,_,XZ=,XL5W'7 +MHP4M+H9TD.>'-*H@_0+C17D%)P0)RW]IX%[$"C`LP1H=G0_&E!1":MS%AMXQ +M.;)R8XI%.5\07+:U'.MSO0WA +M_'JHZ5GH`I2IS+,0&673'.$+E#O<4'!QIOJ(=A]MS$!DF&L:[UWD`N@'J&ON +MK"9F:A.4,3W/HWR,FFXKDW_PF?19FN: +M8JZ13*&(2TCX35$",$])GPS2BLBY&_H+ZVL;>UD^#7W[$P11S-:X+;I&4P_2 +MY2FF=^Q>)*S'RAW&B*<0LART09,I)(!+E@9";HI&D+G>(HXB:0(%/23%!+S@ +M:NR/5*I%\9CK3:PVI^LQ +MG?A&-RQT6VZ[1QPG2Y9UVU<7W0N_@W]>`:ZV*-IMC1AC'"5PL)0>9PPF\/OD +MXXHMUZ\%[^LVOK%3L]),1F)*MOQ5)H +MBOICX8!01L:?$[E<$R]4`!$&F!"N^Z]2A2JH&\;#/"R\MG<`!XM)L2.#%FT@ +MI$"[T#%/63SY#K=LX#RSQ*%:^J%L1'HN2F*:.2LQ2X?"^Y^<9@U@PV(.4E%C +MNS^(@N"/+X8L:9,0T^E]7FZXBP#;&@K7#\.W[MOUG8-`-,<'O._DJ$2@7*?N +M#ZO8W/#Q7.=4KXGTY)7:1_NJ7-3FF\<6*$<^N]+-!C)W4+U60F$@>:,&%)NE +M6U9J0*F"F4!.R,Y`.G*"*P_*(P>-<7WV6=I7*E]98+HG":L0P4Y,?%RY^X[* +MDSY./:P;H^X)PU&L=#"2/IS$=1]X>MB:/!;68(&.EX`-X^"ZW_F;B,/8(F45 +M`_?U2XOBQ%&$HQ1YCTEC[_A42U:75Z>4-MI3.*/X>72/`C"+5Z>0B4C49O3A +M'N8'1IOJC95M$"-@1RC6(E"P.0[ZZ:;J?E0+#0?LP-6OA,=$^`?$]R)HW%K-',>W@ZBCI&CX:1\Q,YOY&7QLLV_J,<-^S+:AAO&'.A;*V +M3AQROQ<,BPB$?,M'!#1QR@/4P#TM5"<*D.$6U5>6NB0:@2#RD1AD +M(26E6M&U7;I&M?><,-FD48N@:,:2[U9AD#8DO`F8G`NH2VG:1>/.?`E4V(J^ +M#5)1$;.G/E?/U7S,RZJL\[5*#)Y+2^&5-1.L.G9B!:4)5`06C6IFT)6E9!%+ +M2NC400RE=1Y`*MU7P;V=>_94Q:13C>/8$;NI99OUE6QCKTH&-VKEJ+?SICS8UQ>&55 +MS!AN47T#J*?QEQ6,M2+`@9;Z#.*GM;18!J.<)J@Q;$1#.I*.KS4[,-,#$>-\ +M)PO&"I/*D7`.K'08;Q%TA8:$`G^HT8=;C3N5_4;IF)[F0N$3TDH*Q0-P<[42 +M,33G53,R)8(253>'`I(/(`C4Y$P=LZC+Z#,LC!<`8.=$0&& +MHL)>D\@AT(1A;6.-(URT`6J-@0P%Y020G3VP:QRCZPM[@QB;&[&]I'1L3HG6 +MHMV_<`9H[#0AJH/K\46RSPQC,;0Z:L!U8.MI-D=NTHM!=_;/K:@AM+BTH-:,C49$_(R)IC&^1Y +M[EI=1!L8O8:A+.F,:91'!%LGX.Y*[9W'8Q>2DN$BIBQ(Y()FJNVR"%T-;E6?@6^'&XMHR)AAWR$C9972+M#HSN- +M);+!>$HVAW9>(FP:U,1BUS180R6S:<$UD99OL@!``!4`P``!P```'1S+F-O;F9" +M6F@Y,4%9)E-9[2!3B0``35^``!1(A>=")0$"`+__W^!``=V;:TML-3U*>D]I +M3`F]4]-(T9`T:&@T8I`F(U'ZIZ@!H`>IZ&`````P``"4T4R9-3U`$T]$`,0: +M9KMEA9T)[OUF#<]_H_:R!%VW(Z5<8GI,92Q$\@_99'.V;0!9M_JQW(Q83@(> +MU(&RK]97O.4,(?9Q\O->C6^K@,2&/*Q^ +MQ&&;:95MR[%C&6KSOTJXV':CWZ0#H?FO+]L:K$L-@\J??66K"[C''`Q(05A5 +MAA`H<]68]Z8WD3,92%/*1-[+TH^&FHBX0;F(=TXT`&\A85\I0".^@K=W'6BV +M6W7I=FG(:-GP/>J!T.)K9&E&`70E\*O7<@B"XC8N],ESUC>J<#0$&LF`$Q"I +MPFSXM(AMXP]!B.B$A)+*A@FQI7N,888C3<<1)$V8APW#69B,^)R&)J6R?!9U +MVQ?F1G9TTEH&P"PV"U$F)!()!JGQ4/1670D)VLF!P-;I.\BHY"F((+!C$AZJ +MZ4L(AR$T$'4N]0NK3DSCOKGQ@9(O?$J:V8>5Q'*QHI5Q[SIJ>$D/XNY(IPH2 +M':0*<2!02P,$+@,```P`/3&13:H[CKHT`@``D`,```4```!V:6UR8T)::#DQ +M05DF4UGG",&2``!,7X``$'3GX%(B(Y<`O__?X4`",3M`&&IZ4RF1H;4T>4#0 +M:-&FFGJ!@:IX1D)H"GJ```>IH](8,```````&`!*:(31I-&1D*:!B:-`#RCY +MHOEMA&9:^A_6YQ\4L6YH5-&(7`A#+:]%S/&IQ$[W/8H;ZFA%01)[[S4:F;&_ +M!D@[I)G#ES0BEILMLL482*!*U'E=+=]T@I+>3,SRC1"`R6N_&W"N!*6`ETU6A!:!+#`TGG.)60$Y.I9/R6HKS@7*V27UYXQ;B=6 +M^!SZ!K319='DPYH]VF.HAV#QAAX%ZW[ +M2:!4(E-IS87%O<(M4.)02QAK:[^IR!Q:E(+-CHU\FTN?0_WYLQ<3,>Y`DDFN;T3HB>SLZQV8J +M<,%2Z,E0FD3`7-]>-Q`H``"L: +M```+`"0`````````(("D@0````!S;6%R=&0N8V]N9@H`(````````0`8`(#N +M7F'.E=0!`,YV<T*``!T)82Q1PWAL +M+U`,N0L_$]^&650C/X$D6#4QFD$\A/"_![4!O/5O/!KH`WCQ*4?T2*]4P#/D +M0'9I?EZG=N69Z0V;H0I=CP*$?".I\ +MGMG/80.A'^W>R4J'S/CZ%P`8`>F=R>R&R$2T@EM#X)"OQH1?A7,`:4IU9WV! +M#2W*DXT',;.4YIN4A:-X)O=IREL201ZSOC=YSAU[C4-::/YV8\)%"L17+>VC +M%/'B]ZCQN$2(Q*9*\KJZ`Y131`]5C&G';@1S-QES_RZF!2OX45@58+??ES%( +MUJ<(\`11M$NO)HK#/MK-9RT"15.2I:IZN8VTM1_?$G\L#BH67]$S%[4 +M%C-$\Q<+./&HV](4,7)OL-@C^M0F"2O!0N$OHOW54H87^QLBQVH*D%A<#SI% +M/#+-5U(W';:KC)RE>0Y^5YI!RECQNR"R4.UW9IR!@:B!UB8?_D5$FT8YCJHJ +M2[2"-&-_D2BJ6#XK[6G=%K"%;'^-+0]FHCY4ER#`^-]>-A`H``"L:```+````?C+)VT(T,`U9;I1/'N)!6C.3M-O9LO]H<^=OZF\ +MXD\QEZIMP!JM4)W+(F]%N;U#&+F]WZ%S9>>]&^75:X)0Q-5>)IRT/%INC887 +ML0UF,PS8J@3;WGH95EW4\KBH,6ZK9,'![3$-(WL/C<[0.6%L9@]0;]_>&[A'#Q^##`6+*PM@YT+]E[B9 +M;QK38:)CO+S]4X2KF'=)BVO-GM+_#]PB0'*J89+$Q11IG\\+84M](58^RW.E +MGM3S6RH]#73P.KV0V,2=,/A!;G>2Y4R!!.4(0U+D<.'5V*%./ZZYD")=W4*4 +M\SL"IUJLN])WM`-&4L&!9&U_-1%%SPU0N4*.^L5&33"A5MZ;@8F+I^D7JB'K +ME&,P>"UI]NT;ZX`UU->]V$?KZV(DX&-6L>? +ME,(J<;DNA6/2.<9$>?-P]_!S8L3/2P@[C'H$FOA$WUOB>;X$)E7;K+7#X9M( +MM+H^AK#CML0LXTO#P#I)J_5"@O +MMGH9+OXD^P7AG@<%QGH<>FIWE]/>L)6Q2Q\CC=[?TCG262"1EO)'_W_C_K4< +M8.7YNR"3\9WTQ"0-!>^,M>WT.Y0A&3AK.4/YN5EW]"&HQ1=8Q-NHQABVH7$% +M24Q:J/<8%F1EI)-!HND&FP_;]E,JI$*AGE&]Y#FJ5R6<(UM;;>U?KU@OQ,M_ +MI/+:;0^0'8JA"==H5/;]Y10P&PS,1_M+D1I>`%>_Y\&F[_R/##784#0AQ1!O +M'8>IBY?;&;(VAOI$+Z[VMI>]WC!['W$&Z5/D+"6["[[.[P!/?9C/;CD66^J/OC) +MO0JXP'Y6]PGW8N@X(#-XXWMP"A%E$A8L+X>QIC[0P,UT0A!IYQ/#+>E60KX/ +M2JC'F-Z@_,-/$7+F5F$1\#^"CTSY;H2EAE,/8>F$^'5H!M2@D?2O@+NJ2&-O +M6IV-[<6F-JG0(,W-YJ%PJ<7T.F")'/+F_`F6B-HG436 +MJM-VK0?>9*R?K#]OHC90!]BCG'^D&%!JZ2T_&[RWV?0Z?T:#_\B`?:MRC$D5 +M1Z.)!/;3/=GT*T;DC/XLR;7S%VEP+U-M)-*V,1?M<377S9-`J]S4[OI4SR>V +M_?_H2WWH>\@*$V@`LZO?-'Z[4_4"T2R@=%! +M!\1]:4M((1B=ATAJ)23HE7>]=,7I#SK5'2OUZLX0LX*_67!+6S:*0UL)^GS6 +MM;$2!(-TY%4ME1>*'M)S'[0JP^&J%"-&A[C1YE;1:X_4Y&K9*2 +M'14NC*]O(JXNLX#M_(LDE6$G<0,+V/;44]-*^S%7G*.\$W&>2=9,EYB)/@XK +MX+E4?>NIP=*X]Z>3T,V?!"5:/M@X-H%DZEV,*9WIK;T_C18UC5$%^Z5PY8;USNB?S]`D-K2V2 +M-C#0-45C'YFF>#*$9JZ!$>%0;=VQ@LGY"J<.C.10"1'JL^B(VXPNK^]Q97E\`HX0MG2)F=>38R,IFGK\R +MVH!.V3_)_D,II8A=SL;X1!2D9E67=.$P+_D6&'X^1;4%:`ERV>!,''7<#=NE +MP')[\,A89G35\;ZN:IIY5DDSB$(8D$[G>7-L!(+(-7=LQ0Q3=S`L-)R9YT;S +M;%ZYVD%6)M=I]*QS>!3='"SB'E/4>J#EQ.&]0*U0H/Y1%DM].)E/UFD0?Q)B;_;;)-K_]6Z<)D_>^[=?49%4D4B@REC` +M,_KU,KX-N,44CL/:*^D>7X(;6_1&C;<;&`K3+4NY[=4%K2K:U"Q%T9)',7," +M+PC``PBBDVNY9V-LN'V')B4@4;OH(\/M%UBX$S."4_!;&=9W\L`!2O\!ANM5 +M8+""#5J&/M[2;M?47WN=0[4C*$7D#QB&`!*@5_0 +M9(5SZ7,Q"N?0W!\0WGQF,0JKH[G-(<6^PMJZV&VHQIT?VKJ^7^!EW?)`EC`$ +M0%,=?X0ME@T_&5UXJIJ9)5(<[=W]M#Y.$.96:_T8WM/H+"T_Q;QZN"`5DC&' +M.+(^")0XO=#&#@KVL2R9+.#F)P02=`6PT[21&>T*Q3Y2WVL><:PHB)R?*'+3 +M+,OK(7PNXHXMVFT"!)??8!H^I,&V +M8BOE(Z/VYY(5WLSG$.+8$;.A>]+PVB`WV,9M@YL='Q_<%2STU?_IG.I17G^L +M>#-=')K6T?C[^&,7Z)0!B[2-)2:&?*5, +M,;&-HI_/O(/V"4-:_E>0Z)=C=1/9HQJO$R:(&H@4RP"3%T[BOBZ`3_2>)K49 +M\ZL-26)MQ\%(>T0S5A4>EUCU[),JI"_/9<5)*6X\97`N:8YSR0%)@*#_'9'? +M/E=RFK"L&(#RS$S_0[?/!DW.`\FH$M\#"JSQ@*&C*W\`4$L#!#\#```.`(@X +MD4TQF^5ZO0$``%0#```'````=',N8V]N9A`"!0!=`"`````1B`2H,L'S88)4 +M*WW.8!PX^&C9J^P?<4.RSJ^WF7L-W#+OB4@(A-83?&.PVF@$.:VM<4\8K(_6 +M`CRI/@MR:,)C8$&U*MZKA\W'KCI%E2GDYYE/SB'*\S@5]NO,ZBF!;HY@ +M3Q=F"]@P02:6^LC>JT$T7BB(.:*G`I.?%/53CC#4I0N_ARFXFG`DZSI*$)UM +M1^X:IZ^^)T*5E,C6L<,L%=ASF-1:0CP6*FX_;?R;#]C:/ +M)C@@M]\FLU4*%H^>8VW4@@%D0U!02P,$/P,```X`/3&13:H[CKHY`@``D`,` +M``4```!V:6UR8Q`"!0!=`"`````1"`2HPGR1L@-2P5SX3!72_Z.U"NE"C>=! +MB(;_TY1PR.H(G3<^QXEA+%'#>&PO4`RY"S\3WX995",_@218-3&:03R$\+\' +MM0&\]6\\&N@#>/$I1_1(KU3`,^1`=FE^7J=VY9GI#9NA"EV-P&-RI:_YA-3= +M!E?#E$05TTLL[G6$<%[`H1\(ZGR>V<]A`Z$?[=[)2H?,^/H7`!@!Z9W)[(;( +M1+2"6T/@D*_&A%^%+WJ/&X1(C$IDKRNKH#E%-$#U6,:<=N +M!',W&7/_+J8%*_A16!5@M]^7,4C6IPCP!%&T2Z\FBL,^VLUG+0)%4Y*EJGJY +MAS2K)[;2U']\2?RP.*A9?T3,7M06,T3S%PLX\:C;TA0Q; +MZC0FI6JUP3:]D^3L->OKP>:G2P_FF8)V>\LVQ^;MC'-2N-,GXD"Z]PZI8A79+7ED*DB,K&N'OH,DT/C4$L!`C\# +M/P,```X`D3B139XWUXV$"@``*QH```L`)``````````@@*2!`````'-M87)T +M9"YC;VYF"@`@```````!`!@`@.Y>8EJ>3<8@F_*<(\B>K]4_(WC8#)`_QSG+`7`B&_11VIJ)@#(K)_I8R^=`VUO2_S,1C=1CAU>-*`]CC+&6Q;EE'#CG-W=^[,F,+UR-TE(9,G +M1_&$NYD`,5WY5QX7@%5XX-*7+7&#W'1#XF+"?6*U!H=B063.]NTA`1+&\J(K +M&U]DE8$ZYUU'UT\$!88I,6+(M/"=_+3@V5H+24 +MBA0>^0?[W07K"B%SQ&OA-R)S9W>DH,<4##E@?5YF:%K$?5U@W[IVG+2#8FC4'GT3-$L%_N,B$^)UY1GAD=.Z(HI#3[T"979&<`+ +M`S]9G+PI';5"==`18UE,>N"?>"_C"MSHX<[!&8$+A6U_7][TK*6.^\-O-UK! +M`)>[*.D1:1H!I,?PEIPW,NQ5CCX2NCY+%UH"T5X!O$&=6+X#"()33:]FPF:F +ME-O)\652#KD$^ZFBJ7K`55]R/A"OA"T7\R6K%B(2&6*/H0>*-:@< +MJ2V9^KJSL+P4A&$`4$L!`C\#/P,``&(`/3&13:H[CKK8`0``D`,```4`)``` +M```````@@*2!`````'9I;7)C"@`@```````!`!@``)=1P<:5U`&`PJ3,QI74 +?`0"74<'&E=0!4$L%!@`````!``$`5P```/L!```````` +` +end diff --git a/libarchive/test/test_read_format_zip_ppmd8_multi.zipx.uu b/libarchive/test/test_read_format_zip_ppmd8_multi.zipx.uu new file mode 100644 index 000000000..6fadcea64 --- /dev/null +++ b/libarchive/test/test_read_format_zip_ppmd8_multi.zipx.uu @@ -0,0 +1,84 @@ +begin 644 test_read_format_zip_ppmd8_multi.zipx +M4$L#!#\#``!B`)$XD4V>-]>-.0D``"L:```+````X%[[99#N5_N2Y@T-?"T%*8O@,-K>QSOR>-).,0-Q:=7)+:*Z&:B9' +M-YXDKOWK"T?"(.;F$3'A>TW+.W#W%Y0`:_5XI+35&4_L<0>95Q!XP,\T=VCM +MZB%S4G,4ED*-4T#5Y'BUCMS!0Z)0*50U=EQWM_F]2)<8#VF(432?^.!'O`I/ +MLP!A8AN1>8;MQ:S?/4-$K07X@["K?T]0<=84:OJ]Y$.D)3MNMK`U.(ORV3M& +MV`2EDVHT&D'F-=\,_"^;UGWW!=,KJI4^NZAF72$S=N.*PW!#ZJ?KQ0;$7E!] +MFC,:ST/]Y(,:K3SK7,%&A_%1`($9H@/\G.RN$J7BA?C\P"[U&[8MITK*`U5E +MVH'&CJ8?%1+H-H0=B*8]_(F;;/P`371(@?QN2T<4C52L;%S-XV),E4-V05&I +MA.V?U+HM63;JH=1]LJ_E"@JM?QH\6#NQ7I\I*%-!)LWB?U):O^--2R=1C):^ +M^@GFC0GETH%>LXD8_D*J'KS5[/M5+89M;6@3?]I.`M=;(S7*1?(?.9G!`S26 +M([JDK/(RCM%WYC`>G*L7:NJ!5TS[7MI^E[4/\RU9$&1XJ5#V;<'F^AX^R@P7< +M/2?UB;-2+I11]^X;,J\:,'^R=&8"@[EI0$J:":*LQ-F(IOSXC?1W>/;?4$P* +M;=V"OEY@S\%3'&F-^K0FME.]%1DTO*OP8.:M('3W)\S"1.@8# +M6Y1&@L+#!YVL]8.,[UP/\HMAR1"O!M/2?DY-8[N'P[T&2J7-V5%=XWMRQ,EM +M7TL^99!FXY0+B1^7K36"#!85"WXE,WFEU=B!+GX2$D<)MW>&A,8A9RPL]CR( +M!=1JX99-:5D;PM@S4(>?BKTFAQ>"%9\Q^Y\6T\ +MTS"*[M.TW7AE:#\(:M_JLJ#*-9-O,UW\T:PVW8G01OR!B59P=[:XNH0Y=OC9`KJ_%.>^)RW-,G^%%(U9HC957G/\@ +MW1\J*,JLX*CF*,-!'^O*E7=V#S8_#G`9YQ)#%,AJ8"-4B,OR8RW4@L)8F$JB +M.3QD;;M_L<:*T3B>8\OECY@*'6?[2VYXB3B!`<:65*HB/.^@;]\`IE)HTIG=_X:*3A'#M\P0L*]^ZFWC\UL`O;WL@#?"!(_]OZ4=B<\QL/ +MWD%/TL5[P\,^_3G\]WD-Q`?XU]K5LFW +MI6J$G\=W(Q+2_U&THGA'_6)`"WK=X74-OD_B%WJU+JH/A"K#JQ@)A<`>A6[H +MC2A>S7]ZV/H?/1KV&OR8B]IS\>Y:\5;B-:+E^V`:5_O[+(!,6LW0&[5(F4UH +MKM;R]_V\?PI6I+"7M1H%E':WL<1#J"O`>]S*UKS02ME,S7]X66@[B`_#?H$^) +MT&07IY[)18RK+`@#?UXT=8B=XC$J'SSJ>XP2RJ]>02[1J:R;0"M&L>KY(3*9QLH1N@K?LL%3J%O? +MR!>+JHS'L3_<*?;*81D1'/A-A"%%P..0%#Q32?L88\C'/'W?D +MG766\,!14W_TY,.;G,\C.+L1LR__7UR*_H!IE,1LZO,_+#?C[OW4+[4I`[7] +MZ;U%`5#@;6A5#JU]H+\E;^^LQBHQ$-T.P5.V/JS69'PI06.%#D4 +M%B.K!F(.G*6@/>[P-FU3-#%N.EY%W*]P_]<>?07:7 +MK&JYBO\0HXN^[/OA#_MP/QQG-+1?Z)7*WRL14*DMWL2795*UHMN/'POIN@Z= +M&&>Z/LT#HX00EWLZ%/G7)X<":!ST&P+Q0R)=V6!#D]TY1RYO!C&5&''Y<9UK +MK=LY46W;=`;(^L<>RMMG%(;??BF+AY,7=+H)6XSZ:2#^`W=RR7KIE>XQ/[_^ +M^$F4*967)DE3/^&\(7W?:EKRJ,PN@HOT:U_.!1-"7A-=LX\Y.=-37_355)/2H +M5V[LT(83$V*>?^YW7'6.SD=MYSM<`[M'IU]?H2ZU8\SLD+R-Q__8G +M2A&2#H5-J+*NQHQ&ZX$YO_(.OC\W[3B6WB#"MKV?*/<=(JBTKB`D_X'X+7"Q +MJ$B]N)L:OTXC':GK/!QYSR.,%%T3XI6(C(3&XZV`.ZLS=\O7>0 +M#N;$J[W)(Q.5<_I5G1K^FB10(.'.#Z!,B8Y'WX_=XOM8\+X/;Z\!"0HTWZG# +M[:0<%(2)`J>6>9=.B5&CWX.,@3RF>FN>'-0))N"!IY,\_?'B`DNYJ4'+!JK] +M\NX-@,,Y%)>#36Z7>CBBL=0S<@W(QRXPD3:VF-W)CS52HO!/T%NV[$0^B01. +MZH\`/9ZTF=$,",`0J\ANSPF&X.0O!?$8\^M%XSM$`D)W".C60(#6;('?*'.` +MTN9>^YI5M0P:-7MCSU2HCCU'`[!+^"9P9BH%1<=K_B#UZ3CDJWX+V\C^S$GD +M]M3M2VJY4''"]RN@>]D^K6B7ZGQG!<4'LZCK,BU]QH,1G(P$AB-W/G%7[.`! +M"P[DH--W#`"^#];7=H09``!02P,$/P,``&(`B#B133&;Y7IM`0``5`,```<` +M``!TCE(/-F<.B*/H_C$3G..AVGS/6(LHZG/Y[UI +MO1E2,[]ZU6`C\+WOIWU\>8)Z?4>]`A@"D*I3B52YGN&M&ME*-/]I!1,P.R@8#3-@,D#_'.`57C@TIL*(7/$:^$W(G-G=Z2@QQ0,.6!]7F9H +M6MRJ>G_FVK['10^FM,>V7UX1]76#?NG:?1,T2P7^XR(3XG7E&>& +M1T[HBBD-/O0)E=D9P`L#/UFL!57W(^$*^$+1?S):L6(A(9 +M8H^A!XHUJ!S!KSA%4.]<`XB*.5^_Q`%$6ON%MH"W>Z9/P^MG.I)@N0WI*EE+ +MX2HV+?E8)N&["2\:UYRI+9GZNK.PO!2$80!02P$"/P,_`P``8@"1.)%-GC?7 +MC3D)```K&@``"P`D`````````""`I($`````8FT!``!4`P``!P`D`````````""`I(%B"0``=',N8V]N9@H`(``````` +M`0`8``#P/%?.E=0!`,YV<5U`$`-+O[QI74`5!+!08``````P`#``T!``#O#``` +"```` +` +end diff --git a/libarchive/test/test_read_format_zip_xz_multi.zipx.uu b/libarchive/test/test_read_format_zip_xz_multi.zipx.uu new file mode 100644 index 000000000..cbb043469 --- /dev/null +++ b/libarchive/test/test_read_format_zip_xz_multi.zipx.uu @@ -0,0 +1,125 @@ +begin 644 test_read_format_zip_xz_multi.zipx +M4$L#!!0```!?`$JCP4K)N%'WP`0``,4(```+````8F%S:"YB87-H%H<2H0"6A +MIYB7=GGPH!_8+Q:$I09*"\)&%HUP2'X^1_;,\T5=!'^A#5>U*#)2^Z>L/*E5 +M%Y=SZRY&`Q_\(3_!//$I\E.SDZXL)WCSE7*0<#ZW/W*WM?^3CV,,)YEK01]T +MT,<7Y)@5O0':C!S_0^WG/9:,_0!\NG>*V2&:&Y6HU["3TUJ[%2F5'QVDJ49T +MXGT,!*J&0,&"B11L_MOP3I7+N6GR9P(1N,#/>TI'$'P9K@^7G77Y&\%&7G#U +M.(1KPE@G:PEDVEKT='HZIP\2CZXN#7@$BKO\_/`A&S$A2;^3%0VJ0B<>6Z7* +M'(W4;J5SB-RW4&#Z`/N2.@8_6$ +M@.Y1A%M"F_0Y9)K*3PGD$TRK,@"9C'8J9SEZHM`Y$LI/?PG%_42+$3TP@+I` +M+:FR6X1[JU^A\'RX,7__;W>6RC#^PQ8?M\"W+\9N+5D6E<,I6YVTKS#\$JVA +MZ;K0DY^(!*H9_',HH[YJO:)!+$6'J^.N'A`#`W;24DHV[N/$M_I"<*S-?4"H#MX!C`%KOF),."`K?#U%$R3O$^-3L$B +M4!;(W2EW.X`!YA8N;NJ=9MQ+0L:V/,NE<$C'^^VK:'5:ISTE3397ZF]A-RNV +M_0['T/F3^RM,C)6LKU.-W*KX4K%_0E/L1N*_EGPJ.$1^?N[5CQ%1QVWW7=PQ +MRBO*[C$);8)1VXE3/D#0,<6UN<]Z2^" +MDK"/$_T4Y*0'L%^0ZU9V*83*/A7Q58HO'/N:&4-[4C;Z"VG4F*YB&&-DUJDN>P.WW_W/(C8EWN-?5&H` +ME(KXMK!7NU+"*$'9`APUBFPA)L:VI),V)&P9=E5XX@LEL0`6F[B]V<#9XN[& +MRCL*9L(.?C[(@XE%VY>?B<'!&6!:!_F-/_H10MNL_AK01%_7!^2_ +MK-;P"DF0A;4'A4:'>AL55U%9:8*0`4^?W8+9DT"V0RD8LG/39(VE@V_M&V+K +MAMS@CU$V!O(Z>CG.!!.(;UXV'W%\:.5'^#G(S+5'+V!>C07US+?DIUL/62Z: +M&B4VHH.FA1S9(^$W)M3*NTHC:"FCSF;3$8P4Z@>L4)!K1Y*\<:F0-]$9L=HB +MY)E#TDD5?I!S4M5OOCCC'R/!D.`9%'K&$F[4E5VH"YP$-,%9!F'3#+:9192$ +M+UB^3S-M%M@;228EH`````````&9"<41``!4)ENHJ``*_`(``````%E:4$L# +M!!0```!?`$DZ/4N(?PNR0`4```8*```+````<&%C;6%N+F-O;F;]-WI86@`` +M`/\2V4$"`"$!`````#[C.DH'JHS8W0 +M#?$V8!+]F!+4KI<$@5HR#OF&_!+?L]N#167^N/Z\RR?GT6KQ]?CNM8>==$!O +M=?@QVU.*S@O@?_!ZYES=OXHQYJ$K_DR%MMLC^2[U):T3="8KH6E\^#TM[(#:A4]+,_]N-'V@&OOT`,J)*`1$6H.A\AQ<:=18H%@ +ML;39WZ/[&[?/?SG0:`)Z@@QK2DRGY04F&/;-K*Z2%^"N_BP_F6O6MO(I/U.T +M,L^0$-SQ!\_E:AA7X(VUV.Y`_]<]=`A6(D?^?S$N7OB; +MT(0U)^NL7%.Y4Y=[UNW8JWFUU2Z%N5G/Z:2V?%%%6/';]$%?%<3`>;=\3"TZ +MOQH>DR`T^1`8>\&>-+PCS8;)$M]WJ2O3&-M5NR) +M$WG>P*D3/:ZM,!QAAXTJ^?^DMH4:%0+):![`U`''+%K.;9,DM-OWB^)49!83 +MO-XVJ-N?)]_]87HXQ6(V0%4!R+DSR&(F=2*,KFTT2V$[`UZVVMLSI3T^B.%" +M3D->'YYCN2Z9EC%"/S^=W%139%7:"`%-3":\73B.TK]C)8:O)4=W6=.WU7J2 +M@Y#"B90[N&9_-Z@-5-#!_EI#N]D^NX!(OD[^I4G/L<]:Y&KD/P[5G?Q)6SR. +MDWT*.9G^KK"3^&1]@*+P6$#\Z4SCR,)E-_G'AVC:U4C%35O[,9B\$>QS9A// +M4-6)6TL?R;D":$KYVE+4" +M&J><3=ZH:XM&"9ZG5-4HEZVM9YO'8X@P#G1HSFI) +M2W.U-!_N1K4*.N#AAUV?:`+5)2Q\ST04<\TLD-92B`V7$2.54`>G]G_GE!`] +M_IBNF"]3MS#G]1S*N!MZE$UQX#1ICZMB[T'>_#A&8:3CP,B^UU\D$[3QNZ\P +M=]=4'BLWD`?H@Z@Z1;^A03I/VJ2$9]*>-&-,(=@TI22.6-'S#KS*+AT&AY64ZN7$Q'B,&TN<$OCO2W@';2/5 +MEKGSV4@OF10@NP4EBO7\Z_Z/ET#+&R::*OH_-<8/2\VF!^Y["EA#=>3&>K=? +M3^N#$@0V@=Y#K\T!ZU-B?E02#N6U>="G!A6&D?[?!.55MDR7@#`#^[T2I%,A +M+@+0O`/8^4,FN!XRSE;S@C93NICEP.5CEB_787X1B`&3L2FW@E,@?%HW,)36 +M&;CMA--B'7O,LRD\RJ3&]S+/5/3#/?F.Y/5_]51?5DUT3'R5UY.+M +M3HA.NI":EB`FI,"L@%9/?/J.E'FE+AIF8,A&N_G:M,4%DNJM5&/OK\3N="=5 +MYWM9VORAD0```9P*AA0```_X7#"H``K\`@``````65I02P,$%````%\`2J/! +M2E3P*2/$"0``R1D```<```!P`!&(!PA/3V#P('SF^\W3.1N]J/".J]^]]U@ZBN`&LJSNNZ7/6,;7A:'$J$`EH:>8EW9Y\*`?V"\6A*4&2@O" +M1A:-<$A^/D?VS/-%701_H0U7M2@R4ONGK#RI51>7<^LN1@,?_"$_P3SQ*?)3 +MLY.N+"=X\Y5RD'`^MS]RM[7_DX]C#">9:T$?=-#'%^28%;T!VHP<_T/MYSV6 +MC/T`?+IWBMDAFAN5J->PD]-:NQ4IE1\=I*E&=.)]#`2JAD#!@HD4;/[;\$Z5 +MR[EI\F<"$;C`SWM*1Q!\&:X/EYUU^1O!1EYP]3B$:\)8)VL)9-I:]'1Z.J-8JFWEH.CT0J&T-]!<(2-ZX3@)RN\J +MD0J4H/6N:=KFT$;\&+`]@#Q%F@SF+-%(65*CB7!$HAZK&W*TS\4$TS&_-!XUFUI$7*A-;',>)+^5/P,ZD`[HWR.A9G^#C +M-6KTNK(S\@+HL$6:2FA5[-R2LQ[3WK4@G[G^^#%XET%6G82[QS;7(S,^%>P. +MYKA-5S)0G[)N+EKS^1K@1W;SU#4':SY-Q*(HA+6WR\"M,7!$$QQVKNVE\P[0 +MBBWM8'\!):GY@^BZZ/>2OU-EQ9/907=8"CMA,Y1-O?0VM$[ +M,QLUL^VK@(FD@-<=6)L`J(Q]$R2&TIS@=/S2CA0:)DA&H\H`$K>#+BZ#L5!& +M]8O1Y2#,@L6`*Z_D^1@3[G%I48$DK(:8@ZH;WXDD0/L?(I^Z'Y4Y+\D8XG`( +MSAG!!38.H:+L%5M[]Y^`-CTFJ#CX<7/IC$#M`L)E=Z68`.Z15.!\)150B=YZ +M<&XZF>5(BYL_(:&P;,V^D';WU2KR?A0+4-E:0%O1F9`\";RR09,+G:>77.C7 +M]KV`K:4D9*B)Q7-\>:1&]N#JUDJ(<;NGJG2;_\!+&29%3S'E+/YN\/.CP_`+ +M2C&6BOI"R]9V%EB^`F/J&*.2?`KS5D\7+CM*1],_W<7&0.MI\<*%EQY7Q@SQ +M]J35[E'P\:C"YF9--<<]IB4/#I9@9S3^CAVB3]UU"M=C#6^!]'`R!N<"BA3D +MN!$ISF;-DTGCY>18H,[#4UN),"U;=7J/F]7_C]K`/LRZ> +ML--U<[]+X!IV+&GCQ]@_E!=D2Y1SD_+@R,2:B,26E+GCGX,S56]3OA<@OBVL +MP`N1,^ZJXI1U"XKCE:I(CGLU!K\^_VU7+TJ2/;81PC?GHW(SPC^[!L2+DA`2 +MY\B!@>P$N?Z%:.G%!3DJDM;N*"DN*/0 +M.&`?D?G/MZ-/H^LA*3<3>4'L">#EY&"!H9TB4M/1;]:<*%/-*%*ZZ2,FF"T` +M1-VLYD+N +M@QH1*4'INB@E6X<9,1=\T'`K<_PIF!F*CV&59B&:2$-.E\%OM2!\T"\X#?W) +M_\E.2DS4TN&FVR:GZQ#7+A4-!3L*KI`(NX#-0JA_OH3'J@OJ/G'C\>/BB@8\ +MO9AR=;IRPX06AZX1'^R/RE>AE,'6#4](`V?FM0=BN]-3P7I*Y1N1VF>PL"IJ +MGF]0@PM>JH[PYGC*F^0"*N9OI-E(9C^LSD&NHWHEQ=_$X4]3I1L][E,)5C%OKU.]ZXIDSD]%)$8-'2>$@08,C&F")$0X!^:[H:2 +M77[$CGQ]J0)I&=%/9U>ZZ!2Y93+,3^<38JQEC-]%YBW9?`S,U]*_N)5*MF1; +M6\2$>DXDY!G'+D3!)U47?ELRTEI9>)&;2]..PL(AM-=O/69L[N:Y:FYU$4OV +M^,N@H"C'ZH#0&TGK?'HWAY=!0Y!^[6!S_P:>A)^@YZ;N82NGABF]X@0;_,-? +M>;?A^0=`'34QXH8VW08:$;&O]MLZDV-KC!/$@J-K$9[N0&GOS+2GI;?LN)L!M0MJD)X0D:^^ +M)C-S@Y6IMQ)#AE,[..:E&_PK.Y=^1&=;H[0VN#H>-LQAW3+GYKV6I=X6B3=8 +M.@Q5EULK;>80C>Q2*R&+D7DQ0E3,-2X+C]$9NN9R62RYTV[PC,4H>F[\5?A# +ML..XN!<+40H+7WGS&4<3H)3#?4%;XFXM&K'OQ*Z'L*"N<J@/Y%[H1GLJ4SH:YM^-V'+5CKLSP:+]J--=`-B,9Z+53^H2Q]->KE@3ZYY +M\OZ`-KB^'V-<@\MC.J++G-TJ>RS-M>)-3OU&NT5CLUM2@\\A+I;`GDR'T11U +MAW1H\ZECZZD*=!@DMO3,60R-`\W7_X6C,8FB3C-2V +M(@S*@2LK1T%!';G7&%\&D>=1/;=64G$92-FPMB(<+37:DLED52:_VFEY47*O +M3ZKVJ;/.\O<_2Z?IB'U9W2:6S6_,X$F=(#VZO.@V-4CY<68\B@A`UKG0$AR1 +MC3-VV!4W/,85QD39F&!67\SGM$KS:=TRN]$V"EC=+O7\PN8.A__Z)$@0>8,] +M!NHPA*?U=-A#U>OEK!#NG,RB2Y@@`5JNHJR0;9A'3FP'V/.UYMG+D*L*">.% +MKY&BW>NHE92\QT;V$D8_X^)[*WA-4GBAS.]:Q0"SGL>J*KNDX)76,LF["/13 +M[SM72F%\\//WRT1^2GJA553G>V*&$Z9ZK.*:"0K;06@IC_N`Z$&Z4?!2=`CN071_Z=#```````!G1/),P``5\:%:*@`"OP"```` +M``!96E!+`0(4`!0```!?`$JCP4K)N%'WP`0``,4(```+`"0```````$`(``` +M``````!B87-H+F)A