#include <string>
#include <stdio.h>
#include <ctype.h>
+#include <stdlib.h>
+#include <string.h>
#include <apti18n.h>
/*}}}*/
unsigned long long Size;
};
+class pkgTagSectionPrivate
+{
+public:
+ pkgTagSectionPrivate()
+ {
+ }
+ struct TagData {
+ unsigned int StartTag;
+ unsigned int EndTag;
+ unsigned int StartValue;
+ unsigned int NextInBucket;
+
+ TagData(unsigned int const StartTag) : StartTag(StartTag), EndTag(0), StartValue(0), NextInBucket(0) {}
+ };
+ std::vector<TagData> Tags;
+};
+
+static unsigned long AlphaHash(const char *Text, size_t Length) /*{{{*/
+{
+ /* This very simple hash function for the last 8 letters gives
+ very good performance on the debian package files */
+ if (Length > 8)
+ {
+ Text += (Length - 8);
+ Length = 8;
+ }
+ unsigned long Res = 0;
+ for (size_t i = 0; i < Length; ++i)
+ Res = ((unsigned long)(Text[i]) & 0xDF) ^ (Res << 1);
+ return Res & 0xFF;
+}
+ /*}}}*/
+
// TagFile::pkgTagFile - Constructor /*{{{*/
// ---------------------------------------------------------------------
/* */
pkgTagFile::pkgTagFile(FileFd *pFd,unsigned long long Size)
+ : d(NULL)
+{
+ Init(pFd, Size);
+}
+
+void pkgTagFile::Init(FileFd *pFd,unsigned long long Size)
{
/* The size is increased by 4 because if we start with the Size of the
filename we need to try to read 1 char more to see an EOF faster, 1
char the end-pointer can be on and maybe 2 newlines need to be added
to the end of the file -> 4 extra chars */
Size += 4;
+ if(d != NULL)
+ {
+ free(d->Buffer);
+ delete d;
+ }
d = new pkgTagFilePrivate(pFd, Size);
if (d->Fd.IsOpen() == false)
}
/*}}}*/
// TagFile::Offset - Return the current offset in the buffer /*{{{*/
-unsigned long pkgTagFile::Offset()
+APT_PURE unsigned long pkgTagFile::Offset()
{
return d->iOffset;
}
*/
bool pkgTagFile::Step(pkgTagSection &Tag)
{
- while (Tag.Scan(d->Start,d->End - d->Start) == false)
+ if(Tag.Scan(d->Start,d->End - d->Start) == false)
{
- if (Fill() == false)
- return false;
-
- if(Tag.Scan(d->Start,d->End - d->Start))
- break;
+ do
+ {
+ if (Fill() == false)
+ return false;
+
+ if(Tag.Scan(d->Start,d->End - d->Start, false))
+ break;
- if (Resize() == false)
- return _error->Error(_("Unable to parse package file %s (1)"),
- d->Fd.Name().c_str());
+ if (Resize() == false)
+ return _error->Error(_("Unable to parse package file %s (%d)"),
+ d->Fd.Name().c_str(), 1);
+
+ } while (Tag.Scan(d->Start,d->End - d->Start, false) == false);
}
+
d->Start += Tag.size();
d->iOffset += Tag.size();
if (Fill() == false)
return false;
- if (Tag.Scan(d->Start, d->End - d->Start) == false)
- return _error->Error(_("Unable to parse package file %s (2)"),d->Fd.Name().c_str());
+ if (Tag.Scan(d->Start, d->End - d->Start, false) == false)
+ return _error->Error(_("Unable to parse package file %s (%d)"),d->Fd.Name().c_str(), 2);
return true;
}
// pkgTagSection::pkgTagSection - Constructor /*{{{*/
// ---------------------------------------------------------------------
/* */
+APT_IGNORE_DEPRECATED_PUSH
pkgTagSection::pkgTagSection()
- : Section(0), TagCount(0), d(NULL), Stop(0)
+ : Section(0), d(NULL), Stop(0)
{
+ d = new pkgTagSectionPrivate();
+#if APT_PKG_ABI < 413
+ TagCount = 0;
memset(&Indexes, 0, sizeof(Indexes));
+#endif
memset(&AlphaIndexes, 0, sizeof(AlphaIndexes));
}
+APT_IGNORE_DEPRECATED_POP
/*}}}*/
// TagSection::Scan - Scan for the end of the header information /*{{{*/
-// ---------------------------------------------------------------------
-/* This looks for the first double new line in the data stream.
- It also indexes the tags in the section. */
+#if APT_PKG_ABI < 413
bool pkgTagSection::Scan(const char *Start,unsigned long MaxLength)
{
+ return Scan(Start, MaxLength, true);
+}
+#endif
+bool pkgTagSection::Scan(const char *Start,unsigned long MaxLength, bool const Restart)
+{
+ Section = Start;
const char *End = Start + MaxLength;
- Stop = Section = Start;
- memset(AlphaIndexes,0,sizeof(AlphaIndexes));
+
+ if (Restart == false && d->Tags.empty() == false)
+ {
+ Stop = Section + d->Tags.back().StartTag;
+ if (End <= Stop)
+ return false;
+ Stop = (const char *)memchr(Stop,'\n',End - Stop);
+ if (Stop == NULL)
+ return false;
+ ++Stop;
+ }
+ else
+ {
+ Stop = Section;
+ if (d->Tags.empty() == false)
+ {
+ memset(&AlphaIndexes, 0, sizeof(AlphaIndexes));
+ d->Tags.clear();
+ }
+ d->Tags.reserve(0x100);
+ }
+#if APT_PKG_ABI >= 413
+ unsigned int TagCount = d->Tags.size();
+#else
+ APT_IGNORE_DEPRECATED(TagCount = d->Tags.size();)
+#endif
if (Stop == 0)
return false;
- TagCount = 0;
- while (TagCount+1 < sizeof(Indexes)/sizeof(Indexes[0]) && Stop < End)
+ pkgTagSectionPrivate::TagData lastTagData(0);
+ lastTagData.EndTag = 0;
+ unsigned long lastTagHash = 0;
+ while (Stop < End)
{
TrimRecord(true,End);
// Start a new index and add it to the hash
if (isspace(Stop[0]) == 0)
{
- Indexes[TagCount++] = Stop - Section;
- AlphaIndexes[AlphaHash(Stop,End)] = TagCount;
+ // store the last found tag
+ if (lastTagData.EndTag != 0)
+ {
+ if (AlphaIndexes[lastTagHash] != 0)
+ lastTagData.NextInBucket = AlphaIndexes[lastTagHash];
+ APT_IGNORE_DEPRECATED_PUSH
+ AlphaIndexes[lastTagHash] = TagCount;
+#if APT_PKG_ABI < 413
+ if (d->Tags.size() < sizeof(Indexes)/sizeof(Indexes[0]))
+ Indexes[d->Tags.size()] = lastTagData.StartTag;
+#endif
+ APT_IGNORE_DEPRECATED_POP
+ d->Tags.push_back(lastTagData);
+ }
+
+ APT_IGNORE_DEPRECATED(++TagCount;)
+ lastTagData = pkgTagSectionPrivate::TagData(Stop - Section);
+ // find the colon separating tag and value
+ char const * Colon = (char const *) memchr(Stop, ':', End - Stop);
+ if (Colon == NULL)
+ return false;
+ // find the end of the tag (which might or might not be the colon)
+ char const * EndTag = Colon;
+ --EndTag;
+ for (; EndTag > Stop && isspace(*EndTag) != 0; --EndTag)
+ ;
+ ++EndTag;
+ lastTagData.EndTag = EndTag - Section;
+ lastTagHash = AlphaHash(Stop, EndTag - Stop);
+ // find the beginning of the value
+ Stop = Colon + 1;
+ for (; isspace(*Stop) != 0; ++Stop);
+ if (Stop >= End)
+ return false;
+ lastTagData.StartValue = Stop - Section;
}
Stop = (const char *)memchr(Stop,'\n',End - Stop);
-
+
if (Stop == 0)
return false;
// Double newline marks the end of the record
if (Stop+1 < End && Stop[1] == '\n')
{
- Indexes[TagCount] = Stop - Section;
+ if (lastTagData.EndTag != 0)
+ {
+ if (AlphaIndexes[lastTagHash] != 0)
+ lastTagData.NextInBucket = AlphaIndexes[lastTagHash];
+ APT_IGNORE_DEPRECATED(AlphaIndexes[lastTagHash] = TagCount;)
+#if APT_PKG_ABI < 413
+ APT_IGNORE_DEPRECATED(Indexes[d->Tags.size()] = lastTagData.StartTag;)
+#endif
+ d->Tags.push_back(lastTagData);
+ }
+
+ pkgTagSectionPrivate::TagData const td(Stop - Section);
+#if APT_PKG_ABI < 413
+ APT_IGNORE_DEPRECATED(Indexes[d->Tags.size()] = td.StartTag;)
+#endif
+ d->Tags.push_back(td);
TrimRecord(false,End);
return true;
}
for (; Stop > Section + 2 && (Stop[-2] == '\n' || Stop[-2] == '\r'); Stop--);
}
/*}}}*/
-// TagSection::Exists - return True if a tag exists /*{{{*/
+// TagSection::Exists - return True if a tag exists /*{{{*/
+#if APT_PKG_ABI >= 413
+bool pkgTagSection::Exists(const char* const Tag) const
+#else
bool pkgTagSection::Exists(const char* const Tag)
+#endif
{
unsigned int tmp;
return Find(Tag, tmp);
/* This searches the section for a tag that matches the given string. */
bool pkgTagSection::Find(const char *Tag,unsigned int &Pos) const
{
- unsigned int Length = strlen(Tag);
- unsigned int I = AlphaIndexes[AlphaHash(Tag)];
- if (I == 0)
+ size_t const Length = strlen(Tag);
+ unsigned int Bucket = AlphaIndexes[AlphaHash(Tag, Length)];
+ if (Bucket == 0)
return false;
- I--;
-
- for (unsigned int Counter = 0; Counter != TagCount; Counter++,
- I = (I+1)%TagCount)
+
+ for (; Bucket != 0; Bucket = d->Tags[Bucket - 1].NextInBucket)
{
- const char *St;
- St = Section + Indexes[I];
- if (strncasecmp(Tag,St,Length) != 0)
+ if ((d->Tags[Bucket - 1].EndTag - d->Tags[Bucket - 1].StartTag) != Length)
continue;
- // Make sure the colon is in the right place
- const char *C = St + Length;
- for (; isspace(*C) != 0; C++);
- if (*C != ':')
+ char const * const St = Section + d->Tags[Bucket - 1].StartTag;
+ if (strncasecmp(Tag,St,Length) != 0)
continue;
- Pos = I;
+
+ Pos = Bucket - 1;
return true;
}
Pos = 0;
return false;
}
- /*}}}*/
-// TagSection::Find - Locate a tag /*{{{*/
-// ---------------------------------------------------------------------
-/* This searches the section for a tag that matches the given string. */
bool pkgTagSection::Find(const char *Tag,const char *&Start,
const char *&End) const
{
- unsigned int Length = strlen(Tag);
- unsigned int I = AlphaIndexes[AlphaHash(Tag)];
- if (I == 0)
+ unsigned int Pos;
+ if (Find(Tag, Pos) == false)
return false;
- I--;
-
- for (unsigned int Counter = 0; Counter != TagCount; Counter++,
- I = (I+1)%TagCount)
- {
- const char *St;
- St = Section + Indexes[I];
- if (strncasecmp(Tag,St,Length) != 0)
- continue;
-
- // Make sure the colon is in the right place
- const char *C = St + Length;
- for (; isspace(*C) != 0; C++);
- if (*C != ':')
- continue;
- // Strip off the gunk from the start end
- Start = C;
- End = Section + Indexes[I+1];
- if (Start >= End)
- return _error->Error("Internal parsing error");
-
- for (; (isspace(*Start) != 0 || *Start == ':') && Start < End; Start++);
- for (; isspace(End[-1]) != 0 && End > Start; End--);
-
- return true;
- }
-
- Start = End = 0;
- return false;
+ Start = Section + d->Tags[Pos].StartValue;
+ // Strip off the gunk from the end
+ End = Section + d->Tags[Pos + 1].StartTag;
+ if (unlikely(Start > End))
+ return _error->Error("Internal parsing error");
+
+ for (; isspace(End[-1]) != 0 && End > Start; --End);
+
+ return true;
}
/*}}}*/
// TagSection::FindS - Find a string /*{{{*/
return Result;
}
/*}}}*/
+// TagSection::FindB - Find boolean value /*{{{*/
+// ---------------------------------------------------------------------
+/* */
+bool pkgTagSection::FindB(const char *Tag, bool const &Default) const
+{
+ const char *Start, *Stop;
+ if (Find(Tag, Start, Stop) == false)
+ return Default;
+ return StringToBool(string(Start, Stop));
+}
+ /*}}}*/
// TagSection::FindFlag - Locate a yes/no type flag /*{{{*/
// ---------------------------------------------------------------------
/* The bits marked in Flag are masked on/off in Flags */
return true;
return FindFlag(Flags, Flag, Start, Stop);
}
-bool const pkgTagSection::FindFlag(unsigned long &Flags, unsigned long Flag,
+bool pkgTagSection::FindFlag(unsigned long &Flags, unsigned long Flag,
char const* Start, char const* Stop)
{
switch (StringToBool(string(Start, Stop)))
return true;
}
/*}}}*/
+void pkgTagSection::Get(const char *&Start,const char *&Stop,unsigned int I) const
+{
+ Start = Section + d->Tags[I].StartTag;
+ Stop = Section + d->Tags[I+1].StartTag;
+}
+APT_PURE unsigned int pkgTagSection::Count() const { /*{{{*/
+ if (d->Tags.empty() == true)
+ return 0;
+ // the last element is just marking the end and isn't a real one
+ return d->Tags.size() - 1;
+}
+ /*}}}*/
+
+#include "tagfile-order.c"
+
// TFRewrite - Rewrite a control record /*{{{*/
// ---------------------------------------------------------------------
/* This writes the control record to stdout rewriting it as necessary. The
override map item specificies the rewriting rules to follow. This also
takes the time to sort the feild list. */
-
-/* The order of this list is taken from dpkg source lib/parse.c the fieldinfos
- array. */
-static const char *iTFRewritePackageOrder[] = {
- "Package",
- "Essential",
- "Status",
- "Priority",
- "Section",
- "Installed-Size",
- "Maintainer",
- "Original-Maintainer",
- "Architecture",
- "Source",
- "Version",
- "Revision", // Obsolete
- "Config-Version", // Obsolete
- "Replaces",
- "Provides",
- "Depends",
- "Pre-Depends",
- "Recommends",
- "Suggests",
- "Conflicts",
- "Breaks",
- "Conffiles",
- "Filename",
- "Size",
- "MD5Sum",
- "SHA1",
- "SHA256",
- "SHA512",
- "MSDOS-Filename", // Obsolete
- "Description",
- 0};
-static const char *iTFRewriteSourceOrder[] = {"Package",
- "Source",
- "Binary",
- "Version",
- "Priority",
- "Section",
- "Maintainer",
- "Original-Maintainer",
- "Build-Depends",
- "Build-Depends-Indep",
- "Build-Conflicts",
- "Build-Conflicts-Indep",
- "Architecture",
- "Standards-Version",
- "Format",
- "Directory",
- "Files",
- 0};
-
-/* Two levels of initialization are used because gcc will set the symbol
- size of an array to the length of the array, causing dynamic relinking
- errors. Doing this makes the symbol size constant */
-const char **TFRewritePackageOrder = iTFRewritePackageOrder;
-const char **TFRewriteSourceOrder = iTFRewriteSourceOrder;
-
bool TFRewrite(FILE *Output,pkgTagSection const &Tags,const char *Order[],
TFRewriteData *Rewrite)
{
return true;
}
/*}}}*/
+
+pkgTagSection::~pkgTagSection() { delete d; }