Add support for writing by-hash dirs in apt-ftparchive

This option is enabled via the APT::FTPArchive::DoByHash switch.
It will also honor the option APT::FTPArchive::By-Hash-Keep that
controls how many previous generation of by-hash files should be
kept (defaults to 3).

Merged from https://github.com/mvo5/apt/tree/feature/apt-ftparchive-by-hash

5 files changed, 219 insertions, 2 deletions

diff --git a/ftparchive/byhash.cc b/ftparchive/byhash.cc
new file mode 100644
index 000000000..04f8f1629
--- /dev/null
+++ b/ftparchive/byhash.cc
@@ -0,0 +1,63 @@
+// -*- mode: cpp; mode: fold -*-
+// Description								/*{{{*/
+/* ######################################################################
+
+   ByHash 
+   
+   ByHash helper functions
+   
+   ##################################################################### */
+									/*}}}*/
+// Include Files							/*{{{*/
+#include <config.h>
+
+#include<algorithm>
+#include<string>
+
+#include <unistd.h>
+#include <sys/stat.h>
+
+#include <apt-pkg/fileutl.h>
+#include <apt-pkg/hashes.h>
+#include "byhash.h"
+
+// Delete all files in a directory except the most recent N ones
+void DeleteAllButMostRecent(std::string dir, int KeepFiles)
+{
+   struct Cmp {
+      bool operator() (const std::string& lhs, const std::string& rhs) {
+         struct stat buf_l, buf_r;
+         stat(lhs.c_str(), &buf_l);
+         stat(rhs.c_str(), &buf_r);
+         if (buf_l.st_mtim.tv_sec == buf_r.st_mtim.tv_sec)
+            return buf_l.st_mtim.tv_nsec < buf_r.st_mtim.tv_nsec;
+         return buf_l.st_mtim.tv_sec < buf_r.st_mtim.tv_sec;
+      }
+   };
+
+   if (!DirectoryExists(dir))
+      return;
+
+   auto files = GetListOfFilesInDir(dir, false);
+   std::sort(files.begin(), files.end(), Cmp());
+
+   for (auto I=files.begin(); I<files.end()-KeepFiles; I++) {
+      unlink((*I).c_str());
+   }
+}
+
+// Takes a input filename (e.g. binary-i386/Packages) and a hashstring
+// of the Input data and transforms it into a suitable by-hash filename
+std::string GenByHashFilename(std::string Input, HashString h)
+{
+   std::string ByHashOutputFile = Input;
+   std::string const ByHash = "/by-hash/" + h.HashType() + "/" + h.HashValue();
+   size_t trailing_slash = ByHashOutputFile.find_last_of("/");
+   if (trailing_slash == std::string::npos)
+      trailing_slash = 0;
+   ByHashOutputFile = ByHashOutputFile.replace(
+      trailing_slash,
+      ByHashOutputFile.substr(trailing_slash+1).size()+1,
+      ByHash);
+   return ByHashOutputFile;
+}
diff --git a/ftparchive/byhash.h b/ftparchive/byhash.h
new file mode 100644
index 000000000..ce05397ad
--- /dev/null
+++ b/ftparchive/byhash.h
@@ -0,0 +1,23 @@
+// -*- mode: cpp; mode: fold -*-
+// Description								/*{{{*/
+/* ######################################################################
+
+   ByHash 
+   
+   ByHash helper functions
+   
+   ##################################################################### */
+									/*}}}*/
+#ifndef BYHASH_H
+#define BYHASH_H
+
+class HashString;
+
+// Delete all files in "dir" except for the number specified in "KeepFiles"
+// that are the most recent ones
+void DeleteAllButMostRecent(std::string dir, int KeepFiles);
+
+// takes a regular input filename
+std::string GenByHashFilename(std::string Input, HashString h);
+
+#endif
diff --git a/ftparchive/makefile b/ftparchive/makefile
index e67272e1e..c80487c3f 100644
--- a/ftparchive/makefile
+++ b/ftparchive/makefile
@@ -12,7 +12,7 @@ PROGRAM=apt-ftparchive
 SLIBS = -lapt-pkg -lapt-inst -lapt-private $(BDBLIB) $(INTLLIBS)
 LIB_MAKES = apt-pkg/makefile apt-inst/makefile apt-private/makefile
 SOURCE = apt-ftparchive.cc cachedb.cc writer.cc contents.cc override.cc \
-         multicompress.cc sources.cc
+         multicompress.cc sources.cc byhash.cc
 include $(PROGRAM_H)
 else
 PROGRAM=apt-ftparchive
diff --git a/ftparchive/writer.cc b/ftparchive/writer.cc
index 7f09a3758..82049836a 100644
--- a/ftparchive/writer.cc
+++ b/ftparchive/writer.cc
@@ -40,11 +40,13 @@
 #include <sstream>
 #include <memory>
 #include <utility>
+#include <algorithm>
 
 #include "apt-ftparchive.h"
 #include "writer.h"
 #include "cachedb.h"
 #include "multicompress.h"
+#include "byhash.h"
 
 #include <apti18n.h>
 									/*}}}*/
@@ -1018,7 +1020,9 @@ ReleaseWriter::ReleaseWriter(FileFd * const GivenOutput, string const &/*DB*/) :
    Fields["Architectures"] = "";
    Fields["Components"] = "";
    Fields["Description"] = "";
-
+   if (_config->FindB("APT::FTPArchive::DoByHash", true) == true)
+      Fields["Acquire-By-Hash"] = "true";
+   
    for(map<string,string>::const_iterator I = Fields.begin();
        I != Fields.end();
        ++I)
@@ -1070,6 +1074,31 @@ bool ReleaseWriter::DoPackage(string FileName)
    CheckSums[NewFileName].Hashes = hs.GetHashStringList();
    fd.Close();
 
+   // FIXME: wrong layer in the code(?)
+   // FIXME2: symlink instead of create a copy
+   if (_config->FindB("APT::FTPArchive::DoByHash", true) == true)
+   {
+      std::string Input = FileName;
+      HashStringList hsl = hs.GetHashStringList();
+      for(HashStringList::const_iterator h = hsl.begin();
+          h != hsl.end(); ++h)
+      {
+         if (!h->usable())
+            continue;
+         std::string ByHashOutputFile = GenByHashFilename(Input, *h);
+
+         std::string ByHashOutputDir = flNotFile(ByHashOutputFile);
+         if(!CreateDirectory(flNotFile(Input), ByHashOutputDir))
+            return _error->Warning("can not create dir %s", flNotFile(ByHashOutputFile).c_str());
+
+         // write new hashes
+         FileFd In(Input, FileFd::ReadOnly);
+         FileFd Out(ByHashOutputFile, FileFd::WriteEmpty);
+         if(!CopyFile(In, Out))
+            return _error->Warning("failed to copy %s %s", Input.c_str(), ByHashOutputFile.c_str());
+      }
+   }
+
    return true;
 }
 
@@ -1107,4 +1136,40 @@ void ReleaseWriter::Finish()
       printChecksumTypeRecord(*Output, "SHA256", CheckSums);
    if ((DoHashes & Hashes::SHA512SUM) == Hashes::SHA512SUM)
       printChecksumTypeRecord(*Output, "SHA512", CheckSums);
+
+   // go by-hash cleanup
+   map<string,ReleaseWriter::CheckSum>::const_iterator prev = CheckSums.begin();
+   if (_config->FindB("APT::FTPArchive::DoByHash", true) == true)
+   {
+      for(map<string,ReleaseWriter::CheckSum>::const_iterator I = CheckSums.begin();
+	 I != CheckSums.end(); ++I)
+      {
+         if (I->first == "Release" || I->first == "InRelease")
+            continue;
+
+         // keep iterating until we find a new subdir
+         if(flNotFile(I->first) == flNotFile(prev->first))
+            continue;
+
+         // clean that subdir up
+         int keepFiles = _config->FindI("APT::FTPArchive::By-Hash-Keep", 3);
+         // calculate how many compressors are used (the amount of files
+         // in that subdir generated for this run)
+         keepFiles *= std::distance(prev, I);
+         prev = I;
+
+         HashStringList hsl = prev->second.Hashes;
+         for(HashStringList::const_iterator h = hsl.begin();
+             h != hsl.end(); ++h)
+         {
+
+            if (!h->usable())
+               continue;
+
+            std::string RealFilename = DirStrip+"/"+prev->first;
+            std::string ByHashOutputFile = GenByHashFilename(RealFilename, *h);
+            DeleteAllButMostRecent(flNotFile(ByHashOutputFile), keepFiles);
+         }
+      }
+   }
 }
diff --git a/test/integration/test-apt-ftparchive-by-hash b/test/integration/test-apt-ftparchive-by-hash
new file mode 100755
index 000000000..6cda0e415
--- /dev/null
+++ b/test/integration/test-apt-ftparchive-by-hash
@@ -0,0 +1,66 @@
+#!/bin/sh
+set -e
+
+verify_by_hash() {
+    for hash_gen in SHA1:sha1sum SHA256:sha256sum SHA512:sha512sum; do
+        hash=$(echo ${hash_gen} | cut -f1 -d:)
+        gen=$(echo ${hash_gen} | cut -f2 -d:)
+        testsuccess stat aptarchive/dists/unstable/main/binary-i386/by-hash/$hash/$($gen aptarchive/dists/unstable/main/binary-i386/Packages | cut -f1 -d' ')
+        testsuccess stat aptarchive/dists/unstable/main/binary-i386/by-hash/$hash/$($gen aptarchive/dists/unstable/main/binary-i386/Packages.gz | cut -f1 -d' ')
+    done
+}
+
+#
+# main()
+#
+TESTDIR=$(readlink -f $(dirname $0))
+. $TESTDIR/framework
+setupenvironment
+configarchitecture 'i386'
+configcompression 'gz' '.'
+
+# build one pacakge
+buildsimplenativepackage 'foo' 'i386' '1' 'unstable'
+buildaptarchivefromincoming
+
+# verify initial run
+verify_by_hash
+previous_hash=$(sha256sum aptarchive/dists/unstable/main/binary-i386/Packages | cut -f1 -d' ')
+
+# insert new package
+buildsimplenativepackage 'bar' 'i386' '1' 'unstable'
+# and build again
+buildaptarchivefromincoming
+
+# ensure the new package packag is there
+testsuccess zgrep "Package: bar" aptarchive/dists/unstable/main/binary-i386/Packages.gz
+
+# ensure we have the by-hash stuff
+verify_by_hash
+
+# ensure the old hash is still there
+testsuccess stat aptarchive/dists/unstable/main/binary-i386/by-hash/SHA256/$previous_hash
+
+# ensure we have it in the Release file
+testsuccess grep "Acquire-By-Hash: true" aptarchive/dists/unstable/*Release
+
+# now ensure gc work
+for i in $(seq 3); do
+    buildsimplenativepackage 'bar' 'i386' "$i" 'unstable'
+    buildaptarchivefromincoming
+done
+
+hash_count=$(ls aptarchive/dists/unstable/main/binary-i386/by-hash/SHA256/|wc -l)
+# we have 2 files (uncompressed, gz) per run, 5 runs in total
+# by default apt-ftparchive keeps three generations (current plus 2 older)
+msgtest "Check that gc for by-hash works… "
+if [ "$hash_count" = "6" ]; then
+    msgpass
+else
+    echo "Got $hash_count expected 6"
+    msgfail
+fi
+
+# ensure the current generation is still there
+verify_by_hash
+