Bug#869557: apt: please make the output of apt-ftparchive reproducible

[Chris Lamb <lamby@debian.org>]

tags 869557 + patch
thanks

Hi,

> apt: please make the output of apt-ftparchive reproducible

*Extremely* WIP patch attached:

  commit 489251d0bcf32883e9028aa0278168c167f2f6e9
  Author: Chris Lamb <lamby@debian.org>
  Date:   Mon Jul 24 15:04:03 2017 +0100
  
      Make the output of ftp-archive reproducible by sorting non-deterministic file orders prior to output. (Closes: #869557)
  
   ftparchive/writer.cc | 19 ++++++++++++++++++-
   ftparchive/writer.h  |  3 +++
   2 files changed, 21 insertions(+), 1 deletion(-)


Regards,

-- 
      ,''`.
     : :'  :     Chris Lamb, Debian Project Leader
     `. `'`      lamby@debian.org / chris-lamb.co.uk
       `-

>From 489251d0bcf32883e9028aa0278168c167f2f6e9 Mon Sep 17 00:00:00 2001
From: Chris Lamb <lamby@debian.org>
Date: Mon, 24 Jul 2017 15:04:03 +0100
Subject: [PATCH] Make the output of ftp-archive reproducible by sorting
 non-deterministic file orders prior to output. (Closes: #869557)

---
 ftparchive/writer.cc | 19 ++++++++++++++++++-
 ftparchive/writer.h  |  3 +++
 2 files changed, 21 insertions(+), 1 deletion(-)

diff --git a/ftparchive/writer.cc b/ftparchive/writer.cc
index 5773f3aec..9c081c567 100644
--- a/ftparchive/writer.cc
+++ b/ftparchive/writer.cc
@@ -124,7 +124,6 @@ int FTWScanner::ScannerFTW(const char *File,const struct stat * /*sb*/,int Flag)
 int FTWScanner::ScannerFile(const char *File, bool const &ReadLink)
 {
    const char *LastComponent = strrchr(File, '/');
-   char *RealPath = NULL;
 
    if (LastComponent == NULL)
       LastComponent = File;
@@ -140,10 +139,18 @@ int FTWScanner::ScannerFile(const char *File, bool const &ReadLink)
    if (I == Owner->Patterns.end())
       return 0;
 
+   Owner->FilesToProcess.push_back(std::pair<string, bool>(File, ReadLink));
+
+   return 0;
+}
+
+int FTWScanner::ProcessFile(const char *File, bool const &ReadLink)
+{
    /* Process it. If the file is a link then resolve it into an absolute
       name.. This works best if the directory components the scanner are
       given are not links themselves. */
    char Jnk[2];
+   char *RealPath = NULL;
    Owner->OriginalPath = File;
    if (ReadLink &&
        readlink(File,Jnk,sizeof(Jnk)) != -1 &&
@@ -209,6 +216,16 @@ bool FTWScanner::RecursiveScan(string const &Dir)
 	 _error->Errno("ftw",_("Tree walking failed"));
       return false;
    }
+
+   sort(Owner->FilesToProcess.begin(), Owner->FilesToProcess.end(), [](const pair<string, bool>& a, const pair<string, bool>& b) {
+      return a.first.compare(b.first);
+   });
+
+   for (const pair<string, bool> &it : Owner->FilesToProcess)
+   {
+      if (Owner->ProcessFile(it.first.c_str(), it.second) != 0)
+         return false;
+   }
    
    return true;
 }
diff --git a/ftparchive/writer.h b/ftparchive/writer.h
index ea4c66da4..4cb775515 100644
--- a/ftparchive/writer.h
+++ b/ftparchive/writer.h
@@ -23,6 +23,7 @@
 #include <set>
 #include <stdlib.h>
 #include <sys/types.h>
+#include <utility>
 
 #include "contents.h"
 #include "cachedb.h"
@@ -39,6 +40,7 @@ class FTWScanner
 {
    protected:
    vector<string> Patterns;
+   vector<std::pair<string,bool> > FilesToProcess;
    string Arch;
    bool IncludeArchAll;
    const char *OriginalPath;
@@ -50,6 +52,7 @@ class FTWScanner
    static FTWScanner *Owner;
    static int ScannerFTW(const char *File,const struct stat *sb,int Flag);
    static int ScannerFile(const char *File, bool const &ReadLink);
+   static int ProcessFile(const char *File, bool const &ReadLink);
 
    bool Delink(string &FileName,const char *OriginalPath,
 	       unsigned long long &Bytes,unsigned long long const &FileSize);
-- 
2.13.3