[PATCH] fix scalability problems with git-deltafy-script

Current version would spin forever and exhaust memory while attempting
to sort all files from all revisions at once, until it dies before even
doing any real work.  This is especially noticeable when used on a big
repository like the imported bkcvs repo for the Linux kernel.

This patch allows for batching the sort to put a bound on needed
resources and making progress early, as well as including some small
cleanups.

Signed-off-by: Nicolas Pitre <nico@cam.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

diff --git a/git-deltafy-script b/git-deltafy-script
index 21a95692ff3270c35a230c3046dee018109aff79..476d8796ecbb60688a91888f311ac6074eead45b 100755 (executable)
--- a/git-deltafy-script
+++ b/git-deltafy-script
@@ -1,6 +1,6 @@
 #!/bin/bash
 
 #!/bin/bash
 

-# Example script to deltafy an entire GIT repository based on the commit list.
+# Example script to deltify an entire GIT repository based on the commit list.

 # The most recent version of a file is the reference and previous versions
 # are made delta against the best earlier version available. And so on for
 # successive versions going back in time.  This way the increasing delta
 # The most recent version of a file is the reference and previous versions
 # are made delta against the best earlier version available. And so on for
 # successive versions going back in time.  This way the increasing delta


@@ -25,37 +25,51 @@
 
 set -e
 
 
 set -e
 

-depth=
-[ "$1" == "-d" ] && depth="--max-depth=$2" && shift 2
+max_depth=
+[ "$1" == "-d" ] && max_depth="--max-depth=$2" && shift 2
+
+overlap=30
+max_behind="--max-behind=$overlap"

 
 function process_list() {
        if [ "$list" ]; then
                echo "Processing $curr_file"
 
 function process_list() {
        if [ "$list" ]; then
                echo "Processing $curr_file"

-               echo "$head $list" | xargs git-mkdelta $depth --max-behind=30 -v
+               echo "$list" | xargs git-mkdelta $max_depth $max_behind -v

        fi
 }
 
        fi
 }
 

+rev_list=""

 curr_file=""
 
 git-rev-list HEAD |
 curr_file=""
 
 git-rev-list HEAD |

-git-diff-tree -r -t --stdin |
-awk '/^:/ { if ($5 == "M" || $5 == "N") print $4, $6;
-            if ($5 == "M") print $3, $6 }' |
-LC_ALL=C sort -s -k 2 | uniq |
-while read sha1 file; do
-       if [ "$file" == "$curr_file" ]; then
-               list="$list $sha1"
-       else
-               process_list
-               curr_file="$file"
-               list=""
-               head="$sha1"
-       fi
+while true; do
+       # Let's batch revisions into groups of 1000 to give it a chance to
+       # scale with repositories containing long revision lists.  We also
+       # overlap with the previous batch the size of mkdelta's look behind
+       # value in order to account for the processing discontinuity.
+       rev_list="$(echo -e -n "$rev_list" | tail --lines=$overlap)"
+       for i in $(seq 1000); do
+               read rev || break
+               rev_list="$rev_list$rev\n"
+       done
+       echo -e -n "$rev_list" |
+       git-diff-tree -r -t --stdin |
+       awk '/^:/ { if ($5 == "M") printf "%s %s\n%s %s\n", $4, $6, $3, $6 }' |
+       LC_ALL=C sort -s -k 2 | uniq |
+       while read sha1 file; do
+               if [ "$file" == "$curr_file" ]; then
+                       list="$list $sha1"
+               else
+                       process_list
+                       curr_file="$file"
+                       list="$sha1"
+               fi
+       done
+       [ "$rev" ] || break

 done
 process_list
 
 curr_file="root directory"
 done
 process_list
 
 curr_file="root directory"

-head=""

 list="$(
        git-rev-list HEAD |
        while read commit; do
 list="$(
        git-rev-list HEAD |
        while read commit; do