initial commit

bugs/issue-8f36cd6c86489cdaa7349beb2277e947a5d0558a.yaml

@@ -0,0 +1,53 @@
+--- !ditz.rubyforge.org,2008-03-06/issue 
+title: Reduce memory expense of multi-alignment (e.g., by discarding maps between passes, by reducing typical kept size, or by reducing map element size)
+desc: ""
+type: :task
+component: ale
+release: 
+reporter: David Hilvert <dhilvert@auricle.dyndns.org>
+status: :unstarted
+disposition: 
+creation_time: 2009-03-28 22:53:12.231388 Z
+references: []
+
+id: 8f36cd6c86489cdaa7349beb2277e947a5d0558a
+log_events: 
+- - 2009-03-28 22:53:12.832156 Z
+  - David Hilvert <dhilvert@auricle.dyndns.org>
+  - created
+  - ""
+- - 2009-03-28 22:53:46.644669 Z
+  - David Hilvert <dhilvert@auricle.dyndns.org>
+  - commented
+  - One approach that may be worth investigating is compression (e.g., by PNG).
+- - 2009-03-28 22:56:07.802488 Z
+  - David Hilvert <dhilvert@auricle.dyndns.org>
+  - commented
+  - |-
+    Related to assertion failiure bug reported by Jari Juslin in message
+    [ale] Processing color negative scans with ALE (was: Help with aligning images)
+- - 2009-03-29 08:22:52.909320 Z
+  - David Hilvert <dhilvert@auricle.dyndns.org>
+  - commented
+  - |-
+    Patch sent to list, adapted for the current branch:
+    
+    --- a/d2/trans_multi.h
+    +++ b/d2/trans_multi.h
+    @@ -55,7 +55,7 @@ private:
+    static ale_pos _multi_decomp;
+    static ale_real _multi_improvement;
+    
+    -       typedef unsigned int index_t;
+    +       typedef unsigned short index_t;
+    
+    std::vector<trans_single> trans_stack;
+    std::vector<multi_coordinate> coord_stack;
+- - 2009-03-29 08:34:03.662057 Z
+  - David Hilvert <dhilvert@auricle.dyndns.org>
+  - commented
+  - |-
+    One approach to solving this would be restricting the candidate set size
+    for each pixel in the map, so that, e.g., only regions neighboring a local
+    region (allowing any level of decomposition) can be mapped.
+git_branch: 

bugs/issue-d279fa0d83212136ed7fd67e1ea69ef663707282.yaml

@@ -0,0 +1,24 @@
+--- !ditz.rubyforge.org,2008-03-06/issue 
+title: Remove superfluous header 'Overview' from introductory sections of asciidoc documentation.
+desc: ""
+type: :task
+component: ale
+release: 
+reporter: David Hilvert <dhilvert@auricle.dyndns.org>
+status: :closed
+disposition: :fixed
+creation_time: 2009-03-28 08:37:16.047379 Z
+references: []
+
+id: d279fa0d83212136ed7fd67e1ea69ef663707282
+log_events: 
+- - 2009-03-28 08:37:16.512711 Z
+  - David Hilvert <dhilvert@auricle.dyndns.org>
+  - created
+  - ""
+- - 2009-03-28 09:02:48.959063 Z
+  - David Hilvert <dhilvert@auricle.dyndns.org>
+  - closed with disposition fixed
+  - |-
+    'Overview' section should probably also be removed from the user guide, but this isn't asciidoc,
+    so closing.

bugs/issue-ec7a9aea4ad735dc6d5e00fe93335cc4b93f7801.yaml

@@ -0,0 +1,38 @@
+--- !ditz.rubyforge.org,2008-03-06/issue 
+title: Serious memory leak affecting all current versions.
+desc: |-
+  Jari Juslin reports in message '[ale] Processing color negative scans with ALE'
+  an apparent out-of-memory condition affecting 0.9 branch with --ips 0.  Testing
+  with 0.8.11.1 indicates that the 0.8 branch is affected also.  In particular,
+  the following commands result in memory increase without bounds (e.g., visible
+  in top or atop) throughout execution, which is unexpected for this
+  configuration:
+  
+  $ convert -size 1000x1000 xc:gray gray.png
+  $ ale-bin --cache 0 --no-inc --dchain zero --achain zero --perturb-upper=0 --ips 0 `for i in \`seq 1 1000\`; do echo -n "gray.png "; done` null:
+  
+  (Ignore the segfault following execution; this is a consequence of the 'zero'
+  filter.)
+  
+  Running 'valgrind' on an input set of smaller dimensions and frame count reveals
+  a number of memory leaks (reachable and unreachable) which have now been fixed.
+type: :bugfix
+component: ale
+release: 
+reporter: David Hilvert <dhilvert@auricle.dyndns.org>
+status: :closed
+disposition: :fixed
+creation_time: 2009-03-30 18:46:16.735784 Z
+references: []
+
+id: ec7a9aea4ad735dc6d5e00fe93335cc4b93f7801
+log_events: 
+- - 2009-03-30 18:46:22.694825 Z
+  - David Hilvert <dhilvert@auricle.dyndns.org>
+  - created
+  - ""
+- - 2009-03-30 18:47:15.401088 Z
+  - David Hilvert <dhilvert@auricle.dyndns.org>
+  - closed with disposition fixed
+  - ""
+git_branch: 

bugs/issue-fe69b6ceefef3805e29190e1040f9278be84b1f8.yaml

@@ -0,0 +1,23 @@
+--- !ditz.rubyforge.org,2008-03-06/issue 
+title: Investigate recent ImageMagick build procedures, and consider updating build accordingly.
+desc: ""
+type: :task
+component: ale
+release: 
+reporter: David Hilvert <dhilvert@auricle.dyndns.org>
+status: :closed
+disposition: :fixed
+creation_time: 2009-03-30 19:52:00.861315 Z
+references: []
+
+id: fe69b6ceefef3805e29190e1040f9278be84b1f8
+log_events: 
+- - 2009-03-30 19:52:02.141451 Z
+  - David Hilvert <dhilvert@auricle.dyndns.org>
+  - created
+  - ""
+- - 2009-03-31 12:23:33.521227 Z
+  - David Hilvert <dhilvert@auricle.dyndns.org>
+  - closed with disposition fixed
+  - Current configure seems to be OK for ImageMagick 6.4.5 2009-01-22 Q16 (shipped with Ubuntu).
+git_branch: 

bugs/project.yaml

@@ -0,0 +1,8 @@
+--- !ditz.rubyforge.org,2008-03-06/project 
+name: ale
+version: "0.5"
+components: 
+- !ditz.rubyforge.org,2008-03-06/component 
+  name: ale
+releases: []
+