rpattabi / diskpacker Goto Github PK

Infrarecorder is unable to open generated irp project files.

I'm using infrarecorder 0.44.1.0 unicode, x86 version.

Incremental backup support needed.

One possible way to implement this:
1. Backup files first time. Note down the date of backup.
2. Next time, give the date of previous backup (for that particular backup set)
3. diskpacker finds files which are created/modified after that specified
date&time, and backs them up.

As of now (Rivision 22), output files are created at source code directory
(lib). 

Output files should go to a different folder, perhaps a subfolder or
configured by the user.

In Infrarecorder project: 
Whatever 'file' goes into the root directory of a DVD, it comes twice. In
other words, Files in dvd root directory have 2 entries each. 

It does not happen to files in sub folders.

bin_packed txt files are normal. 

Duplicated files inflate the Disk size. if they are removed, project size
comes to estimated size.

Issue exists in irp compilations. other compilations (k3b, brasero) are not
scanned for this problem yet.

Each backup set should have provision for an Exclude List too. 

Files/Folders included in the exclude list will be skipped by DiskPacker.

In MBs now.

This is odd, but I could get files with invalid names e.g.

    file name ---> "01 I�Ve something.wma (invalid encoding)"

Right or wrong I have a file named like this. If I include the
corresponding folder for backup, infra recorder project generator fails.

This can be considered as rare possibility.

It would be useful if disk packer identifies duplicates in the backup set.
There are two possibilities:

1. User might give the same input path twice
2. In different input paths we might have same file or folders

Diskpacker should by default backup only one set and report the details of
the duplicates.

Criteria to identify the duplicates: Files and folders with same name,
size, and other suitable attributes. Checksum could also be used.

Note: There can be a future requirement to identify duplicate photos, music
files, etc which may not have the same file name (e.g. iTunes, and iPod
files have different file names). In these cases checksums, and tag
comparison (in general embedded metadata) could be used.

DiskPacker is currently command line only utility. A GUI would help. If the
system finds out the installed burning softwares it would be great.

For some reason utf-8 to utf-16 conversion after the generated
infrarecorder project does not work in windows. The project file looks okay
in linux, but it looks garbled in windows.

1. The code is not organized well. Everything is in a single file.
2. The design shall be re looked at. The type of bin (CD/DVD/etc), and the
type of compilation output shall be easily pluggable and selectable.
3. Tests???

DiskPacker skips files/folders larger than a single disk. But it doesn't
report about the skipped files/folders.

Batch files do not remove all the files they are supposed to clean up. For
instance, if 10 files go to Disk1 (infra project &/ bins_packed.txt), not
all the 10 files are removed by Disk1.bat. There are omissions.

DiskPacker does not work with files with unicode names.

I'm trying to backup my photos. DiskPacker throwing an error and stops when
 it encounters a filename/folder with unicode characters. I experience this
problem in my windows xp machine, with Tamil and Korean filenames. 

Here's the output:

D:\diskpacker-read-only\lib>ruby main.rb
main.rb:32:in `size': No such file or directory -
f:/photos/august_2007/????006.
jpg (Errno::ENOENT)
        from main.rb:32:in `element_generator'
        from main.rb:11:in `dir_tree_walker'
        from main.rb:16:in `dir_tree_walker'
        from E:/Program Files/ruby/lib/ruby/1.8/find.rb:39:in `find'
        from E:/Program Files/ruby/lib/ruby/1.8/find.rb:38:in `catch'
        from E:/Program Files/ruby/lib/ruby/1.8/find.rb:38:in `find'
        from main.rb:14:in `dir_tree_walker'
        from main.rb:16:in `dir_tree_walker'
        from E:/Program Files/ruby/lib/ruby/1.8/find.rb:39:in `find'
        from E:/Program Files/ruby/lib/ruby/1.8/find.rb:38:in `catch'
        from E:/Program Files/ruby/lib/ruby/1.8/find.rb:38:in `find'
        from main.rb:14:in `dir_tree_walker'
        from main.rb:28:in `element_generator'
        from main.rb:27:in `each'
        from main.rb:27:in `element_generator'
        from main.rb:54:in `generate_elements'
        from main.rb:87

Diskpacker handles big files fine. But when we have very large number of
files (tens of thousands of files e.g. digital photos, mp3s, etc), the
performance is not good.

Currently diskpacker accepts no command line arguments. A neat command line
interface needs to be developed.

Options are

--backup-set <backup_set_name> # formerly input_paths.txt
--disk-type <CD/DVD/DVD9/custom> #cd,dvd,cd650,dvd9,etc.
--disk-capacity <inMB> #optional. not optional for custom disk type. disk
capacity can override disk type as appropriate.
--output-path <output_folder_name> #if the folder is not present,
diskpacker would attempt to create the folder
--backup-title <title=BACKUP> #DivX --> DivX_0.irp
--report <report_name> #packed --> packed.txt in output path
--burner <infrarecorder/brasero/k3b/ALL> #if you want combinations e.g.
only k3b and brasero project files, you can specify --burner option
multiple times.
--dir-to-backup <folder_to_backup> # this is an alternative to --backup-set
if you want to backup one folder (or more if you add this option multiple
times)

Reasonable defaults shall be assumed if the options are not given.

Files belonging to each disk in bins_packed.txt are not sorted in any
particular order. If they are sorted folderwise or so, it'll be easy to see
how many files from a drive/folder go to a particular disk compilation.
It'll also be very useful when creating a cd/dvd compilation manually.

1. bins_packed.txt has forward slashes (/) in paths. Not usable in Windows
2. generates DOS batch files to clean up files. Similarly it should be able
to  create shell scripts for linux.

This is the suggestion from Ananth. DiskPacker could be much more useful if
it also catalogs the disks. There could be two possible scenarios.

1. Before the disk is burned, once the disk burning projects are generated,
there could be an option add the disk to the catalog.
2. Once a disk with data is inserted, it could be added to catalog.

DiskPacker shall index the cataloged disks, and it should be easy to search
for a particular file with regular expression support.

There is a k3b project generator. But it is not tested or not working.