New : in and out saved in param file and restored , verified only newly copied file , other restiling
This is a simple program to copy files or directories and verify the copied files with md5 hash.
How to use :
The first startup is longer because it unpacks the package in a temp dir . From the second run the startup is speeder because it reuses the same temp dir.
Primary window :
in the first line the are various objects:
1) a box for the name of the file where will be written md5 hash when copying a dir, or the extension added to the file name when copying a single file.
2) the indicator to decide if you want to replace existing files with older modification date.
3) the indicator to decide if you want to copy even the initial selected directory.
4) the dimension of the buffer in read and write operation in Megabyte (1 = 1MegaByte of buffer).
5) the indicator to verify only newly copied files of the md5sum.md5 or all the files in the md5sum.md5
6) a box where you select wich file_select_box to use in input and output :
MultiDir = (default) you can select multiple dirs or multiple files with a custom box
multiple = you can select one dir or more than one file with a generic box
Single = you can select one file with an os_native box(now multiple?)
Dir = you can select one dir with an os_native box
in the 2 and 3 line you select sorces files and dir and one destination directory.
In the forth line there are 6 object :
1) Execute button
2) A double box where you can see the length and number of byte copied for every single file during the copy operation.
3) A Sotp indicator that you can activate to interrupt a long copy operation (if it is wrong...)
4) An Exit button
Than, there are 2 text box : one for stdout normal messages, one for stderr warning and error messages.
Hash is saved in and .md5 file compatible with md5sum.
It has a gui(Tk) and works on Linux and windows (the script should work even on Mac).
Md5 is computed while reading, so sources file are read only once.
The executables have been created using pp : it creates an executable similar to an autoextract-zip with the script, the perl engine and all libraries.
I'm using it under linux to copy GigaBytes files from ntfs internal hd to ntfs external hd with an old atom eepc(msi u100). Speed is comparable with the system copier : 8.5MB/sec or a bit better. Obviously With more powerful pc speeds are far better (3 times).
Added ability to manage very long path (>260 char) under windows (under linux is default : magic linux!).
the custom file_select box manges very long path, multiple selection of dirs and files, sort pushing column headers and Unicode(Non_latin_chars).
Choose sort type before opening dirs. After changing sort type, dirs are NOT reloaded.
Added full Unicode(utf8) support (both win and lin).
Now is possible to manage even files with name in non-latin characters (chinese,hebrew,cyrillic,greek...). Added auto-save of messages in a log file at end. Made some cleanings.
The verify phase has been moved at the end of all the copy phase (not after every single file copy as before) to reduce cache influence on md5 verification.
Added possibility to copy files replacing destination (if exists) by modification date; some interface restiling.
To do :
3) verify if very strange(illegal) file names create probleme and if error situations(disk full...) are correctly managed.
4) make a good,standard perl module/widget for FileSelectTree.pm
5) clean sources
p.s. : many thanks to Perl and Tk creators and mantainers, to perlmonks and to all that wrote examples and helps : most of my work is just copy and paste. I hope they won't be offended by this bad code.
