xplorer² online help: Automation of DOS (and windows) commands

xplorer²

Exponential growth to file management productivity

Automation of DOS (and windows) commands

In spite of the drag-drop and click-happy era of file management we're living in, for some tasks you can't beat the efficiency offered by traditional command line execution of programs. In the old days, instead of clicking on icons to start a program, one had to type its path/name in a command processor window (DOS console). This may seem quaint and antiquated but it is very powerful since you can specify extra arguments that modify the behaviour of the program in question, specify which files to operate upon etc.

xplorer² brings the command line into the third millennium, easing its use while maintaining all its power and flexibility. If you are familiar with the DOS console, you can regard the active pane cum addressbar combination as a graphical console that shows the folder contents of the "current directory" at all times. As you browse folders the "current directory" is updated automatically, and the commands you type can access the folder contents directly.

You can start composing text commands using Tools > DOS command menu, or just get into the addressbar with <Shift+Tab> and type the special prompt $ character that instructs xplorer² to interpret the input as a DOS command — otherwise it could be mistaken for a folder path. For instance typing $ dir and pressing <Return> will execute the DOS dir command, producing the listing of the current folder's contents - arguably not the most useful command within xplorer²'s context!

The commands you type correspond to names of executable files that exist on disk - minus the .exe or .com extension - and in the end of the day launch the respective application. So you must make sure you type the names correctly and that the files themselves can be located. If you have a program called windiff.exe located in c:\tools and you want to call it while browsing c:\work, then you must type its complete path in the addressbar, e.g. $ c:\tools\windiff. Alternatively you can add c:\tools to the PATH environmental variable and reference its contained programs just by name

DOS commands create usually text output that goes into a DOS console window. xplorer² uses its own user friendly console window, but you can also use the traditional MSDOS console.

TIP: if you want to open a MSDOS console window at the current folder location, type a solitary $ dollar (without any command) in the addressbar and press <ENTER>. This window is then independent of xplorer² — just close it when you are done with it.

Except for DOS commands ($-prompted) you can also execute generic windows programs using the special character >, e.g. typing > winword will launch MS Word. In this sense you have all the functionality of the standard Start > Run menu straight from the addressbar. This execution mode is mostly for programs that have their own user interface and don't need a console window for their output.

Except for launching both DOS and windows commands, the addressbar assists you during the typing process in a number of ways:

It supports path autocompletion using the <F1> key.
Start typing partial paths and press F1 a few times to complete folder and file names matching the stem
<Ctrl+Enter> inserts the focused filename from the active view straight in the addressbar. Using <Ctrl+arrow keys> you can move the focus around while remaining in the addressbar, so you can insert more than one filenames as part of a command.
Special tokens can be used in place of explicit command arguments. So instead of typing > windiff file1 file2 that compares only these two files you can type > windiff $n $i which is reusable for many situations. All commands are available for future use through Tools > Run history menu.

Commands launched from the addressbar normally apply only to the single focused item in the active folder view. If you want to apply the command to all selected items use a double prompt character, e.g. >> or $$. Then instead of running the command once, it will be repeated for each selected item. This construct is a simple road to scripts without a dialog window.

$-Tokens

Instead of typing specific filenames as parts of a command, you can use special characters that get automatically substituted with the name of the item(s) that happen to be selected when a command is executed. For example the commands $type file1 and $type $N are equivalent if file1 is the focused item in the active folder pane. Using special tokens like $N saves keystrokes and allows for reusable commands. So whereas $type file1 can only be used to type (show the contents of) file1, the version with $N can be used to type any file, as long as it is selected in the active view.

There is a variety of $-tokens, each representing a different part of the active item or even the all selected items taken as a whole. Let's use a hypothetical situation to clarify the meaning of each $-token. Suppose we have a dual-pane arrangement where the right view is active and the contents are:

Left pane: (inactive) shows the contents of c:\music and contains 10 items, out of which only one is selected & active, called titles.txt
Right pane: (active) shows the contents of c:\work\c++ and contains 2 items, both selected; the first one is called file.txt and the second one active.cpp. The focused item — the one with the dotted rectangle around it — is active.cpp

The following table lists all the available tokens. The third column shows what each token would have been substituted for, given the above scenario. Unless stated otherwise all tokens act on the single focused (or active) item in each pane.

Table 3. Reusable command tokens

Token	Description	Example
$N	Local filename	active.cpp
$B	Base name (before .)	active
$E	Filename extension (past .)	cpp
$P	Parent folder path	c:\work\c++
$D	File modification date	13/3/2004 10:12
$C	Parent folder plain name	c++
$F	Full path name	c:\work\c++\active.cpp
$U	UNC path name (network accessible)	\\comp\share\active.cpp
$S	All selected filenames	active.cpp file.txt
$A	All selected filenames (fullpaths)	c:\work\c++\active.cpp c:\work\c++\file.txt
$>	Like $A but selected fullpaths are saved to the list file `%TEMP%\x2tmpList.txt` and this file is used as argument	%TEMP%\x2tmpList.txt
$L	Left (top) pane path	c:\music
$R	Right (bottom) pane path	c:\work\c++
$I	Inactive pane path	c:\music
$Q	Filename from inactive pane	titles.txt
$G	Path and filename from inactive pane	c:\music\titles.txt
$Z	Temporary filename extracted from zipfolder, FTP, etc Within normal filesystem folders it behaves like $F	%TEMP%\x2TMP_001_active.cpp
$nn	Automatically incremented counter starting from number nn Possible counter formats are $1 for the sequence 1,2,... $001 for 001,002,..., $5 for 5,6,... and so on	01
$[s:l]	Extract part of the filename starting at s and for l letters $[4:3] will extract 3 letters starting from character 4 Use a negative start to count from the end of the name	ive (extracted from active.cpp)
$?	Asks for an argument	(whatever you type)
$$	Escape for a single $	$

Many of these tokens are intended for renaming large numbers of files, as they are used in File > Mass rename command. But you can use them for any other purpose too; just type as many tokens as necessary in a single command and they will all be substituted according to the above rules.

Now we are in a position to explain the celebrated windiff command used to compare the focused item in the active pane to its namesake in the inactive pane:

>WINDIFF "$N" "$I"

Consulting table 3 above, $N corresponds to the focused item, its full name without path, and $I is the path of the inactive pane (without explicit filename). Note we enclosed tokens in quotation marks to guard against spaces in paths and filenames. This generic command template can be used to compare whichever file is active — and regardless which pane is active (left or right).

$Z token can be used to extract files from compressed ZIP archives. It can be used to compare files in ZIPs with those in regular folders. Make sure the zip folder is active, and use this command:
windiff $z $g
$z will stand for the temp file extracted and $g is the focused item in the inactive (regular folder) pane. Note you must select the correct file in the inactive pane as well, perhaps using Mirror scrolling command <CTRL+M>. $I wouldn't be useful in this situation as the extracted filename is somewhat mangled with x2TMP_xxx stem.

You may have noticed that all tokens in table 3 were listed in uppercase. If you type a token in lowercase then it is interpreted in a slightly different fashion, using the short 8.3 version of whatever name the normal uppercase token would have produced. To clarify, had $N corresponded to undocumented.h the lowercase $n would have resulted in UNDOCU~1.H, which is the equivalent old DOS filename.

8.3 filenames look obscure but they don't contain any spaces so they can be used in commands without worrying for quotes. In some partitions 8.3 names are disabled (they usually exist for C: but not for extra partitions). In that case xplorer² will add quotation marks to protect against spaces in filenames, when lowercase tokens are used.
Tokens that correspond to multiple items like $A automatically add quotes as necessary (even in uppercase format)

Another kind of token can be used to extract the text of a particular file column. The generic format is ${column_name}, e.g. ${size} will fill in the file size as it appears in detailed view mode. You can use any of the available columns, which comes handy for mass renaming as e.g. you can use MP3 tags as part of the filename.

For date property tokens you can supply your own format specifier, overriding the way dates appear in columns. This is most useful when friendly dates are in force. You can supply a custom date format as ${Modified:yyMMdd HH:mm} where date and time format pictures come after the property name, separated by a colon :

The colon modifier has an application for counter tokens too, signifying a step size. So whereas plain $01 generates the sequence 01,02... the modified version $01:3 would generate the sequence 01,04,07... (numbers increasing by 3).

The addressbar in scrap windows also supports command execution. The only difference is the concept of the "current directory". So whereas in normal browser windows the current directory is synchronized with the active folder, scrap windows contain files from many folders and as a result there is no current directory in the DOS sense. Therefore you should only use tokens that carry full path information, e.g. $F instead of $N, otherwise you may have problems addressing the intended files. The tokens $L, $R and $I don't have any meaning for scraps altogether.

Command output redirection

Most DOS commands don't have a GUI of their own and rely on the console window for their output. This is the traditional black & white system DOS box, which is rather awkward to use. xplorer² comes with a substitute console that is nearly equivalent in functionality and much easier to work with, e.g. it can mark and copy text like a normal editor window, search for text, etc.

Figure 45. User-friendly vs traditional DOS console

This modern version of the DOS console automatically follows the browser window, changing its current directory as you browse folders with xplorer². The working path is shown on its title bar. It receives the output of all $-prompted commands issued from the addressbar and from command scripts, too.

The top row toolbar lets you easily copy/paste and even search for text in the console output. You can also select the font to use.

You can also type commands directly in the console window - that's the purpose of the input field ("addressbar") shown in figure 45, as well as deliver input to running programs (e.g. Y/N responses). This input area supports path autocompletion with <F1> key and maintains a history of past commands too.

The xplorer² console doesn't have a "prompt" line, but you can type extra commands in its addressbar — they will operate on the current directory (folder) you browse in the main xplorer² window. You can also type anywhere in the console window and press <ENTER> key to execute commands. This is a bit weird, but it works!

If despite all the extra functionality you prefer the traditional DOS console, you can disable it unticking "Redirect DOS command output..." program option. Note that in such a case you'll get a separate console for each DOS command you execute.

Automatic script generation

Commands launched from the addressbar operate only on the single focused item within the folder being browsed (unless a double prompt as >> is typed as discussed above). In some occasions you may need to apply the same command to a number of files as for instance for batch conversions of MP3 files. For this task it is more convenient to use Tools > Command script menu to treat all selected files in one stroke.

Figure 46. Dialog for batch file generation

In this dialog you specify a command Template that will be applied to each and every selected file name. The syntax is identical to the addressbar commands, using the familiar $-tokens instead of actual file names. Command templates may contain multiple comma-separated commands as the above snapshot illustrates. If you want to add a comma verbatim in a command escape it with two ,, commas.

If you don't remember the token names to use, you will find most of them under Special tokens drop down list. For extended information click $-token help button. As you compose your command template use the preview button to double check the command looks like what you intend it to! It will apply the template on the first selected filename and show the command(s) in the preview window. The example above inserts a counter at the end of the filename by a laborious route (copy then delete instead of a simple REN) — it was selected for demonstration rather than for its usefulness :)

The script generator replicates the command template for each selected file in succession. Plain text parts are reproduced verbatim and $-tokens are replaced on a per-file basis. If you have 100 files selected then 100 sets of commands will be generated, effortlessly. Let's have an example using the template ren "$N" "$B_$01$$.$E" on a selection that comprises 2 files, x2help.htm, x2help.ccs.

This template has four special $-tokens, N for the whole name, B for the base, E for the extension and an automatic counter starting from 1 with a leading zero ($01). The double dollar is an escape for a single $. For each file the generator copies the constant parts of the template including spaces and quotation marks, and substitutes the variable parts using the file in question. For our two files this translates to (notice the counter in the second line):

ren "x2help.htm" "x2help_01$.htm"
ren "x2help.ccs" "x2help_02$.css"

The commands are executed on each item in the order they appear in the active pane. Most of the time this isn't important unless you use automatic counter tokens; in this case make sure your selected items are sorted in the appropriate sequence before executing the script.

The use of quotation marks around the filenames covers for situations where names contain spaces. An alternative is to use lowercase token versions like $n that are guaranteed to be free of spaces, albeit rather mangled.

Once the script is generated, there are a number of alternative modes of execution, each activated by a dialog button:

Execute (DOS). This is the natural mode, running a batch file in DOS. Behind the scenes a file called x2tmpScript.bat is created in the %TEMP% directory.
Execute (shell). This is the script equivalent of >-prompted windows program execution. In this case templates cannot contain commands like ren that require the DOS interpreter.
Edit. Instead of immediate execution you can edit the script file, perhaps adding a few final touches. If you are in dual-pane mode the inactive pane will focus onto x2tmpScript.bat so that you can easily execute the final script by double-clicking, after you save your changes.

You should note that the wizard is merely manipulating strings and is oblivious about their meaning or command syntax. If you make any mistakes you will notice during execution! Always preview the template before executing the commands.

Batch scripts using DOS commands are very last century. Nowadays administrators are more likely to use powershell to work with files on the command line. xplorer² scripts can be used to execute powershell commands on selected items, as long as you enclose everything in powershell -command "...". Use full path tokens to address files as xplorer² current directory won't be transmitted to powershell. Here is an example:

powershell -command "& {(get-item '$F').LastWriteTime = get-date '${Picture date}'}"
This uses $-tokens to change the modified time of selected JPGs to the date the photo was taken!

Further reading
◪ Advanced file management using windows scripting host (WSH)

Favorite user commands

Instead of typing commands in the addressbar, you can store your favorite commands in a list for quick and frequent access. The easiest way to add a command is to first find the executable program (browse its container folder and select the EXE file), then use Customize > User commands > Add current menu command. This will pick the executable's full path and you just add whatever command line arguments you need and store it in the list.

For more customization use Customize > User commands > Organize menu, which will open a dialog like we have seen for organizing bookmarks. Edit the user command you just added to change the icon, add a keyboard shortcut for quick access and so on:

Figure 47. Setting individual user command properties

The Name is just a mnemonic for your own use, as it appears on the user command list. The important part is the Description which is the command line, including the program and the arguments. Notice you must start the command with the correct prompt character ($ for DOS and > for windows GUI tools). Click on Info button to see a reminder of available $-tokens.

The "description" field is rather narrow for most commands with rich arguments. But never mind, long commands will wrap in multiple lines. As long as you don't add newlines, you should be fine. You cannot use commas to define multi-line commands either.

For really reusable command templates, you can use environmental variables like %WINDIR% in the command paths, these are recognized and substituted for the correct full path before execution. The special environmental variable %X2DIR% stands for the full path of xplorer² installation folder, and can be used for portable commands, assuming you save your extra tools where you keep your ultimate xplorer².

Once you define a command, you can access it from Customize > User commands list. For easier access assign a keyboard shortcut to it, or add it on a toolbar through customization. User commands in toolbars are drop targets, meaning you can drag drop a file on a command button and it will open in the respective user command.

The order of command items determines their position in the respective submenu. Items in the top of the pile are easier to access so you should keep the most frequently used ones near the beginning (use organize command to reorder them). Each submenu shows up to 20 user commands; however you can define up to 100 items for each category and access them using the More... menu command.

Here are some examples of really useful external commands. The first uses the $A token to join multiple selected PDFs in a single file called OUT.PDF. The second can convert any video file to the MP4 format. Obviously you need to download the tools first and change the paths to match your installation.

Table 4. Useful external commands

Name	Description
join pdfs	$ "C:\Program Files (x86)\tools\pdftk bin\pdftk.exe" $A cat output out.pdf
convert to MP4	> C:\PROGRA~2\tools\FFMPEG.EXE -i "$N" "$B.mp4"

Optional command line arguments

When executed from the command line xplorer² accepts a number of optional arguments that control some aspects of the program. The complete command line is:

xplorer2.exe /F:n /P /R:rootFolder /S:registryKey /L:searchFromFolder /I:settings.REG /M /1 /2 leftFolder rightFolder

All arguments are optional and are described in the following table. The last two arguments — without the / switch character — specify starting left & right pane folders for the main window. These can also be the names of folder groups saved in Customize menu if you want to reinstate a particular set of tabs upon startup.

If you start with a scrap container (/F:1 switch) leftFolder can be a cida file or a name of a folder to flatten.

Table 5. xplorer² command line arguments

Switch	Function
/F	Starting window type. By default a normal browser window is opened, equivalent to /F:0. If you want to start with a scrap window use /F:1
/P	By default xplorer² runs in a system resource-friendly single process mode, even when you launch separate instances. Using /P you force new instances to open as separate processes.
/R	Roots the program on a folder other than the default desktop. For instance `/R:c:\winnt` will force both the tree and view panes on c:\winnt and won't allow users to reach other parts of the namespace unless underneath the root node; e.g. d:\ will be off-limits whereas c:\winnt\system will be accessible
/S	Registry key extension for storing program options. The default key is `HKCU\Software\ZabaraKatranemia Plc\xplorer2` but you can have multiple parallel keys with different options. Keys can be changed dynamically with Window > Save layout menu. For instance `/S:preview` will load options from the key `...\ZabaraKatranemia Plc\xplorer2_UC.preview`
/L	Causes a scrap window to issue a "Find Files" command when it first comes up. The search is rooted at "searchFromFolder"; alternatively it could be a saved search `.X2FND` file. Can only be used in conjunction with /F:1
/M	Does not allow window to open in minimized mode on startup
/N	Do not restore slow folders from last session (e.g. network) in case they are unavailable
/1	Forces single pane mode
/2	Forces dual pane mode
/T	Doesn't restore folder tabs on startup (by default you get all tabs reopened - as you left them at last use)
/D	Force xplorer² to nag you for a registration key
/I	(portable version) Initialize registry with a custom settings file. The settings file should be in the same folder as xplorer² and be created with Actions > Extract settings on your "home" PC. For ultimate version this can also be a INI file bypassing the registry
/E	Run in windows explorer replacement mode
/Z	Silently start and stay in the background without a window
/B	Execute one of the saved macros (see Customize > Macros menu) when xplorer² starts. For instance `"/B:sample macro"` will launch the saved "sample macro" (assuming the name exists). This switch is useful for automating tasks using the task scheduler

You can take advantage of these options by creating a desktop shortcut for xplorer² (or use the one added by the installer). Right click on the shortcut icon and pick Properties from the context menu. Switch to Shortcut property page, and add the required command line arguments in the Target field, past the executable name. You must leave a space between the executable and each option you add. The snapshot to the right illustrates a shortcut that launches the program using the options /1 C:\, which means start with a single pane and browse folder c:\

Likewise you can create multiple desktop shortcuts, each for a different starting folder or starting layout (passing the layout name with /S: switch). Rename these desktop shortcuts to recognize what each one does.

if the folder path you want to pass as an argument contains spaces, enclose it in "quotes". Likewise, if a layout name contains spaces, enclose the entire switch in quotes, as such:

xplorer2_64.exe "c:\some path\with spaces" "/s:layout name"