Is there possibility of copy ONLY difference of two files? Like in winmerge, but, I can't find this option
Just like on this screen- i want to copy only 'yellow part' . Of course I can do that manually, but in big file it's not too funny :-)
WinMerge has a built-in and simple way to generate such "diff only" files, that they called "patches".
Click on "Tool", then on "Generate Patch...", and enter where you want to store the result:
You will obtain (for your example) the file
4,8c4,8
< HELLO WORLD
< HELLO WORLD
< HELLO WORLD
< HELLO WORLD
< HELLO WORLD
---
>
>
>
>
>
That uses the standard, compact way of representing diff and is easy to manipulate.
There is a free app called DiffMerge.
Note : The option 'Show differences only' will be disabled until you switch to the bottom tab labeled Referenced View(Files as loaded). After that you should be able to use Differences Only view.
Here is how you do it in WinMerge:
Disable any context lines:
[View] > [Context] > [0 Lines].
Select pane of interest: Click on pane.
Select all text: [Edit] > [Select All] or [Ctrl]+[a].
Copy selection: [Edit] > [Copy] or [Ctrl]+[c].
In winmerge there is very nice feature- in tools you can generate raport in html by Tools-> Raport. After that operation you can parse generated html- and in that way get only differenece (but maybe it's not simplest solution)
Most diff tools have a patch generation functionality which can work toward the clipboard, this is mostly what you want as you will get only changed text on left and right side (if you remove the context lines). If you want only those of one side you can easily filter the patch with regard to the first character (removing all lines, using an editor, matching something like ^[+>].*$ to keep removed lines or ^[-<].*$ to keep added lines).
The diff tool in command line will output just that, piped with a grep and one of the above regular expression you are done.
WinMerge offers the "Copy Left to..." or "Copy right to..." that will copy only the differences to a folder of choice.
Related
Is there an equivalent command in microstation for the copyclip with basepoint from autocad to copy an element from one dgn to another dgn?
I tried attaching a reference and copying it that way, but then I can not detach that reference.
I get an error message about not being able to detach a model from a managed model.
Well, have you considered using the FF (FILEFENCE) command?
As you can see from the help topic it can be used in batch processing.
In short, you place a fence around the content to be saved to another file and then in the command window:
ff=<path_to_new_dgn>
Then, once you have your file, open the other DGN, and attach this new file as a reference.
Now go into the reference manager and right-click the attachment. Choose Merge to Master
Now the contents will be part of your DGN file.
Select the element(s) you want to copy.
Set a tentative snap point on the basepoint about which you'd like to copy. To set a tentative point, the default key is middle mouse button. If that button is not assigned, see this link.
Menu Bar > Edit > Copy
Open the drawing you'd like to copy into.
Menu Bar > Edit > Paste
Ensure your settings in the "Paste from Clipboard" dialog box are correct (angle, x scale, y scale).
Click where you'd like to paste.
Briefly, I would like to show a moderately complicated Photoshop action in a forum. Saving the .atn file is easy, but it is encrypted by adobe.
I found a 25,475 line .jsx file which will apparently convert it to XML but is unusable without any usage or documentation
http://ps-scripts.cvs.sourceforge.net/viewvc/ps-scripts/xtools/apps/ActionFileToXML.jsx
What is the easiest way, other than read action word, type word in text editor, to get the 6 inches of action (as seen in Photoshop) into plain text?
GORY DETAILS:
I have a large number of files which I inadvertently damaged by using perfectlyclear on them. It enhances some of the areas but pathologically destroys all darkish areas by converting them to pure black and near zero contrast. When printed, the pictures look like somebody took a black magic marker and redacted large areas. They are damaged beyond use as-is.
The Photoshop fix is to
duplicate layer
select color range, click on a black area, set fuzziness to ~12, range=100%
select expand 4, feather 3
make new mask channel
select backward (original) layer
delete (nukes blackened area under mask)
save as PNG with transparency
This leaves a PNG file with the redacted areas transparent and with feathering around them. By placing the original file beneath it, the original non-blackeded areas are shown.
I would like to document this modest solution in an ImageMagick forum but can not believe how far adobe has gone to lock my action into adobe-only tools. I want to jailbreak this and all of my other actions.
NOTE: There is a one line usage in ActionFileToXML.jsx: "This script reads an ActionFile and converts it to XML" and no documentation of any type. An alert I stumbled upon states that it will only work in CS2/3/4 and I have CS6. It has a 2007 date on it.
I have read that this .JSX is adobe's version of JavaScript and that you run them from inside Illustrator (which I don't have).
I want to figure out how to decrypt my actions and write a useable script:
USAGE: decrypt.atn.to.txt.pl encrypted.atn [-o text_file_name] <enter>
Supply fully qualified path to a .atn" file and it will be deciphered
into a useable .txt file with the same path/basename and a .txt
extension unless you use the -O option which will attempt to write to
the file name you supply.
Perhaps, I could even make a CPAN module?!
Good thing the .JSX writer had the foresight to include 0.0039% documentation or the program would be completely useless! :)
SOLUTION == and STEP by STEP instructions:
The link:
http://ps-scripts.cvs.sourceforge.net/viewvc/ps-scripts/xtools/apps/ActionFileToXML.jsx
points to a gigantic adope extend-script. Reading the file, line 3 has the ~only documentation:
// This script reads an ActionFile and converts it to XML.
The filename already tells you this: ActionFileToXML.jsx
Without wading through 25,000 lines of largely uncommented, 8-year-old code/data/??? it is completely unusable.
What the link poster failed to include was the PACKAGE containing the other 300 files which includes the README.txt, INSTALLATION.txt, /docs, etc.
The PACKAGE supplying context, install, usage, etc can be found at
http://sourceforge.net/projects/ps-scripts/files/xtools/v2.2betas/
How to Decrypt adope's .atn file, step by step:
download README.txt and xtools*.zip from http :// sourceforge.net|projects|ps-scripts|files|xtools|v2.2betas
READ README.txt and unzip zip to any place you like (and REMEMBER where you put it). NOTE: evilnet explorer will by default hide it under some mile long, incredibly ugly file path where you may never find it so use FIREFOX: set tools -> options -> general -> downloads to Always_Ask_Me (or set a reasonable download directory)
Photoshop -> actions, click on action set you want to decipher and click the "arrow box" to the right of actions -> save_actions and put them where you can find them
Photoshop file -> scripts -> browse and navigate to where you stashed ActionFileToXML.jsx and execute. This pops up a GUI as shown at http :// ps-scripts.sourceforge.net|xtools.html
Navigate to where you hid you .atn file, the XML file box will be populated with the same path/file_BASE_name and an XML extension as a default. Adjust name/location to suit
hit PROCESS and in a delightfully brief period (in my case), it was done
Get ready to marvel at the succinct efficiency with which adope stores an action like [select->color_range, localized, fuzziness=14, range=100%] (56 bytes written by hand) in only 3635 bytes of unfathomably labyrinthine XML with no default values left underspecified. It look a lot like IRS regulation fine print! ;)
The main difficulty in trying to make sense of the XML is that it is written in some funky interpreter psycho-code which bears absolutely no resemblance to the keys/clicks you actually used to create it.
One of the steps I was attempting to elucidate was was simply layer (I NEVER ToucheDER) -> layer_mask -> hide_selection. It is diabolically obfuscated as (and I quote):
<ActionItem key="TEXT" expanded="false" enabled="true" withDialog="false" dialogOptions="2" identifier="TEXT" event="make" name="Make" hasDescriptor="true"><ActionDescriptor key="make" count="3"> <DescValueType.CLASSTYPE key="1316429856" id="1316429856" symname="New" sym="Nw " classString="Channel" class="Chnl"/><DescValueType.REFERENCETYPE key="1098129440" id="1098129440" symname="At" sym="At "><ActionReference key="1098129440" id="1098129440" symname="At" sym="At " count="1">
make .. new .. channel .. at .. mask .. hideSelection? Huh?
I had to scratch my head and fiddle around with the Channels panel options before I found the menu solution.
According to the generous and personable developer, Xbytor (who patiently answers emails from agitated would-be users), this XML can be hacked (carefully), translated back into a .ATN file and used by Photoshop. A very powerful possibility.
Brian
say i have this line of code :
Object.Property.field;
Object.Property:FunctionName();
in all my sublime languages if i was to double click "Property" on either line, it would select just that word.
For some reason my lua/lua love2d syntax highlighting selects the whole line up to the ":"
How can I change this behavior, so it will only select the single word?
The reason this is occurring is because of a somewhat strange addition to the Lua Love plugin, which I assume you're using. You're using Sublime 2, so select Preferences -> Browse Packages... to open up your Packages folder, then open the Lua Love subfolder. There is a file called completions.py, which has this content:
#completions.py
import sublime
import sublime_plugin
import re
class LoveCompletions(sublime_plugin.EventListener):
ST = 3000 if sublime.version() == '' else int(sublime.version())
def on_query_completions(self, view, prefix, locations):
if self.ST < 3000 and ("lua" in view.scope_name(locations[0])):
seps = view.settings().get("word_separators")
seps = seps.replace('.', '')
view.settings().set("word_separators", seps)
Even if you don't know Python, the logic is pretty easy to follow. It sets the variable ST to Sublime's version, which is 3000+ if you're using ST3 (current build is 3061), and is 2221 (I think) for ST2. It then sets up an event listener (the process is always running in the background) checking to see if the Sublime version is less than 3000 (you're using ST2) and you have lua in your current scope (basically, your file is source.lua or source.lua.love, if you're using the plugin's language definition). If both of those are true, it removes the . character from your "word_separators" setting, which is defined in Preferences -> Settings-Default and can be overridden in Preferences -> Settings-User.
The word_separators setting controls what characters are considered to be word separators when double-clicking to select a word. Its default value is ./\\()\"'-:,.;<>~!##$%^&*|+=[]{}`~? so, for example, if you double-click on the foo part of foo-bar Sublime will only select foo, but if you double-click on the foo part of foo_bar Sublime will select the whole thing (since - is in word_separators). . is in word_separators by default, so double-clicking on foo in foo.bar will only select foo, which is expected behavior for most people, I would assume. However, this cute little plugin removes . from word_separators in Sublime Text 2, so in your case clicking on Property selects everything from the beginning of the "word" (the whitespace before Object) to the next word separator - the :, in the case of your second example.
OK, so we know what the problem is, how do we fix it? First, while you're in Packages/Lua Love, just delete completions.py altogether. There's no harm in doing so, and in fact it's actually causing harm by being there. Make sure you restart Sublime after deleting the file. Next, open Preferences -> Settings-User and add . back into the word_separators list, anywhere between the beginning and ending double-quotes. Save that file, go back to your source code, and double-clicking should once again behave normally.
Good luck!
EDIT
I submitted this pull request to delete the completions.py file from the plugin's Github repo, and it was just merged, so hopefully users in the future won't have to deal with this :)
I use Beyond Compare (version 3.1.10) to compare different versions of Delphi Form Files, but I don't want to see differences concerning ExplicitTop, ExplicitLeft, ExplicitHeight and ExplicitWidth.
Details:
These lines will always begin with a number of whitespace characters, then "ExplicitXXX = " and a number. Older versions of Delphi didn't have these lines, so I want to ignore differences where these lines are added to the newest version, and I also want to ignore differences where the number has changed.
Does anyone know how to do this?
Edit:
Duplicate (more or less) of:
How do I configure BeyondCompare to ignore SCM replaced text in comments?
Load a pair of DFM files showing the difference.
Click the Session Settings button (aka Rules w/ umpire icon) or use the Session->Session Settings menu item.
Switch to the Importance tab then click the Edit Grammar... button to open a second dialog.
Click the New... button below the top listbox to open a third dialog.
Change the Element Name option to something like Explicit*, change the Text Matching to Explicit(Left|Top|Width|Height) = \d+ and check the Match character case and Regular expression checkboxes, then click Ok, then click Ok again in the second dialog.
Explicit* should now appear in the original dialog's Grammar Elements list. Uncheck it, then change the combobox at the bottom of the dialog from Use for this view only to Update session defaults.
I don't use Beyond Compare, but if you want to have newer versions of Delphi stop adding the (IMO useless) Explicit* properties, you can use Andreas Hausladen's DDevExtensions
In my case (C#), I wanted to ignore the entire line which contained namespaces (and thus, using's) which I changed.
(Referenced Walkthrough - Ignore entire line if text exist in line
ie.
namespace INSERT.NAMESPACE.HERE
changed to
namespace INSERT.NAMESPACE.HERE.NEW
To do that
In step 5. of Craig's solution, change the Text Matching to
" .\*INSERT.NAMESPACE.HERE.\* "
(include the quotes)
That's it.
Craig Peterson's answer is correct.
N.B. However! The tab 'importance' is not always visible from Session/Session-settings. Always, from inside a Folder list view, it will not be there. It seems there are certain filetypes that do not have it either, though I'm less clear on that. BC has so many options and plugins I bet there is a workaround, but for me I have been ok so far.
http://www.scootersoftware.com/vbulletin/showthread.php?t=8457
i'm searching for a pretty print program (script, code, whatever) for Informix-4GL sources.
Do you know any ? Than you, Peter.
Have you looked at the IIUG (International Informix User Group) software archive? There are two pretty printers there (of indeterminate quality).
The other place to look would be the Aubit4GL site - an open source variant of I4GL. Again, I'm not sure that they have a pretty-printer, but it might be something they have (though a casual check doesn't show one).
I don't know if anyone is reading this post anymore, but the easiest way to get some kind of nice "pretty print" of 4gl code is to view it in the Openedge Developer Studio, then use ctrl-I to set indention. You can adjust indention in the editor settings by saying the length of "tabs". (default is 4, I use 3)
Then do a ctrl-shift-f to make all command words uppercase.
Next, you can condense the code a few lines by moving all the "DO:" statements up a line next to the "THEN" statement with this regular expression search and replace.
ctrl-f:
search "\s*\n\s*DO[:]"
replace " DO:"
make sure you click the checkbox marked regular expressions.
At this point the code is nice and tidy.
Do a ctrl-a and ctrl-c to copy it to the clipboard.
paste it in Outlook as an email without sending. Print it in color.