It's a known "not trivial". You have layer.textItem.width/height, but they don't always report correct values. I.e. if the layer or doc has been resized, those values are missing scale multipliers. Worse it's works differently wrong in different versions of CC and Mac/Win. What I use is below (with some utility functions missing), but it's just CC 2012 and 2014 (CS6 is just bounds). Alternatively you can get this stuff on Action Manager only, but I'm not familiar enough and let other comment on it. Photoshop = {   getTextExtents: function (layer) {   if (layer && layer.textItem) {   var text_item = layer.textItem   if (Photoshop.isCC2014() || (Photoshop.isMac() && Photoshop.isCC2012())) { // in mac the newer method is correct also in CC2012   return Photoshop._getTextExtentsCC2014(text_item)   } else if (Photoshop.isCC2012()) {   return Photoshop._getTextExtentsCC2012(text_item)   } else {   return Photoshop._getTextExtentsCS6(text_item)   }   }   },   _getTextExtentsCC2014: function (text_item) {   app.activeDocument.activeLayer = text_item.parent   var ref = new ActionReference()   ref.putEnumerated( charIDToTypeID("Lyr "), charIDToTypeID("Ordn"), charIDToTypeID("Trgt") )   var action = executeActionGet(ref)   //Photoshop._debugActionKeys(action)   var textKey = action.getObjectValue(stringIDToTypeID('textKey'))   var bounds = textKey.getObjectValue(stringIDToTypeID('bounds'))   var width = bounds.getUnitDoubleValue (stringIDToTypeID('right'))   var height = bounds.getUnitDoubleValue (stringIDToTypeID('bottom'))   var x_scale = 1   var y_scale = 1   if (textKey.hasKey(stringIDToTypeID('transform'))) {    var transform = textKey.getObjectValue(stringIDToTypeID('transform'))   x_scale = transform.getUnitDoubleValue (stringIDToTypeID('xx'))   y_scale = transform.getUnitDoubleValue (stringIDToTypeID('yy'))   }   x_scale *= width / text_item.width   y_scale *= height / text_item.height   return {   x:Math.round(text_item.position[0]),   y:Math.round(text_item.position[1]),   width:Math.round(width*x_scale),   height:Math.round(height*y_scale) }   },   _getTextExtentsCC2012: function(text_item) {   app.activeDocument.activeLayer = text_item.parent   var ref = new ActionReference()   ref.putEnumerated( charIDToTypeID("Lyr "), charIDToTypeID("Ordn"), charIDToTypeID("Trgt") )   var action = executeActionGet(ref)   var textKey = action.getObjectValue(stringIDToTypeID('textKey'))   //Photoshop._debugActionKeys(textKey)   var bounds = textKey.getObjectValue(stringIDToTypeID('bounds'))    var width = bounds.getUnitDoubleValue (stringIDToTypeID('right'))    var height = bounds.getUnitDoubleValue (stringIDToTypeID('bottom'))    var x_scale = 1   var y_scale = 1   if (textKey.hasKey(stringIDToTypeID('transform'))) {    var transform = textKey.getObjectValue(stringIDToTypeID('transform'))   x_scale = transform.getUnitDoubleValue (stringIDToTypeID('xx'))   y_scale = transform.getUnitDoubleValue (stringIDToTypeID('yy'))   }   return {   x:Math.round(text_item.position[0]),   y:Math.round(text_item.position[1]),   width:Math.round(width*x_scale),   height:Math.round(height*y_scale)   }   },   _getTextExtentsCS6: function (text_item) {   var layer = text_item.parent   return {   x:Photoshop.getLayerLeft(layer),   y:Photoshop.getLayerBottom(layer),   width:Photoshop.getLayerWidth(layer),   height:Photoshop.getLayerHeight(layer)   }   } } ... View more

Try setting 'epsOpts.width="500px", I think I hit some scenario where app.preferences.rulerUnits was not followed as default units (to be fair, it's rulers units). Also I guess it could be that EPS has coordinates in inches/mm (=static) and so the two are alternative ways of defining rendering accuracy with .resolution having higher priority. ... View more

Yes, (in Javascript API) there is the collection document.selection that offers methods for manipulating the selection. I had these functions in my library that you can use as an example isAreaEmpty: function (left, top, right, bottom) {   var selection_area = new Array(new Array(left, top), new Array(right, top), new Array(right, bottom), new Array(left, bottom))   var selection = app.activeDocument.selection   selection.select(selection_area)   try {   selection.copy()   } catch (e) {   return true   } finally {   selection.deselect()   }   return false }, clearArea: function (left, top, right, bottom) {   var selection_area = new Array(new Array(left, top), new Array(right, top), new Array(right, bottom), new Array(left, bottom))   var selection = app.activeDocument.selection   selection.select (selection_area)   selection.clear()   selection.deselect() }, deselect: function () {   app.activeDocument.selection.deselect() }, ... View more

I don't really know, but sounds cool so have to say something One issue is frame order, will your first scan provide frames 1-10, 11-20, 21-30... or 1-10, 372-381, 748-757... I.e. will you be cutting strips or moving them? To process them to a "stack" you need to have a handle on this. One option is to use very large numbering for each strip like 1-10000, 10001-20000, etc. So the aim would be to get frames 00001-00371 from first strip, then 10000-10371, etc. It's not sequential, but correctly ordered. Bigger issue is likely accuracy, i.e. how much jitter is there in each position and if there is rotation. Would be good if vertical is stable as the space between frames is constant so moving to next frame should be easy. I.e. you need to find beginning of first frame of strip and then start copying. A very simple solution would be to manually select first frame and then let script continue to scrape the rest of strip. Or if you can crop a bit, capture a bit smaller frames so you can easily hit to each. In any case I'd start with something easy, run shake removal algorithm in video editing and see what you get. ... View more

Just meant the way you access and manipulate the document is totally different. Certainly different rules apply to what can and cannot be done. Might be that the Javascript APIs are just a subset of AM, i.e. when you call a function it fires a bunch of AM code for you. Also I jumped to the Javascript wagon and the AM stuff is kinda scary ... View more

Yeah, Javascript and ScriptListener are two different worlds and usually you can do more with latter (in a bit more difficultly controlled manner). One issue might be that creating animated gifs in the Photoshop UI has changed at least in CS5, CS6 and CC. I.e. you might need different code for each. ... View more

Yes that's my understanding for Javascript, with clarification that with "compile them all into animated GIF through script" using some external program (e.g. check out ImageMagick). Whether that is an OK solution depends on what you are trying to do (just for own internal use or distribution). For the Javascript docs see http://wwwimages.adobe.com/content/dam/Adobe/en/devnet/photoshop/pdfs/photoshop-cc-javascript-ref.pdf ... View more

You can create scripts using Javascript (and a couple of other choices) that has access to most features. So you could open a bunch of PNGs and copy them as layers to your PSD in the right order. But saving an animated gif in Photoshop required you to create and setup the timeline or the the animation option is not available. And timeline is something you can't access in Javascript so I would guess it's not possible in Javascript. Personally I'd try to find some tool that can automatically convert a layered PSD file to animated gif and then make a script that compiles the layered PSD in the way you want. ... View more

Great work on tracking this down and reporting it as it's probably a pain to do! ... View more

Sounds like a general "something went wrong and dont know what to do"-error, which can be difficult to solve. Try to narrow it down in exactly what conditions it happens: - Comment that line, run script and save from PS menu (sometimes PS can give more informative errors) - Try rebooting computer and running freshly started Photoshop - Try copying file in Finder and running it for that - Try running from different folder like Desktop - Try CC 2012 - Try commenting out script operations that change the doc one at the time Idea is to see what things could make the issue go away ... View more

Well, I never had issues before CC 2014.2 (still works fine in CC 2012 and CS6), have you tested that? It might also be dependant on the layer internal state, i.e. that merge might matter. ... View more

Thanks for the feedback! The sample code is meant for the sample psd, so it's playing on assumption that there is only one layer that is not background, is visible and not locked. In the production code it's always inside a group so it's not a background and we check for locked. Visibility we don't check but it has not been the case in the PSDs we've replicated this in. However playing with this I did manage to find a workaround: it's possible to duplicate the layer to beginning of doc and then move it like below. try {     var original_doc = app.activeDocument     var  copy_doc = original_doc.duplicate("copy_doc", false) // duplicate original     var export_doc = app.documents.add(original_doc.width, original_doc.height, original_doc.resolution, "export_doc", NewDocumentMode.RGB, DocumentFill.TRANSPARENT) // create empty export doc     app.activeDocument = copy_doc   // select copy and first layer of it     var layer = copy_doc.artLayers[0]     // create dummy layer to merge to to flatten all vector & style data     var tmp_layer =  copy_doc.artLayers.add()     tmp_layer.move(layer, ElementPlacement.PLACEAFTER)     layer = layer.merge()     var duplicate = layer.duplicate (export_doc , ElementPlacement.PLACEATBEGINNING)     app.activeDocument = export_doc     var dummy = export_doc.artLayers.add()     var target = export_doc.artLayers.add()     duplicate.move(target, ElementPlacement.PLACEAFTER) } catch (e) {     alert(e.message) } finally {     copy_doc.close(SaveOptions.DONOTSAVECHANGES)     export_doc.close(SaveOptions.DONOTSAVECHANGES)    } var original_doc = app.activeDocument; var  copy_doc = original_doc.duplicate("copy_doc", false); // duplicate original var export_doc = app.documents.add(original_doc.width, original_doc.height, original_doc.resolution, "export_doc", NewDocumentMode.RGB, DocumentFill.TRANSPARENT); // create empty export doc app.activeDocument = copy_doc;   // select copy and first layer of it var layer = copy_doc.artLayers[0]; if (!layer.isBackgroundLayer) {     //alert ("is not background") if (layer.visible == true) {     //alert ("visible") if (!layer.allLocked) {     //alert ("not locked")     alert("Did you check this all?")     } } } Also if I try to move duplicated layer without setting the active document Photoshop raises the same exception "The object " <unknown> of layer 2" is not currently available". I.e. it's hitting some internal issue instead of giving the right error of "not allowed to move layers in non-selected document". It kind of makes sense, i.e. duplicating to different doc after certain layer is a combo operation that duplicates to different doc and moves and they different requirements (what is the active doc). ... View more

Ok, now I get it Accessing clipboard directly is probably hard, at least I don't know a way. Exceptions "not currently available" often mean "option in menu is disabled". A good reference is trying to do same through PS menus. In the case of having raw text (from apps like notepad) put text as ascii to clipboard and PS seems not to support it. If copy from Word you can actually paste because Word puts text and image and PS will paste the graphics as layer (i.e. you'd not get the text data). I'd try creating a Javascrip dialog window with a text area and pasting there. You could even detect when it gets text and automatically close eliminating an OK-button. ... View more

If the layer is a text layer, you can use layer.textItem.contents to get & set the text. Before setting you might have to sanitize line changes like my_text = my_text.replace(/\r\n/g, "\\r").replace(/\n/g, "\\r") ... View more

Ditto ... View more

Different things that you can do - Make a dialog window in javascript that can exists while a script is running. These can have buttons etc. to control the script, but might not be very responsive if lots of things going on - You can use Adobe Configurator to create a simple panel in Photoshop that intended to be personal collection of tools but are also good for starting scripts (~bookmark). For example you can put information, pictures and a button to run your script, but can't access Photoshop side. - You can use the CC html panels that can do full html+css+js and talk to Photoshop side (e.g. fetch/set info, run scripts). Most capable and future proof and not so hard after the beginning, but CC only - Use C SDK to code a panel We've done a Configurator panel for CS6 that is basically just a place to run the scripts (but you can do things like settings by running "update_settings.jsx" that creates a dialog window in Photoshop to edit settings). Then we have a HTML panel CC that offers more capabilities (editing in panel). A hard part is also "sharing code", i.e. keep your jsx scripts so they can be run from both CS6 and CC. ... View more

And you can save the data by putting it to an action descriptor and saving to custom options. I have assumption about JSON encoding, but many other ways of stuffing things to action descriptors are possible. savePreferences: function (metadata) {   var metadata_json = serialize_to_json(metadata)   var desc = new ActionDescriptor()   desc.putData(0, metadata_json)   app.putCustomOptions (APP_ID, desc, true) }, loadPreferences: function () {   try {   var desc = app.getCustomOptions (APP_ID)   if (desc) {   var metadata_json = desc.getData(0)   if (metadata_json) {   eval("var result = (" + metadata_json + ")")   if (result) {   return result   }   }   }   } catch (e) {   var foo = e.message   } }, ... View more

Thanks. They are checked for being locked and can't be background since in the real case it'll always be a layer in group. Invisible it might be, but has not mattered so far. And in the example PSD (and any that I've tried) that causes an exception these are all OK. ... View more

It's quite complex.. Before the offending line several hundreds of lines code has been run. But I managed to squeeze it to a small replicable format. I.e. below code runs in CC2012 but fires an exception in CC2014.2. Code is funny, but it's basically the steps of doc manipulation that happens in minimal case. Should work on any PSD, but you can get an example from here: Dropbox - bug_report.psd try {     var original_doc = app.activeDocument     var  copy_doc = original_doc.duplicate("copy_doc", false) // duplicate original     var export_doc = app.documents.add(original_doc.width, original_doc.height, original_doc.resolution, "export_doc", NewDocumentMode.RGB, DocumentFill.TRANSPARENT) // create empty export doc     app.activeDocument = copy_doc   // select copy and first layer of it     var layer = copy_doc.artLayers[0]     // create dummy layer to merge to to flatten all vector & style data     var tmp_layer =  copy_doc.artLayers.add()     tmp_layer.move(layer, ElementPlacement.PLACEAFTER)     layer = layer.merge()     // copy layer to export doc     layer.duplicate (export_doc.artLayers[0] , ElementPlacement.PLACEAFTER) } catch (e) {     alert(e.message) } finally {     copy_doc.close(SaveOptions.DONOTSAVECHANGES)     export_doc.close(SaveOptions.DONOTSAVECHANGES)   } ... View more

Not exactly your question since I've never tried to set it, but reading the text size is a known challenge.. The thing is that text_item.size does not include the transforms either in the object or in the document. So you have to check if there are transforms and calculate result. I have the following code for doing that getTextSize: function (text_item) {   return Math.round(text_item.size*getTextScaleHorizontal()) // .size is a unit value and we want to round it first to a regular number  }   getTextScaleHorizontal: function () {     var ref = new ActionReference()    ref.putEnumerated( charIDToTypeID("Lyr "), charIDToTypeID("Ordn"), charIDToTypeID("Trgt") )    var desc = executeActionGet(ref).getObjectValue(stringIDToTypeID('textKey'))    if (desc.hasKey(stringIDToTypeID('transform'))) {      var transform = desc.getObjectValue(stringIDToTypeID('transform'))    var mFactor = transform.getUnitDoubleValue (stringIDToTypeID('yy'))    return mFactor    }      return 1  } ... View more

Thanks, but what I meant was that I start script normally from panel and then minimize Photoshop so I can get back to browsing while script takes a few minutes to execute. So if I don't minimize and leave Photoshop as the open app, everything OK. If I minimize, dialog frame is left open. ... View more

Thanks also, saved my day ... View more

And if you're talking Javascript, you have the LayerSet.visible for that. Just remember you have multiple ways of going through a document ArtLayer is the thing called "layer" in Photoshop, i.e. ~content LayerSet is a group in Photoshop Layer is either, you'll know by Layer.typename .artLayers and .layerSets are collections of just those and .layers is a collection of either. ... View more

Note that CC2014 works differently from CC2012 and I've not found a reliable way for CS6. This is the code I have for getting the paragraph text box extents. getTextExtentsCC2014: function (text_item) {         app.activeDocument.activeLayer = text_item.parent         var ref = new ActionReference()         ref.putEnumerated( charIDToTypeID("Lyr "), charIDToTypeID("Ordn"), charIDToTypeID("Trgt") )         var action = executeActionGet(ref)         var textKey = action.getObjectValue(stringIDToTypeID('textKey'))         var bounds = textKey.getObjectValue(stringIDToTypeID('bounds'))         var width = bounds.getUnitDoubleValue (stringIDToTypeID('right'))         var height = bounds.getUnitDoubleValue (stringIDToTypeID('bottom'))         var x_scale = 1         var y_scale = 1         if (textKey.hasKey(stringIDToTypeID('transform'))) {              var transform = textKey.getObjectValue(stringIDToTypeID('transform'))             x_scale = transform.getUnitDoubleValue (stringIDToTypeID('xx'))             y_scale = transform.getUnitDoubleValue (stringIDToTypeID('yy'))         }         x_scale *= width / text_item.width         y_scale *= height / text_item.height         return {             x:Math.round(text_item.position[0]),             y:Math.round(text_item.position[1]),             width:Math.round(width*x_scale),             height:Math.round(height*y_scale) } }, getTextExtentsCC2012: function (text_item) {         app.activeDocument.activeLayer = text_item.parent         var ref = new ActionReference()         ref.putEnumerated( charIDToTypeID("Lyr "), charIDToTypeID("Ordn"), charIDToTypeID("Trgt") )         var action = executeActionGet(ref)         var textKey = action.getObjectValue(stringIDToTypeID('textKey'))         var bounds = textKey.getObjectValue(stringIDToTypeID('bounds'))          var width = bounds.getUnitDoubleValue (stringIDToTypeID('right'))          var height = bounds.getUnitDoubleValue (stringIDToTypeID('bottom'))          var x_scale = 1         var y_scale = 1         if (textKey.hasKey(stringIDToTypeID('transform'))) {              var transform = textKey.getObjectValue(stringIDToTypeID('transform'))             x_scale = transform.getUnitDoubleValue (stringIDToTypeID('xx'))             y_scale = transform.getUnitDoubleValue (stringIDToTypeID('yy'))         }         return {             x:Math.round(text_item.position[0]),             y:Math.round(text_item.position[1]),             width:Math.round(width*x_scale),             height:Math.round(height*y_scale)             } }, ... View more

I would not waste too much effort into it. I mean in the end it can be broken and chances are you'll annoy your paying customers. I think licensing should be just good enough to keep honest people hones. I.e. it should not be possible to just open file, change "payed=true" to get full access but you'd need to download a hacked version. I'm using just license file with license type, expiration date and a signature that is just HMAC with shared secret. The secret can be extracted, but not by 99% Also pirating is not necessarily an issue if there is a path to converting to paid. I.e. you have to "re-pirate"every time there is a new version  (e.g. every new version has a new key) means that most people (that would pay in anycase) are willing to pay if they actually need it. ... View more

Sorry, forgot this. suspendHistory tells Photoshop to run a part of script without saving the undo state for each step. Basically everything is one big undo (which is typically what you'd want in any case). Calling it gives one info text (becomes title or something) and the second parameter is a peace of Javascript you want to run like calling one function. Not 100% sure if passing parameters as references/values works. ... View more

Two things you could try: Instead of Copy/Paste use layer.duplicate() When doing lots of operations, maintaining history can be a bottleneck. Try using app.activeDocument.suspendHistory("Executing script, please wait...", "savetAllLayers(app.activeDocument, mainFolder)") ... View more