I'm writing a Quicklook generator plugin for a GPS logging file format (.fit files). I've shoved something together which loads the file and plots the path using the NSGraphics/NSBezier stuff:
(the horrible code for the above can be found here)
This is fine, but it would be much nicer to display this over Google Maps data or similar (the recorded tracks cover a fairly small distance - e.g the above covers about 20km by 10km - too small to display on a single map of the earth, so the map needs to be dynamically loaded somehow)
Things I've tried so far:
Returning HTML which shows a Javascript GMap instance. Doesn't work as the Quicklook HTML rendering doesn't allow Javascript (or Flash, Java etc)
Returning HTML containing a static Google Maps image. Quicklook wont load remote images.
Getting a NSGraphicsContext and displaying a standard WebKitView into this, modified from some thumbnailing code - this errors because you cannot initialise WebKit from a secondary thread (which seems to work fine for thumbnails, but not previews?)
With 2. I could possibly download the static Google Map image in ObjC, then display that with HTML, but the static map isn't ideal (fixed size so no zooming, and potentially slow to initially load)
I came across MacMapKit which I've not tried yet, but since it's WebKit based, I suspect it will have the same issue as 3.
Is there some way of displaying a map in a Quicklook that I am overlooking?
trying to get Google data directly in your generator (whatever way you try to do it) is not possible as the host of your plug-in is sandboxed and denies any connection to the network
trying to return HTML data and some clever javascript won't work either as the Quick Look panel also protects against network access for security reason and there is no public way to work that around
I am afraid you won't find a real solution to your problem, except by using your own map data.
Still, your use case is interesting and you should file a bug to Apple's Bug Report website.
Related
One thing I find Delphi doesn't manage well is the link between image "source" files and the image components in my project (D2006 here but I'm assuming it applies to all versions).
Say I have various static images in my project - backgrounds, toolbar button glyphs, various bits of eye candy, etc. I have a corresponding collection of PNG/BMP etc. files that I have sourced, scraped or created, and these have been loaded into the image components at design time.
One problem is that there seems to be no automated way of finding out three months later what source image file was used to load an image component. This becomes more of an issue when you need to edit one or more images.
So what I would really like, is some way of including the image source files in a build of a project. I.e. when I did a build, all of the image components would be reloaded from the sources first. At present, when I modify some of my images with PhotoShop or similar, I'm faced with a lengthy and error-prone process to reload the altered images into their respective image containers.
Is creating a resource with all of the images bound into it and adding code to load the image components at startup a viable way to go? Do others have this problem, and how do you manage it?
None of the standard components support what you are asking for. However, Thany's TPngImageList component does support the ability to associate a user-defined string with each image in the list.
Otherwise, to know which file belongs with which image, I would suggest using an .rc file to compile the external image files into the app's resources at compile-time, and then you can load the image resources into your components dynamically at run-time instead of at design-time. That way, you can manage the image files however you want. When you change a file, it will automatically be linked into the app on the next compile.
You might load the same images in runtime, and add these images in version control to ease maintenance. There are also command line tools for Lazarus (I guess you might find similar for Delphi too) which can put files into resources that you app can later use.
I load the images from ico files, populate a imagelist and then use it in my Virtual string tree for various nodes. Easy to change the ico files.
/Mikael
I'm creating a small Windows application (C++) which create some windows an such (CreateWindowEx and the like). And in one window I want to add a control that shows MHTML loaded from memory, i.e. the whole MHTML content is in a string, not in a file.
I already tried the Shell.Explorer component, but I can't find a way to inject to it anything other than HTML (see AtlAxCreateControl).
Is it possible to use the Shell.Explorer component or any other to show in-memory MHTML??
That is, without writing the content to a file and then giving the file path. That works, but it's not optimal and it'll have to be the last alternative if I can't do it any other way.
PS: If this can be done in another language or environment like C#, JS, VBS, WSH, HTAs, I'm all ears as well.
A bit late to answer, but still - there is a way to do it : use Microsoft's WebBrowser control (COM component, can be used in .Net apps as well).
Take a look an what the proposed solutions here - How to load mht from stream/string into a WebBrowser control? and here - How to display the string html contents into webbrowser control?. some of them refer to mhtml content but the solutions are applicable to ordinary html files as well.
Good luck.
I need to extract the positions of all images from a pdf-file on the Mac. For now I am considering PDFKit, but that if there is a better way of doing that, i'm open to suggestions.
So the question is, how can i parse a PDF file on the Mac, receiving (images, not necessarily), their positions within a page and page numbers using PDFKit or, probably, another library or even a program (I don't have to build it into my program, i just need it done).
UPD Seems like it is possible to do that using Adobe's Acrobat SDK.
Thanks in advance,
Timofey.
I have an old program that shows an embedded browser using the HTML Rendering library from Carbon. I am migrating it from codewarrior to Xcode, using the 10.4 SDK. While the HTML is displayed correctly, including links, the images just don't show up. I can see the alt content, and dimensions are properly set with the width and height fields.
I am doing the initialization with:
OSErr err = HRNewReference(m_HRRef, kHRRendererHTML32Type,
GetWindowPort((WindowRef) m_pWindow));
And then I open my local HTML file with:
err = HRGoToFSRef(m_HRRef, &f, false, false);
My images are also stored locally but just do not appear, it was working fine previously on my ppc-only codewarrior compilation.
I tried with web pages on Internet with HRGoToURL, and I tried replacing my pictures src fields with http:// or file:// links to images, in jpg, gif and png, always with the same result.
Are you aware of any issue like this? I know I could, and probably should, migrate to WebKit but that would me more involved.
Sylvain
This is a shot in the dark (I am completely unfamiliar with HTMLRenderingLib), but it reminds me of this. Maybe here too it doesn't load images, even local ones, asynchronously; have you tried letting the run loop loop (whichever way is most appropriate to your app: WaitNextEvent, return back to the main runloop, spin a sub event loop, …) and see if the images load?
I'm trying to load an image from the Firefox cache as the title suggests. I'm running Ubuntu, so the location of my cache is /home/me/.mozilla/firefox/xxxxxx.default/Cache
However, in the Cache (and this is on Mac, too) the filenames are just ridiculous combinations of letters and numbers. Is there a way to pinpoint a certain file?
You should take a look at the source code of the CacheViewer Add-on.
Download the file instead of installing it (right click and save as) and then extract it (it's just a Zip file, even though it has a .xpi extension), then extract the cacheviewer.jar file inside the resulting chrome folder. Finally go into content and then cacheviewer to find the javascript and XUL files.
From my brief investigation, the useful routines are in the cacheviewer.js file, though if you were hoping there would be a simple javascript one liner for accessing cached items you're probably going to be disappointed. The XUL files (which are just XML) are helpful in working out which JS functions are called to perform particular tasks. I'm not too sure how all this maps into Greasemonkey, rather than the extension environment, but hopefully there's enough code to get you started.
Ummm, that really is an internal implementation detail. But I suggest looking at how about:cache?device=disk and about:cache-entry?client=HTTP&sb=1&key=https://stackoverflow.com/Content/img/wmd/blockquote.png are implemented.
Also, http://www.securityfocus.com/infocus/1832 gives details, too. Note that Firefox doesn't use a separate file for everything...
And of course, Firefox may change the format at any time.
Just give your img src= attribute the full URL. If the image happens to be cacheable (the server sends an appropriate Expires: or Cache-control: header, for example) and it's already in the cache, Firefox will not hit the network.
HTTP caching is supposed to be invisible. When you're generating content, you generally shouldn't worry about it.
You can point REDbot at a URL to see all sorts of delicious information about its cacheability.