How to turn off the new file-picker widget in Opera

In the latest version of Opera desktop Web browser, a widget pops up whenever you want to upload a file to WordPress or a service. It adds an extra distraction and a ‘dismiss’ click, on the way to seeing your actual Windows Explorer view and your target file. This unwanted pop-up-like widget is going to become very tiresome. So let’s turn if off…

1. Go to Menu | Settings.

2. In the top-right search-box, search for “Show pop-up with clipboard and recent downloads when uploading files”, or just a fragment such as “pop-up with clipboard” will do it.

3. Turn this feature’s control-button off, via the blue button-slider.

That’s it. Exit the browser’s Settings, and your Opera browser should be back to normal.

Pinterest, begone

New and useful for picture researchers of various types, a UserScript to “Hide pinterest.com in Google Images” search results. Uses a simple ‘if result contains pinterest.com, do not display’ CSS method. The script is easy to tweak and as such it could be adapted for other image sites that you find are verbose/useless (e.g. Alamy and its ilk), without the need for a full-blown URL blocker add-on.

UserStyle to UserScript Converter for Windows

Stanley Lim’s free UserStyle to UserScript Converter as a Windows .EXE file. Version 1.5, 2017. Download, unzip, drill down to ..\CSStoUserScriptConverter-1.5\CSStoUserScript\bin\.. and run.

My test-convert of a two column layout I cooked up a while back. Works fine. Although the column-splitting method used may be Opera / Chrome specific, as well as the conversion.

eTools is especially useful now that DuckDuckGo searches turn into irrelevant mush 60% of the time.

How to kill Opera instantly

Google Search results are becoming an increasingly dangerous place, once you get past the first page or two, and for even the most experienced and protected surfer with Safe Surf enabled in both search-engine and browser, and various blockers running. There can be times when the searcher, on a perfectly normal search, finds that a simple click has ‘trapped’ them on an unwanted page.

Why “trapped”? Because some rogue pages give you what appears to be a regular browser alert, this spurious alert being deliberately triggered to prevent your exit. It prevents you using the usual ‘Back’ button. Behind its CSS styling, that alert button could be anything, and if clicked will likely spawn another and another alert… and thus drive the clueless surfer into a panic where they will click on anything. In such cases the savvy surfer will know not to click on the first alert, for fear of triggering further problems, or giving the rogue site some unwanted permission or data such as geo-location.

Best then just to kill the browser. So you Crtl Alt Del and… oh, wait. The Opera browser is running as 32 x Task Manager instances of Opera.exe. You thus have to click down each one and every one to kill the browser. There is no shift-select available, to select them all at one go. (You may have a few less instances than I do, depending on how many CPU cores your desktop PC has).

Right then… so let’s go with ‘the nuclear option’ for such rare cases. A Windows .BAT file. This then is what your .BAT file’s lines should look like….

Type the lines exactly into a normal .TXT file, save to your desktop. Rename it there as a Windows .BAT file. When you find yourself at such a rogue page, click the .BAT to kill the Web browser immediately. “Immediately” here means… in a micro-second.

In practice I find that just one termination of opera.exe is enough to kill it instantly, rather than 32. I’ve no firm idea why, when it takes multiple clicks in the Windows Task Manager. I guess it just keeps on closing instances of opera.exe until there are none left to close. So just one works via a .BAT file, and the opera_crashreporter.exe goes automatically when it no longer has an opera.exe to work with.

The above should also work for other Web browsers on Windows desktops. Just change the name of the .EXE file.

Screaming Frog SEO Spider as a Linkbot replacement

Can Screaming Frog SEO Spider be used as a Linkbot Pro replacement? SEO Spider is mature desktop PC software meant for SEO-focused webmasters that, in its free form will check 500 Web URLs at a time. Are they broken or moved, and if the latter then where are they now?

Yes, it can be used in this way, and here’s a short tutorial for absolute beginners who wouldn’t know the difference between SEO and a sea-otter.

1. Extract a plain-text list of Web URLs from your source page of bookmarks, or find the URL list you’ve already prepared and cleaned. Sobolsoft’s “Extract Data & Text From…” will extract without fuss from a saved HTML page and give you a one-per-line list, but there are also various bits of Windows freeware to do it.

2. Install Screaming Frog SEO Spider. Turn off: Top menu | Help | ‘Auto Check for Updates’.

3. Now whitelist SEO Spider in your Firewall, so the new software can access the Web…

C:\Program Files (x86)\Screaming Frog SEO Spider\jre\bin\javaw.exe
C:\Program Files (x86)\Screaming Frog SEO Spider\ScreamingFrogSEOSpider.exe
C:\Program Files (x86)\Screaming Frog SEO Spider\ScreamingFrogSEOSpiderCli.exe

4. Set Screaming Frog SEO Spider user-agent (i.e. its internal Web browser) to emulate Chrome. Top menu: Configuration | User Agent | Chrome. This is apt as its main internal browser agent is actually Chrome, in the form of a headerless Chrome browser driven by SEO Spider. To a web server, SEO Spider now looks like a normal Chrome browser.

5. Enable Always Follow Redirects in SEO Spider, so we have the required redirect details in the results. Top menu | Configuration | User Agent, as seen here…

6. While you’re in that panel also ensure HSTS is off (otherwise you may get SSL security certificate errors). It should be off by default. Having this off enables the Chrome browser to still grab the status and redirect URL, even if it can’t otherwise access the https:// website due to a squiffy SSL certificate. It’s the only link-checking software I found that can do this.

7. On the top menu bar, find the “Mode” item and use this to switch to List mode. This will enable you to load your list of Web URLs to be checked for viability. By default it is set to a zero depth for following links, except when Always Follow Redirects is on. In that case it makes an exception for redirects.

8. Load and then run the list, by pressing the “Start” button.

9. Let it run until completed. Now filter the results by the Response Codes tab. This removes a whole lot of unwanted columns from the results table.

10. Then adjust the table’s columns so you only see the columns you need. Which in this case would be…

Address | Status Code | Status | Redirect URL

11. Just the one time, then save the current UI configuration as the Default. Top Menu | File Configuration | Save as Default. Now, when you switch through to the Response Codes tab in future, you should see the arrangement of columns you just saved.

12. Nearly done. Now sort the “Status” column by clicking once on it, so all the various status codes are grouped thus…

And save the project file for later reloading, or passing to a colleague for manual checking, as needed.

You can also do things like have a Dark UI (Config | User Interface | Theme | Dark). It appears you can’t make the font a touch bigger, but you can get a .CSV out which you can then do what you want with. That’s done via Top Menu | Reports | Redirects | All Redirects.

---

So basically SEO Spider is now Linkbot Pro, albeit with the regrettable loss of that fab 1996 Windows UI feel. It has nice new touches such as ignoring SSL errors, and even the ability to check (one at a time) for presence of the URL on Google Search…

SumatraPDF adds extensive annotation tools

Extensive new support for editing PDF annotations, in the latest version (June 2021) of the popular freeware SumatraPDF PDF reader.

Also here on the JURN blog, my short guides to how to set up Sumatra for book and magazine reading (cover-page + double-page spreads, no gutters), and enable keyword search for .ePub files opened in SumatraPDF.

Regrettably it’s also removed support for embedded media. You used to at least be able to right-click on an embedded video, and “save as…” then play.

Setting up QuietRSS

I finally made the move from the trusty old FeedDemon to QuietRSS. Both are free desktop PC feedreaders for RSS feeds, but FeedDemon is increasingly old now (2013) and was starting to refuse to acquire perfectly valid Atom feeds. QuietRSS is maintained (Jan 2021) and has no such problems. It does however take some initial wrangling to set up — in terms of font choice, font sizes and CSS styles — to get it looking good and working as required. The basic steps are:

1. Tools | Options | change Font types and sizes. Arial Unicode and Segoe UI work well together.

2. Then View | Applications Style and choose a new CSS style for the UI. A number of pre-made CSS styles ship with the software, and ‘Dark’ will be the choice of many. This can then be opened with Notepad++ and hacked re: changing colours. There’s also an online ‘replace theme’ utility that can extend the Dark theme by writing a new variant of it.

3. You may find the dark blue ‘new items’ links stubbornly remain, even with a new CSS style, and thus ruin your tweaked colour scheme. Note that there are also font colours to hack in the QuiteRss.ini file, but the easier way is to use the tab tucked away at the back of the Fonts…

4. It appears the only way to compact the database is to have this done automatically at each shutdown. FeedDemon could do it manually, and compacting is important because it speeds up a feedreader running many feeds and keeps it fast. In QuietRSS the switch for compaction is reached via Tools | Options (or just press F8), and then you work through to this tab. Enable cleanup on shutdown, and also DB optimisation on shutdown…

It usually takes some time for a database to ‘learn’ what’s new and what are old posts dredged up from the past, and for the feedreader to settle down into showing you only the most recent posts.

Un-squish the line-height of the Google Search results

In recent days the snippet text in the Google Search results has been squished down in its line-height. Here’s how to fix it in a Stylus script…

#search .g {
         line-height: 1.4; !important;
    }

I simply paste this into my already-installed Classic Links fix, which runs inside the browser add-on Stylus.

Freeware to automatically screenshot every time you click

Free software that automatically takes a screenshot every time you click somewhere with your mouse? You’re in luck, there’s one built into Windows 7 and later. Or there was until the Windows 10 Creators Update, when it was bjorked.

In Windows 7 it was called Problem Steps Recorder (PSR), later just Steps Recorder, and can be found by via typing psr or steps into the Start menu search-box. It automatically makes one whole-screen screenshot per user click, but is limited in the number of screenshots it can make. It’s meant to be a quick tool that helps IT technicians see what a user is getting hung up on, without having to record and send video or launch a Remote Desktop connection.

Slightly more advanced is the Windows freeware Imago Recorder 1.2, which has no cap on the number of screenshots. You do need to manually hack its config file to get full-size screenshots (open imago.conf.xml and change resize to 0). Automatically captures the whole screen only. Although I’m guessing it may be able to capture a region if you delve into hacking the .XML config further?

The freeware Snappy can capture the whole screen on a click, and is a bit more friendly and fully-featured than Imago. I had no success with getting it to repeatedly capture a pre-defined region on a click, only the whole screen. Though it can grab a region in the usual way.

StepsToReproduce 1.0 was freeware meant as a taster for the more fully featured and paid StepShot, later StepShot Guides. StepShot was bought out for the underlying technology in 2019 and since 2020 is no longer available for purchase. However this cut-down freeware still works, and is a rarity in freeware in that it can do more than full-screen… but it appears to be limited to 800 x 600px in its region capture. The cursor being captured was a feature that could not be turned off, as StepShot was meant to be for rapidly producing software how-to documentation. But if you need the cursor gone, then try a temporary ‘one dot’ or thin ‘bar’ cursor that won’t be noticed. The full StepShot could automatically capture a custom region of unlimited size, simultaneously with user mouse clicks (or a looping macro emulating the same).

Beyond that you start to quickly get into expensive/subscription corporate territory.

Those have lots of full-screen screenshots would then need to crop their repeating target-region from each one, by using a friendly freeware batch image-cropper such as Image Tools.

How to extract Windows Explorer thumbnail previews

How to extract Windows Explorer thumbnail previews from a specific Windows folder, using old-school desktop freeware:

1. Install the freeware Q-Dir, a quad-view Windows Explorer replacement. This has a useful “print the folder as you see it” feature, something lacking in Windows Explorer.

2. Use Q-Dir to navigate to your chosen folder. Show the folder with medium or large thumbnails, as you prefer. Then print the folder view to a Microsoft .XPS file, using Q-Dir. The .XPS format was Microsoft’s attempt at a .PDF rival, and all Windows installations should be able to print to it.

3. Now install the little freeware utility STDU Extractor and load the .XPS file you just printed. This utility can extract images inside several formats, including from .XPS files. STDU will show you a preview of the available thumbnails and let you extract as .PNG files or in other image formats. For some reason its batch extract is very slow, but the individual select-and-extract is fast.

For batch processing of a folder with thousands of Windows thumbnail previews, you’re probably looking at an overnight job — due the slowness of batch in STDU Extractor. The workflow is more useful if you just want a few dozen at a uniform size, and without faffing around trying to manually take exactly-sized screenshots. As you can see from the above final-output example, the drop-shadow is also extracted. But neatly so.

What you don’t get is the extracted thumbnail being given the name of the file it represents. So far as I can tell, no such software exists for that sort of extraction.

So long as you have software that gives you Windows Explorer previews for its file-types, the above workflow can work even on files that are not images. For instance, the above test is with an E-on Vue 2016 3D scene file.

---

Saving the .XPS to .PDF and extracting images from that will not work. The preview thumbnails become fragmented into strips by the PDF printing process.

There are also freeware extractors that will attempt to load the Windows thumbnails .db database in the Windows ../System folder and extract from that. But that’s ‘pot luck’, even if you can get them to open. The above can target a specific folder and a few specific icons.