linux remove old kernels

Recently Ubuntu-based linux SO doesn’t remove automatically old kernels, with the command

sudo apt-get autoremove --purge

This because kernels, even automatically installed, are considered manually installed, and so they are not automatically removed.

To solve this, you should do beforehand this command

sudo apt-mark auto $(apt-mark showmanual | grep -E "^linux-([[:alpha:]]+-)+[[:digit:].]+-[^-]+(|-.+)$")

With this command you say that all kernels are automatically installed, and therefore the previuous command (sudo apt-get autoremove –purge) will work.

nvidia grafical card and linux kernel

It could happen that a new kernel doesn’t work with your nvidia (geforce) grafical card. The problem could de the linux-modules, such as linux-modules-nvidia-460-5.8.0-53-generic.

You can check it by the following command

dpkg --list | egrep -i --color 'linux-image|linux-headers|linux-modules' | awk '{ print $2 }'

To solve this, if you find that the last kernel does’t have a matchintg nvidia module, you have to install manually that module

sudo apt-get install linux-modules-nvidia-460-5.8.0-53-generic

import an xml into phpmyadmin

Given that WordPress doesn’t allow to import a local file (and this is quite unfair, in my opinion), you can transform a wp site into a xml file and then import it in your local database via phpmyadmin.

But you have to format carefully the xml: look how phpmyadmin export an xml file and format your xml according to that model.

regex replace html tags

In exporting a odt file to epub LibreOffice can make many mistakes, such as get a 2nd level title not with <h2>, but with <p class=”para0″>. To fix this error, you can use regex, in this way:

find: <p class="para0">(.*?)</p>
replace: <h2>\1</h2>

and so on for similar cases.

regex “whatever”

If you want to select “Whatever” (word or character), regardless of its length, you can simply use

(.*?)

For example if you want delete all the words between <span> and </span>, as in the following row

many words <span>many other words here</span> other words

you can use

delete <span>(.*?)</span>. 

The result will be:

many words other words

automatic crop pdf margins

I found this excellent tool: PdfCropMargins, a very light app (both for Linux and Windows), which can crop automatically pdf white margins (top,bottom,left,right) even if they are very irregular (different from page to page). With a great accuracy. And without growing the pdf size.

You can use the gui, starting from a command line, such as pdf-crop-margins -gui “original-pdf.pdf” -o “target-pdf.pdf”.

save only some pdf pages from a pdf

You can use Okular, of course: print to file -> the pages you want. But thus you get an image pdf. If you want keep a text pdf, you can use a program like PdfArranger, which allows to save (only) the pages you want from a whole pdf, keeping them as searchable pdf (text).

subtitles framerate

Correcting subtitles framerate may be very important to sync it with audio/video. You can select all subtitles with an editor, and change framerate (for example to 25 from 23.976).

Thereby you should see the video correctly (with VLC or others).

quick type special characters

In Linux you can use ComposeKey, setting it for example (in System settings) as RightCtrl (the right-Ctrl key). RightCtrl is better than AltGr in Italian keyboard, to keep AltGr for some characters like ‘[‘, or ‘]’, or ‘@’, or ‘#’, otherwise unaccessible.

In that way, when you type 1) first RightCtrl 2) then ^ 3) then o, you will get ô. You don’t need to press simultaneously all the keys.

To sum up, the main simbols :

  • RightCtrl+^+o = ô
  • RightCtrl+”+o = ö
  • RightCtrl+’+o = ó