For this purpose, I like to use swiper ¹ which I bind to C-s.

swiper searches inside current buffer and list all matches in minibuffer. To search in all open buffers, use swiper-all instead.

Sometimes I just prefer simple search like the classic vim way, then I use / which maps to evil-ex-search-forward from evil mode ².

It behaves exactly like what you do with / in vim normal mode.

For this purpose, this great blog post ⁷ describes it in detail. It use counsel-ag but same thing works for counsel-rg which is the one I use.

In short, workflow is something like this:

• C-/ (counsel-rg)
• type search pattern
• C-c C-o (ivy-occur) which saves result into a buffer
• C-x C-q (ivy-wgrep-change-to-wgrep-mode) in ivy-occur buffer
• :%s/from/to/g to do the replace
• C-c C-c (wgrep-finish-edit) save the changes

This is made possible via several Emacs packages:

• Magit
• evil-magit
• Git time machine
• Git gutter fringe

Among all these packages, Magit is the most critical one and does most of the heavy lifting. It alone enables me to perform all the git operations I can think of.

Magit website has an excellent guide to walk you through all the basic things like:

• show the current repository status
• stage/commit files/changes
• show git log history and move through logs
• rebase/merge/branching
• diffing

There are a few things from Magit I particularly like because they’re MUCH easier with magit than the git CLI or even some other git tools.

Here’s my list:

• stage a particular hunk of a file
• quickly discard changes to files/hunks
• list all the commits affecting a particular file
• list all the commits affecting a particular function

Magit is really sweet for git, but as I’m so used to Vim editing mode(that’s where the evil package really shines), I want to use Vim keybinding for operating magit buffer (for eg, use hjkl for navigating). evil-magit makes me feel right at home.

I realized there are times I want to see a particular version for a given file or just traverse through versions; git time machine is the package that helps me with that. However I need a little tweak so I can select the particular version as the starting point (other than current head).

After starting the git time machine, you can use p to go to the previous version while n goes to the next version. There is also a line at the top of the buffer to indicate which version you’re looking at along with other relevant information.

See it in action below:

Sometimes we may also find it helpful that Emacs will show the indication that some lines are changed since the last commit on the fringe of the buffer for whatever reason, that’s what git gutter fringe package is for.

It will give you the visual indication on the left fringe of the current buffer if there is change, something like this:

Again, with a little tweak we can easily get a list of changes in the current buffer and jump to the one I’m interested in:

There are basically two workflows for me.

If I just want to do things like the following with the current file:

• get diff for changes
• get list of commits
• get list of commits affect a function
• stage current file

Then I just invoke the key binding corresponds to the magit command without openning magit status buffer.

However if I want to:

• get the status of a repository
• stage/unstage certain files
• commit current staged changes
• pull/push from remote
• branching/merging

then I would use magit-status (I bound to SPC g s) to launch the magit status buffer and start from there.

Here is a demo showing some of the operations mentioned above:

Hopefully by now you have seen how awesome Emacs is as a git client.

It’s pretty much a no-brainer for anyone who already uses Emacs day to day to adopt those packages (IF they have not done so).

But even for non-Emacs user, if you need to work with git, I would still recommend it even this is the ONLY thing you use Emacs for.

With all these packages and a little bit of customization Emacs turns into a top-notch and FREE (also across platform) git client.

Pythonista provides a module called photos which can be used to do a LOT of things to your photos on your ios device.

Before we diving into the final code, I would like to explain few concepts involved really quick to help understand the code.

AssetCollection is the corresponding python type for a photo album. It holds a list of Asset object which contains two pieces information: metadata and image data for a photo.

photos module provides get_albums and get_smart_albums method to retrieve an photo album which is AssetCollection object as aforementioned.

Since I want to clear the Camera Roll and only get_smart_albums return would include it, so we will be using it to begin with.

Next step is straightforward, get the Camera Roll from return result of get_smart_albums by checking the title of an album.

After that we can just delete all the photos from the album by invoking photos.batch_delete(). That method requires a list of Asset object and AssetCollection.assets property is what we need.

That’s it!

Here’s all the code:

import photos
# NOTE: return of this will not include camera roll
# a = photos.get_albums()

# return of this will include camera roll
albums = photos.get_smart_albums()
camera_roll = [a for a in albums if a.title == 'Camera Roll'][0]
photos.batch_delete(camera_roll.assets)

# NOTE: this will not work! see photos module documentation
#camera_roll.remove_assets(camera_roll.assets)

After you run the script, iOS will popup a confirmation dialog before it actually deletes all the photos.

Once you tap Delete button, all photos are gone. I figure it’s nice and easy enough for me, if you own the Pythonista already, it might be worth a try.

This is tested with my iPhone6 (running iOS 11.3).

photos module provides much more functionality than this, you can explore its documentation to find out other use cases.