I have a lot of things to template, so I use Amber in a script. Every time, Amber wants to replace the values found in that script which is currently telling amber to replace those values, so I have to be very careful not to ruin my script!

It'd be create that if amber could detect if, for instance, it is being called by /foo/bar/create_template.sh then /foo/bar/create_template.sh would be omitted from the list of files to be replaced.

Related to https://github.com/dalance/amber/issues/47

It would be helpful if amber provided also support for renaming files.

There is a common tool called rename (from the linux-utils package), but it doesn't support regex.

On the other hand, simply doing a for loop in bash is error prone (non-interactive & requires working bash knowledge) and doesn't support recursion.

Thus, I think that it would be a nice functionality for amber to provide. Any chances?

I'm calling amber from a batch script, it executes amber.exe tens of times in order. Some of them will randomly fail. The script:

The random fails looks like this:

Example

Issues

1. ^ 1. 「(.+)\n(?! 2.) expected to work but failed to match, does amber support Negative Look Ahead?
2. ^ 1. 「(.+)\n matching works but replacing does not, it failed to capture. >「$1 replacement result:

   Expected: 1. 「ことの葉は強 (こは) く見ゆれどすまひ草露には―・つるものにざりける」〈[源順集](https://dictionary.goo.ne.jp/word/person/%E6%BA%90%E9%A0%86/#jn-212920)〉 Returned: > 1. 「

3. As other have reported, amber does not support \n in replacement string.

Use case:

Sometimes I want to do an ambitious search or search/replace but I get a bunch of junk files I don't want to touch, like a CHANGELOG.md or a package-lock.json.

It would be nice to add an argument to ambs/ambr to ignore these files, e.g.

and

The current workarounds I've been using are grep -v and trying to manually select 'no' on the replace query for the files I don't want to replace text inside. I can assure you, this is a very error prone process :sweat_smile:

It would be great to be able to install amber with Homebrew on macOS. Please provide a way to do that.

The Freedesktop standard has an environment variable XDG_CONFIG_DIRS which sets the location for config directories. On Linux for example it's usually under ~/.config/. It's nice to keep all the config files in one place and avoid cluttering the home dir.

When I use \n in the replacement string, it doesn't work:

Suggestion: -y flag to replace all without requiring any user input

I often use ambr for multiple passes of mass replacements, and it would be useful to have such a flag.

Amber gives a lot of option to ignore things for different use cases which is great. However it also comes with cost that there is a high chance that amber will run differently on different environments. And that could become hard to debug what things are being ignored and why.

In my use case is that running amber on Github Actions and somehow the target files being ignored and I just can't find why and the nature of Github Actions makes it hard to debug. Which is not a scope for Amber of course but I think having verbose option to see what amber is processing and skipping or ignoring and why would be awesome nonetheless.

If we have a symlink, e.g.

then ambr replaces 1 and then wrongly replaces 2.

When replacing, the replaced lines have wrong line ending \r\n when the file uses line ending \n, on Windows when running in cmd.exe.

In msys it doesn't react to entering y. Instead, I have to Ctrl-C and then it tries to execute y on the shell.

How to specify the number of surrounding lines to show as context with ambr? For making better replacing decisions.

EDIT: It doesn't seem to have that option, at least according to -h. I suggest adding this option with -c <N> / --context <N> It would also be useful for ambs.

Suggestion: If the number is specified with a plus (+) sign in front, only show N following lines, none behind.

sometimes, instead of wanting to replace a text with another, I want to delete the entire line of a matched value. Adding support to this would be great!

As of now, the LTO is disabled for the release build. Is this intentional?

The line matched repeatedly is duplicated.

• detect file encoding
• change keyword encoding
• change output encoding

• .gitignore search to lower directory
• .gitignore search to upper directory
• ! support
• ** support