mirror of
https://github.com/deadc0de6/dotdrop.git
synced 2026-02-05 17:53:52 +00:00
50 lines
1.6 KiB
Markdown
Vendored
50 lines
1.6 KiB
Markdown
Vendored
# Handle special chars
|
|
|
|
* [Detect encoding](#detect-encoding)
|
|
* [Special chars](#special-chars)
|
|
* [Re-encode](#re-encode)
|
|
|
|
---
|
|
|
|
## Detect encoding
|
|
|
|
Text file encoding can be identified using, for example, `file -b <file-path>` or in vim
|
|
with `:set fileencoding`.
|
|
|
|
Here's an example of encoding that will fully work with dotdrop:
|
|
```bash
|
|
$ file -b <some-file>
|
|
UTF-8 Unicode text, with escape sequences
|
|
```
|
|
|
|
and another that will mislead the `compare` command and return false/inaccurate results:
|
|
```bash
|
|
$ file -b <some-file>
|
|
ISO-8859 text, with escape sequences
|
|
```
|
|
|
|
## Special chars
|
|
|
|
### CRLF
|
|
|
|
The use of dotfiles with DOS/Windows line endings (CRLF, `\r\n`) will result in
|
|
the comparison (`compare`) returning a difference where there is none.
|
|
This is due to Jinja2 stripping CRLF.
|
|
|
|
One solution is to use `dos2unix` to re-format the dotfiles before adding them
|
|
to dotdrop.
|
|
|
|
See <https://github.com/deadc0de6/dotdrop/issues/42>.
|
|
|
|
### Non-Unicode chars
|
|
|
|
Jinja2 is not able to process non-Unicode chars (<https://jinja.palletsprojects.com/en/2.11.x/api/>). This means that dotfiles using non-Unicode chars can still be fully managed by dotdrop; however, when comparing the local file with the one stored in dotdrop, `compare` will return a difference even if there is none.
|
|
|
|
Either replace the non-Unicode chars (see below [Re-encode](#re-encode)) or accept the fact the comparison shows a difference while there's none.
|
|
|
|
See <https://github.com/deadc0de6/dotdrop/issues/42>.
|
|
|
|
## Re-encode
|
|
|
|
To change an existing file's encoding, you can use `recode UTF-8 <filename>` (see [recode](https://linux.die.net/man/1/recode)) or in vim `:set fileencoding=utf-8`.
|