Allow specifying file start position

[pickfire]

Like helix-term/src/commands.rs:3426:15

[archseer]

Is this really necessary? It's handled a few lines lower

[pickfire]

Yes, or otherwise it will panic when it is out of bounds.

[Rational-Curiosity]

Linux allows : in the file name. With this feature, how to open a file named helix-term/src/commands.rs:3426:15 from the command line?

[cole-h]

EDIT: Oh, I see your point now.

Then I suppose we should probably check if there's a file with the same name before starting to parse it for line and column.

[Rational-Curiosity]

EDIT: Oh, I see your point now.

Then I suppose we should probably check if there's a file with the same name before starting to parse it for line and column.

Then... How to open a file named helix-term/src/commands.rs:3426:15 existing other file named helix-term/src/commands.rs?

[cole-h]

There's no perfect solution. Even for your PR of +line:column, what if I have a file named exactly +10:10? How do I open file asdf first, alongside +10:10 without going to line 10 and column 10 of file asdf?

Do you find yourself editing files named test:10:10 alongside a file named test that often? To me, this sounds like an edge case that should be acknowledged, but not supported.

IMHO, I'd much rather support this syntax because that's what almost every compiler outputs in their diagnostics.

[Rational-Curiosity]

There's no perfect solution. Even for your PR of +line:column, what if I have a file named exactly +10:10? How do I open file asdf first, alongside +10:10 without going to line 10 and column 10 of file asdf?

Linux way: hx -- +10:10 opens a file named +10:10.
hx -- +10:10 asdf opens two files named +10:10 and asdf. hx +10:10 asdf opens asdf at point 10, 10.

Do you find yourself editing files named test:10:10 alongside a file named test that often? To me, this sounds like an edge case that should be acknowledged, but not supported.

I wrote an example, there are others: How to open a new file named helix-term/src/commands.rs:3426:15?
I have found applications that store old logs with the time at the end, for example: file.log2021-11-21_08:03:01. I think helix is awesome, why to start with annoying behaviors? it isn't necessary. Can you imagine a new user trying to open a file named file.log2021-11-21_08:03:01 and not understanding what happens?

IMHO, I'd much rather support this syntax because that's what almost every compiler outputs in their diagnostics.

This problem has been well solved by other applications. I believe that it is convenient to take advantage of their experience. For example, reading compiler output with helix and implementing file_at_point function that supports file.ext:10:10 format.

[pickfire]

I am aware of this but I haven't encounter it myself.

@Curiosidad-Racional please choose better wording, you made it sound like I purposely write this patch to annoy users.

My original thoughts was check if it exists before splitting, but I think since file extension is usually at the back rather than the syntax here so I think it won't be an issue.

Or maybe I should just add the check before splitting it?

[Rational-Curiosity]

I am aware of this but I haven't encounter it myself.

@Curiosidad-Racional please choose better wording, you made it sound like I purposely write this patch to annoy users.

As you may have noticed I'm not a native English speaker. My intention was not to offend but to clarify. I was referring to annoying behaviors:

1. opening a file named file.log2021-11-21_08:03:01
2. creating a file named file.log2021-11-21_08:03:01

I didn't refer to the PR. Altruistic contributions are admirable.

My original thoughts was check if it exists before splitting, but I think since file extension is usually at the back rather than the syntax here so I think it won't be an issue.

IMO users espect hx [filename] to open literal [filename]. Other behaviors are usually with modifiers hx --compiler <filename:line:col>. This is an example, I don't pretend that it is implemented like this... I'm giving ideas not solutions.

Or maybe I should just add the check before splitting it?

Whatever you prefer. We just don't make a possible use of helix impossible (1. or 2.).

:-)

[pickfire]

I implemented a check before splitting.

[archseer]

Suggested change

	<files>... Sets the input file to use or with pos file[:row[:col]]
	<files>... Sets the input file to use, position can also be specified via file[:row[:col]]

[pickfire]

Not very nice looking on small terminals.

[archseer]

Looks like there's a few merge conflicts