Improves Printer #428

kpp · 2017-03-30T19:27:57Z

I added coloring for columns and separators. Also I fixed some bugs, when paths in context were not colored.

BurntSushi

I think I'm mostly in favor of the changes here, although I noted a couple places where an additional allocation was added.

If you think you fixed bugs in the printer, then I think we need at least one regression test for each bug.

Finally, I'm not totally sold on adding colors for columns and separators. It seems a bit much, although, I suppose it's not a big deal. See #377.

BurntSushi · 2017-03-30T19:38:07Z

src/app.rs

+              (for fg and bg) or a text style. A special format, {type}:none, \
+              will clear all color settings for {type}.\n\nFor example, the \
+              following command will change the match color to magenta and \
+              the background color for line numbers to yellow:\n\n\


These docs also need to be added to doc/rg.1.md and the doc/rg.1 file needs to be regenerated. (If it's hard for you to install pandoc, then I can do the regeneration.)

I will update doc/rg.1.md. There is only pandoc=1.16.0.2 in my system, so do the regeneration on your on, please.

BurntSushi · 2017-03-30T19:39:38Z

src/printer.rs

-        // N.B. We can't use `write` here because of borrowing restrictions.
-        if self.context_separator.is_empty() {
-            return;
+        if !self.context_separator.is_empty() {


Why change the coding style here? I much prefer early returns.

U r the boss.

BurntSushi · 2017-03-30T19:40:30Z

src/printer.rs

-        if self.context_separator.is_empty() {
-            return;
+        if !self.context_separator.is_empty() {
+            let separator = self.context_separator.clone();


The previous code called out the borrowing restriction, but you removed that comment and added a clone. The whole point of the previous code was to avoid the clone. ;-)

BurntSushi · 2017-03-30T19:41:09Z

src/args.rs

@@ -731,8 +731,10 @@ impl<'a> ArgMatches<'a> {
        let mut specs = vec![
            "path:fg:magenta".parse().unwrap(),
            "line:fg:green".parse().unwrap(),
+            "column:fg:yellow".parse().unwrap(),


Please see #377, which asks for colors on column numbers. What's the use case? Similarly for separators.

The use case is to improve readability. Color helps humans to orient. See attached image in the description of PR

Right, but if you read #377, then I'm specifically calling into question why humans are reading column numbers, especially when the match itself is colored.

I don't know =( I wanted to color everything in the output...

I think I'd prefer if you backed out the addition of column and separator color styles. Whether we add that or not should really be decided in #377.

But... the pic... prettier... =(

More features means more code that I need to maintain. For example, if we ever added code to auto-detect the best color settings (which seems like something Windows users might appreciate), then the more color settings we have, the harder that code will be to write.

Features need to be decided upon carefully. Once they're added, we can essentially never remove them. Sorry.

BurntSushi · 2017-03-30T19:45:29Z

src/printer.rs

-                self.write(msg.as_bytes());
-                let _ = self.wtr.reset();
+                let msg = format!("[Omitted long line with {} replacements]", count);
+                self.write_colored(msg.as_bytes(), |colors| { colors.matched() });


I think the braces in your closure here are superfluous.

BurntSushi · 2017-03-30T19:45:33Z

src/printer.rs

            let msg = format!("[Omitted long line with {} matches]", count);
-            self.write(msg.as_bytes());
-            let _ = self.wtr.reset();
+            self.write_colored(msg.as_bytes(), |colors| { colors.matched() });


I think the braces in your closure here are superfluous.

BurntSushi · 2017-03-30T19:45:50Z

src/printer.rs

-        self.write_path(path.as_ref());
-        let _ = self.wtr.reset();
+    fn separator(&mut self, sep: &[u8]) {
+        self.write_colored(&sep, |colors| { colors.separator() });


I think the braces in your closure here are superfluous.

OK, this happens a lot. I'll stop picking on it.

BurntSushi · 2017-03-30T19:45:56Z

src/printer.rs

-                let _ = self.wtr.set_color(self.colors.matched());
-                self.write(&buf[m.start()..m.end()]);
-                let _ = self.wtr.reset();
+                self.write_colored(&buf[m.start()..m.end()], |colors| { colors.matched() });


I think the braces in your closure here are superfluous.

BurntSushi · 2017-03-30T19:47:06Z

src/printer.rs

+                    } else {
+                        b
+                    }
+                }).collect();


This introduces an additional allocation that wasn't there before, probably as a result of refactoring path_to_str_native to return a slice of bytes. Maybe it should return a cow instead?

https://doc.rust-lang.org/std/primitive.slice.html#method.to_vec

The above let mut path = path.to_vec(); was that allocation. I replaced for with map. No additional allocations are made by this code.

D'oh. I missed the to_vec in the old code.

kpp · 2017-03-30T20:39:19Z

src/args.rs

@@ -731,8 +731,10 @@ impl<'a> ArgMatches<'a> {
        let mut specs = vec![
            "path:fg:magenta".parse().unwrap(),
            "line:fg:green".parse().unwrap(),
+            "column:fg:yellow".parse().unwrap(),


I don't know =( I wanted to color everything in the output...

kpp · 2017-03-30T20:42:30Z

src/printer.rs

+                    } else {
+                        b
+                    }
+                }).collect();


https://doc.rust-lang.org/std/primitive.slice.html#method.to_vec

The above let mut path = path.to_vec(); was that allocation. I replaced for with map. No additional allocations are made by this code.

kpp · 2017-03-30T20:46:05Z

src/printer.rs

-        self.has_printed = true;
-        let _ = self.wtr.write_all(&self.context_separator);
-        let _ = self.wtr.write_all(&[self.eol]);
+        let separator = self.context_separator.clone();


I thought that possibility to color it overcomes 1 additional allocation?

I don't think those are our only choices. You should be able to add color without additional allocations. The code just might be a bit more circuitous.

Will you help me? I am a rust newbie)

The problem is that self.separator borrows self mutably, so you can't also send in a reference to self.context_separator. There are a few ways to solve this, but the simplest is to do what the old code did: duplicate the code and inline the call to separator manually.

kpp · 2017-03-30T21:28:20Z

See how it looks without colors for column and separator:

It is easy to mix up column :22: with the text itself.

BurntSushi · 2017-03-30T21:38:28Z

I'm not debating whether colors makes things easier to read. ripgrep has colors already, so I obviously see the utility. What I'm trying to ascertain is whether it's worth it to add more colors. I'm trying to do that by asking for use cases for reading column numbers. I don't understand why a human would find a column number to be useful information on a regular basis. Please see the discussion in #377.

kpp · 2017-03-30T22:23:13Z

@BurntSushi +regressions

kpp · 2017-03-31T18:19:11Z

@BurntSushi is there any unresolved question about this PR?

BurntSushi · 2017-03-31T18:44:16Z

Looks good, thank you!

kpp force-pushed the improve_printer branch from 9be4961 to 1789325 Compare March 30, 2017 19:35

BurntSushi requested changes Mar 30, 2017

View reviewed changes

kpp force-pushed the improve_printer branch 2 times, most recently from 9c5d469 to 8bb6d21 Compare March 30, 2017 20:43

kpp commented Mar 30, 2017

View reviewed changes

kpp force-pushed the improve_printer branch 3 times, most recently from c08095d to 691a468 Compare March 30, 2017 21:25

Improves Printer, fixes some bugs

164c2e5

kpp force-pushed the improve_printer branch from 691a468 to 164c2e5 Compare March 30, 2017 22:22

BurntSushi approved these changes Mar 31, 2017

View reviewed changes

BurntSushi merged commit aed3ccb into BurntSushi:master Mar 31, 2017

kpp changed the title ~~Improves Printer, adds ColorSpecs for column and separator~~ Improves Printer Mar 31, 2017

kpp deleted the improve_printer branch March 31, 2017 19:52

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Improves Printer #428

Improves Printer #428

kpp commented Mar 30, 2017

BurntSushi left a comment

BurntSushi Mar 30, 2017

kpp Mar 30, 2017

BurntSushi Mar 30, 2017

kpp Mar 30, 2017 •

edited

Loading

BurntSushi Mar 30, 2017

BurntSushi Mar 30, 2017

kpp Mar 30, 2017 •

edited

Loading

BurntSushi Mar 30, 2017

kpp Mar 30, 2017

BurntSushi Mar 30, 2017

kpp Mar 30, 2017

BurntSushi Mar 30, 2017

BurntSushi Mar 30, 2017

BurntSushi Mar 30, 2017

BurntSushi Mar 30, 2017

BurntSushi Mar 30, 2017

BurntSushi Mar 30, 2017

BurntSushi Mar 30, 2017

kpp Mar 30, 2017

BurntSushi Mar 30, 2017

kpp Mar 30, 2017

kpp Mar 30, 2017

kpp Mar 30, 2017

BurntSushi Mar 30, 2017

kpp Mar 30, 2017

BurntSushi Mar 30, 2017

kpp commented Mar 30, 2017

BurntSushi commented Mar 30, 2017

kpp commented Mar 30, 2017

kpp commented Mar 31, 2017

BurntSushi commented Mar 31, 2017

Improves Printer #428

Improves Printer #428

Conversation

kpp commented Mar 30, 2017

BurntSushi left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kpp Mar 30, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kpp Mar 30, 2017 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

kpp commented Mar 30, 2017

BurntSushi commented Mar 30, 2017

kpp commented Mar 30, 2017

kpp commented Mar 31, 2017

BurntSushi commented Mar 31, 2017

kpp Mar 30, 2017 •

edited

Loading

kpp Mar 30, 2017 •

edited

Loading