Strip header whitespace. Fix #1890. Code by @matthewd

[nateberkopec]

Needs a java equivalent.

[olleolleolle]

Great start!

Nitpick: is that whitespace uniform to the rest of its file?

[Jesus]

You might know this already but just in case, I want to advertise a little known feature of git. You can create a commit with several authors, just add this to your commit message:

Co-authored-by: Matthew Draper <matthew@trebex.net>

More at: https://help.github.com/en/articles/creating-a-commit-with-multiple-authors

[headius]

@nateberkopec This patch should do it. It's larger than yours because there were duplicate loads of the parse buffer that I didn't want to duplicate a third time.

diff --git a/ext/puma_http11/org/jruby/puma/Http11.java b/ext/puma_http11/org/jruby/puma/Http11.java
index 59dde372..7e38c2b2 100644
--- a/ext/puma_http11/org/jruby/puma/Http11.java
+++ b/ext/puma_http11/org/jruby/puma/Http11.java
@@ -87,7 +87,9 @@ public class Http11 extends RubyObject {
                 validateMaxLength(flen, MAX_FIELD_NAME_LENGTH, MAX_FIELD_NAME_LENGTH_ERR);
                 validateMaxLength(vlen, MAX_FIELD_VALUE_LENGTH, MAX_FIELD_VALUE_LENGTH_ERR);
 
-                ByteList b = new ByteList(Http11.this.hp.parser.buffer,field,flen);
+                ByteList buffer = Http11.this.hp.parser.buffer;
+
+                ByteList b = new ByteList(buffer,field,flen);
                 for(int i = 0,j = b.length();i<j;i++) {
                     if((b.get(i) & 0xFF) == '-') {
                         b.set(i, (byte)'_');
@@ -105,7 +107,9 @@ public class Http11 extends RubyObject {
                   f.cat(b);
                 }
 
-                b = new ByteList(Http11.this.hp.parser.buffer, value, vlen);
+                while (vlen > 0 && Character.isWhitespace(buffer.get(value + vlen - 1))) vlen--;
+
+                b = new ByteList(buffer, value, vlen);
                 v = req.op_aref(req.getRuntime().getCurrentContext(), f);
                 if (v.isNil()) {
                     req.op_aset(req.getRuntime().getCurrentContext(), f, RubyString.newString(runtime, b));

Looking at this code has made me realize how much it can be improved. The ByteList constructions here are doing full copies of the parse buffer each time, for example, and going out to java.lang.String just for a comparison is surprisingly expensive.