encoding/base64: ignore \r \n during decode

[gopherbot]

by jason.swank:

The base64 package does not decode base64 encoded content with newlines, as is typical
w/ email attachments and most base64 implementations every 76 characters.

The Mime RFC (2045) section 6.8 specifies all non-alphanumeric characters must be
ignored by a compliant decoder, but warnings/errors might be appropriate under some
circumstances.

Attached is a small go base64 program- base64.go.  Using it along w/ the system's base64
command, it is easy to reproduce the issue:

1) WORKS, no newlines: 
   $ echo -n "Hello, World" | base64 -w 0 | ./base64-go
   Hello, World

2) BROKEN, newlines every 12 characters:
   $ echo -n "Hello, World" | base64 -w 12 | ./base64-go
   Hello, Wo
   illegal base64 data at input byte 12

Which compiler are you using (5g, 6g, 8g, gccgo)? 8g

Which operating system are you using? Linux i686

Which revision are you using?  0beb796b4ef8+ weekly/weekly.2011-12-02

Please provide any additional information below.

Attachments:

1. base64-go.go (407 bytes)

[rsc]

Comment 1:

Labels changed: added priority-later, removed priority-medium.

[rsc]

Comment 2:

Labels changed: added priority-go1.

[robpike]

Comment 3:

Owner changed to @robpike.

Status changed to Accepted.

[robpike]

Comment 4:

assigning to brad, cc to adam. i honestly do not understand what the correct behavior
should be. encoding is particularly unclear.

Owner changed to @bradfitz.

[agl]

Comment 5:

FWIW, I think I agree that newlines should be ignored in decoding - they certainly are
in Python. I'll fix this unless someone beats me to it.

[robpike]

Comment 6:

Owner changed to builder@golang.org.

[rsc]

Comment 8:

Labels changed: added go1-must.

[dsymonds]

Comment 9:

Owner changed to @dsymonds.

[dsymonds]

Comment 10:

Status changed to Started.

[dsymonds]

Comment 11:

This issue was closed by revision 2f2b6e5.

Status changed to Fixed.