Buffer.alloc v4: Incorrectly returns zero-filled buffer when encoding is passed

[CrabDude]

• v4.6.1, 4.5.0, 4.6.0:
• Darwin Kernel Version 16.0.0; root:xnu-3789.1.32~3/RELEASE_X86_64 x86_64:
• Buffer.alloc(size, data, enc):

$ node
> let s = 'YQ=='
undefined
> Buffer.alloc(s.length, s, 'base64')
<Buffer 00 00 00 00>
> Buffer(s, 'base64')
<Buffer 61>
> Buffer.alloc(1, s, 'base64')
<Buffer 00>
> Buffer(s, 'base64').toString('base64')
'YQ=='
> let b = Buffer(s.length); b.write(s, 'base64'); b
<Buffer 61 00 00 00>
> b.toString('base64')
'YQAAAA=='

Also, the Buffer.alloc(size, data, enc) docs example fails:

> Buffer.alloc(11, 'aGVsbG8gd29ybGQ=', 'base64');
<Buffer 00 00 00 00 00 00 00 00 00 00 00>

[mscdex]

/cc @nodejs/buffer

[not-an-aardvark]

Working on this. The issue is that Buffer.alloc calls Buffer.fill with another Buffer as an argument. However, passing another Buffer as an argument is not supported in Node 4.

[ChALkeR]

Ow, that was my backport. Btw, it probably means that there is no testcase for that in master, if it wasn't added recently, as I backported with testcases. Will take a look at that.

[ChALkeR]

Update: the coverage was recently (2016-09-19) fixed in master in 8699ecd (#8552). That probably needs to be backported to v4.x.

[MylesBorins]

@ChALkeR would you be able to do a Pr?

On Sun, Oct 23, 2016, 10:45 AM Сковорода Никита Андреевич <
[email protected]> wrote:

Update: the coverage was recently (2016-09-19) fixed in master in 8699ecd
8699ecd
(#8552 #8552). That probably needs
to be backported to v4.x.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#9226 (comment), or mute
the thread
https://github.com/notifications/unsubscribe-auth/AAecV-1819JoJtStID1Du2txIiEqX0dQks5q2yymgaJpZM4Kdfcn
.

[ChALkeR]

@thealphanerd You mean the testcases update? I will take a look at that.

[MylesBorins]

fixed in dc3e45f and will land in v4.6.2