correct error message

[calvinmetcalf]

we we're throwing an error about public keys when we needed a private key

[calvinmetcalf]

though node makes the same mistake

[dcousens]

though node makes the same mistake

What about iojs?

[calvinmetcalf]

It was iojs I was testing against

[dcousens]

@calvinmetcalf can you make a PR upstream?

[calvinmetcalf]

maybe depends if it's an issue in node/iojs (in which case yes) or one in
opnessl (in which case no)

On Mon, Apr 13, 2015 at 8:55 PM Daniel Cousens notifications@github.com
wrote:

@calvinmetcalf https://github.com/calvinmetcalf can you make a PR
upstream?

—
Reply to this email directly or view it on GitHub
#12 (comment)
.

[dcousens]

Maybe they were the same, because in the end, both error messages would technically be correct?

[calvinmetcalf]

it is a technically correct error message because it is an issue with the public key cryptosystem type, but since it uses the phrase public key by it's self when you are actually give it an incorrect private key means it implies the issues is with the public key file. Confused me to no end while I was debugging and issue that ended up being caused by passing an rsa key to something that wanted an EC one.

[dcousens]

Maybe before we publish this, you should establish if this is actually
consistent with iojs. I know it sucks, but, people will rely on this being
consistent.

On Tue, Apr 14, 2015 at 11:49 PM, Calvin Metcalf notifications@github.com
wrote:

it is a technically correct error message because it is an issue with the
public key cryptosystem type, but since it uses the phrase public key by
it's self when you are actually give it an incorrect private key means it
implies the issues is with the public key file. Confused me to no end while
I was debugging and issue that ended up being caused by passing an rsa key
to something that wanted an EC one.

—
Reply to this email directly or view it on GitHub
#12 (comment)
.

[calvinmetcalf]

this is not consistent with iojs when I said node in the second comment I meant iojs (which I was testing against)

[dcousens]

@dominictarr thoughts?

[dominictarr]

If we have to choose we should be consistent with iojs. If there is a problem with iojs I am confidant we can fix it... and eventually that might get into joyent/node

[calvinmetcalf]

so the issue stems from here so more or an openssl thing then a node thing