Prometheus text parser barfs on UTF-8 BOMs

[christianlavoie]

UTF-8 BOMs at the beginning of a text format string confuse the text parser (prometheus will complain about an invalid metric name). Tested with 1.9.1.

[beorn7]

Yeah, that's kind of expected. The text format is fully specified and doesn't include BOMs.
Understanding BOMs would imply Prometheus could act according to whatever the BOM specifies. But it can only understand the one text format specified in https://prometheus.io/docs/instrumenting/exposition_formats/ .

What would be the rationale for supporting BOMs?

(BTW: The code that handles this is in prometheus/common, so we should arguable move this issue over there. But let's see what comes out of it...)

[juliusv]

To be honest, I didn't know about BOMs yet, but research turned up:

"A protocol SHOULD forbid use of U+FEFF as a signature for those textual protocol elements that the protocol mandates to be always UTF-8, the signature function being totally useless in those cases."

(https://tools.ietf.org/html/rfc3629)

[grobie]

Lucky you @juliusv. I can remember dark times with PHP, Windows and UTF-8
... I didn't know they're still being used.

On Fri, May 27, 2016 at 5:02 PM, Julius Volz notifications@github.com
wrote:

To be honest, I didn't know about BOMs yet, but research turned up:

"A protocol SHOULD forbid use of U+FEFF as a signature for those textual
protocol elements that the protocol mandates to be always UTF-8, the
signature function being totally useless in those cases."

(https://tools.ietf.org/html/rfc3629)

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
#1679 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AAANaHNiUV_3Z1JFAsaPwFOMVpWct1Opks5qF1v0gaJpZM4IotOs
.

[christianlavoie]

As far as I can tell, most "modern" Windows tools require them.
On May 27, 2016 5:13 PM, "Tobias Schmidt" notifications@github.com wrote:

Lucky you @juliusv. I can remember dark times with PHP, Windows and UTF-8
... I didn't know they're still being used.

On Fri, May 27, 2016 at 5:02 PM, Julius Volz notifications@github.com
wrote:

To be honest, I didn't know about BOMs yet, but research turned up:

"A protocol SHOULD forbid use of U+FEFF as a signature for those textual
protocol elements that the protocol mandates to be always UTF-8, the
signature function being totally useless in those cases."

(https://tools.ietf.org/html/rfc3629)

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
<
#1679 (comment)
,
or mute the thread
<
https://github.com/notifications/unsubscribe/AAANaHNiUV_3Z1JFAsaPwFOMVpWct1Opks5qF1v0gaJpZM4IotOs

.

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
#1679 (comment),
or mute the thread
https://github.com/notifications/unsubscribe/AAGUKgS8Nblx-E42Idf9c7o2i1uhoOKVks5qF15ngaJpZM4IotOs
.

[beorn7]

We have a number of issues around the exposition format by now. How about moving them all into the docs repo, with a dedicated label? We should have clarity on the spec before starting any work, and the spec lives in the docs repo. So it might be the best point to track those issues.

[beorn7]

Moved to prometheus/docs#542

[lock]

This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.