Re: [PATCH 0/2] teach am and rebase -q/--quiet

Stephen Boyd [off-list ref] writes:

I am not sure why "current branch $branch is up to date" and
"Fast-forwarded $branch to $onto" is echoed to stderr, but I've
wrapped those in a quiet check because it doesn't seem to be
an error. Maybe those should be changed to stdout?

There are many valid cases where it makes sense to use stderr for messages
that are not errors (e.g. diagnostics, prompts, progress reports, and
informational messages that otherwise would clutter machine parsable
output meant to go to stdout).

I do not understand why some people seem to think stderr is only for
errors.  I think we even saw a broken interpretive language environment
where the system considers it an error if a program it launched said
anything to stderr, instead of correctly diagnosing the exit status from
it?

It is a disease.

Junio C Hamano wrote:

There are many valid cases where it makes sense to use stderr for messages
that are not errors (e.g. diagnostics, prompts, progress reports, and
informational messages that otherwise would clutter machine parsable
output meant to go to stdout).

I do not understand why some people seem to think stderr is only for
errors.  I think we even saw a broken interpretive language environment
where the system considers it an error if a program it launched said
anything to stderr, instead of correctly diagnosing the exit status from
it?

It is a disease.
  

If I understand you correctly, wrapping them in quiet checks is fine.
Also, thanks for the explanation. I'll consider my self inoculated.

Heya,

On Sun, Jun 14, 2009 at 01:05, Junio C Hamano[off-list ref] wrote:

Stephen Boyd [off-list ref] writes:
I do not understand why some people seem to think stderr is only for
errors.

The name might be setting people off? ;)

-- 
Cheers,

Sverre Rabbelier