Bug: Commit fails when no global email address is set even though --author is used

Bug: Commit fails when no global email address is set even though --author is used

[Marcus Tillmanns]

What did you do before the bug happened? (Steps to reproduce your issue)

* Set your machines hostname to a name that does not contain "." (e.g. "ihavenodotinmyhostname")
* Make sure you have no name or email configured in your global git config
* Create a new repository and "git add" a file
* Run: git commit -m "Test" --author "My Name <my@email.com>"

What did you expect to happen? (Expected behavior)

A commit should be created with author name "My Name", and author email "my@email.com"

What happened instead? (Actual behavior)

An error is thrown, complaining about not being able to determine the email address

What's different between what you expected and what actually happened?

The email should have been taken from the "--author" argument, but instead the commit failed.

Anything else you want to add:

This does not happen if your hostname contains a ".", e.g. "myhostname.local"

[System Info]
git version:
git version 2.40.1
cpu: x86_64
no commit associated with this build
sizeof-long: 8
sizeof-size_t: 8
shell-path: /bin/sh
uname: Linux 6.5.0-15-generic #15-Ubuntu SMP PREEMPT_DYNAMIC Tue Jan  9 17:03:36 UTC 2024 x86_64
compiler info: gnuc: 12.3
libc info: glibc: 2.38
$SHELL (typically, interactive shell): /bin/bash

Re: Bug: Commit fails when no global email address is set even though --author is used

[Phillip Wood]

Hi Marcus

On 08/02/2024 15:26, Marcus Tillmanns wrote:
> What did you do before the bug happened? (Steps to reproduce your issue)
> 
> * Set your machines hostname to a name that does not contain "." (e.g. "ihavenodotinmyhostname")
> * Make sure you have no name or email configured in your global git config
> * Create a new repository and "git add" a file
> * Run: git commit -m "Test" --author "My Name <my@email.com>"
> 
> What did you expect to happen? (Expected behavior)
> 
> A commit should be created with author name "My Name", and author email "my@email.com"
> 
> What happened instead? (Actual behavior)
> 
> An error is thrown, complaining about not being able to determine the email address

This is expected as "git commit" needs an identity to use for the 
committer as well as for the author. To set the committer you can use 
the GIT_COMMITTER_NAME and GIT_COMMITTER_EMAIL environment variables if 
you don't have the relevant config set and git cannot extract a domain 
from your hostname.

Best Wishes

Phillip

Re: Bug: Commit fails when no global email address is set even though --author is used

[Marcus Tillmanns]

Uff, that’s a mean trap. Especially since there is no “—committer” option as far as I can see.
Also the difference between the two error messages is just one (the first) word.

Maybe that could be made more obvious, especially if the user specifies “—author” already.

Thanks for the info!

- Marcus 

> On 8. Feb 2024, at 16:50, Phillip Wood <phillip.wood123@gmail.com> wrote:
> 
> [You don't often get email from phillip.wood123@gmail.com. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
> 
> Hi Marcus
> 
> On 08/02/2024 15:26, Marcus Tillmanns wrote:
>> What did you do before the bug happened? (Steps to reproduce your issue)
>> 
>> * Set your machines hostname to a name that does not contain "." (e.g. "ihavenodotinmyhostname")
>> * Make sure you have no name or email configured in your global git config
>> * Create a new repository and "git add" a file
>> * Run: git commit -m "Test" --author "My Name <my@email.com>"
>> 
>> What did you expect to happen? (Expected behavior)
>> 
>> A commit should be created with author name "My Name", and author email "my@email.com"
>> 
>> What happened instead? (Actual behavior)
>> 
>> An error is thrown, complaining about not being able to determine the email address
> 
> This is expected as "git commit" needs an identity to use for the
> committer as well as for the author. To set the committer you can use
> the GIT_COMMITTER_NAME and GIT_COMMITTER_EMAIL environment variables if
> you don't have the relevant config set and git cannot extract a domain
> from your hostname.
> 
> Best Wishes
> 
> Phillip

Re: Bug: Commit fails when no global email address is set even though --author is used

[Konstantin Khomoutov]

On Fri, Feb 09, 2024 at 07:43:01AM +0000, Marcus Tillmanns wrote:

> Uff, that’s a mean trap. Especially since there is no “--committer” option
> as far as I can see. Also the difference between the two error messages is
> just one (the first) word.
> 
> Maybe that could be made more obvious, especially if the user specifies
> “--author” already.

There's a way to specify the author and e-mail on the command-line during a
single invocation:

  git -c user.name=marcus -c user.email=user@example.com commit ...

This "-c" command-line option is explained in the "root" Git manual page
(run `git help git` for instance).

The reasoning behind having different author and committer is that Git has
been created to handle the development of Linux where most of the code flowed
(and still does, I presume) in in the form of patches sent to a set of mailing
lists, so the person creating a Git commit out of such a patch in most cases
not the person who has authored the change introduced by the patch (the
author).

Also note that you do not have to set the name and e-mail globally: Git has
three levels of configuration, one of which is "local" meaning it's stored in
the repository's metadata. So you can basically run

  git config --local --add user.name marcus

to have this setting apply only to the repository which you "are in"
currently. This value will override any "global" (per-user) settings and any
"system" settings.

Re: Bug: Commit fails when no global email address is set even though --author is used

[Kristoffer Haugsbakk]

On Fri, Feb 9, 2024, at 08:43, Marcus Tillmanns wrote:
> Uff, that’s a mean trap. Especially since there is no “—committer”
> option as far as I can see.

Your report would have been more clear if you included the error:

```
$ git commit -m "Test" --author "My Name <my@email.com>"
Committer identity unknown

*** Please tell me who you are.

Run

  git config --global user.email "you@example.com"
  git config --global user.name "Your Name"

to set your account's default identity.
Omit --global to set the identity only in this repository.

fatal: unable to auto-detect email address (got '<machine>')
```

I guess there is no `--committer` since there are use-cases for
attributing the commit to someone else (like you just received a
patch or diff from someone) but there isn’t really a common use-case for
attributing the committer to someone else—the point of the “committer”
is to identify who did the commit itself.

And the error tells you about something that you would want to do in any
case. Because passing in the committer with

```
GIT_COMMITTER_NAME
GIT_COMMITTER_EMAIL
```

every time is a lot of work.

I guess the most common workflow is to configure your user globally. But
it’s often the case that you might have a work identity (like the
email). What I do is to try to remember to configure it locally in those
repositories as well. But in the most recent case I forgot about it and
had to rewrite the commits. :) So an alternative is to not set any
global user. And then you will get this complaint every time you start
work in a new repository:

```
Committer identity unknown

*** Please tell me who you are.
```

And then you can copy–paste one of your identities into the local
config. And be certain that you won’t make any commits with the wrong
identity.

That’s more work but also safer.

So I guess that “Please tell me who you are” might be a part of some
peoples’ usual repository setup workflow.

-- 
Kristoffer Haugsbakk

Re: Bug: Commit fails when no global email address is set even though --author is used

[Marcus Tillmanns]

Everything you said is true. But it misses the point.

If I try to commit for the first time, don’t want to set the global config, look at man git-commit and search for “author", it says “—author” to specify the author.
Since “comitter” is such a hidden feature that even long time users of git don’t necessarily know about it, when I then specify “—author” and get the “same” error message again, I have no clue what’s going on, since I just specified my user name and email, and still I’m told it cannot be determined.



> On 9. Feb 2024, at 09:37, Kristoffer Haugsbakk <code@khaugsbakk.name> wrote:
> 
> On Fri, Feb 9, 2024, at 08:43, Marcus Tillmanns wrote:
>> Uff, that’s a mean trap. Especially since there is no “—committer”
>> option as far as I can see.
> 
> Your report would have been more clear if you included the error:

Had I had any idea that the report was different between with / without —author I probably would have added it, or found out what the issue was.

Cheers,
Marcus

Re: Bug: Commit fails when no global email address is set even though --author is used

[Kristoffer Haugsbakk]

On Fri, Feb 9, 2024, at 09:46, Marcus Tillmanns wrote:
> Everything you said is true. But it misses the point.
>
> If I try to commit for the first time, don’t want to set the global
> config, look at man git-commit and search for “author", it says
> “—author” to specify the author.

That reminds me of a mistake I once did. I was setting things up on a
different machine and I set `author.*` in my config. Then I got very
confused when `git commit` complained. I totally failed to see that I
need to set `user.*`, not `author.*`.

Ideally I think I should have gotten this error:

```
Committer identity unknown

The `author` config is set but not `committer`.

Did you mean to set `user` (author and committer)?
```

> Since “comitter” is such a hidden feature that even long time users of
> git don’t necessarily know about it, when I then specify “—author” and
> get the “same” error message again, I have no clue what’s going on,
> since I just specified my user name and email, and still I’m told it
> cannot be determined.

This is a very good point. The _committer_ is pretty well-hidden in
normal workflows. And probably irrelevant.

So when a user gets this error:

```
Committer identity unknown

*** Please tell me who you are.
```

And have never heard of “committer” before… what is she to think? I
think it’s very natural to conclude that “author” and “committer” are
the same thing. So she thinks:

  “ Okay, so the program is complaining about the author not being
    set. (It calls it “committer” here for some reason.) But I have set
    the author…

Maybe this is another case of: it all makes perfect sense if you already
know all the concepts.

>> Your report would have been more clear if you included the error:
>
> Had I had any idea that the report was different between with / without
> —author I probably would have added it, or found out what the issue was.

You don’t know what you don’t know. That’s why it’s best to include all
context.

Cheers

-- 
Kristoffer Haugsbakk

Re: Bug: Commit fails when no global email address is set even though --author is used

[Junio C Hamano]

"Kristoffer Haugsbakk" <code@khaugsbakk.name> writes:

> So when a user gets this error:
>
> ```
> Committer identity unknown
>
> *** Please tell me who you are.
> ```
>
> And have never heard of “committer” before… what is she to think? I
> think it’s very natural to conclude that “author” and “committer” are
> the same thing. So she thinks:
>
>   “ Okay, so the program is complaining about the author not being
>     set. (It calls it “committer” here for some reason.) But I have set
>     the author…
>
> Maybe this is another case of: it all makes perfect sense if you already
> know all the concepts.
>
>>> Your report would have been more clear if you included the error:
>>
>> Had I had any idea that the report was different between with / without
>> —author I probably would have added it, or found out what the issue was.
>
> You don’t know what you don’t know. That’s why it’s best to include all
> context.

So, now, let's be productive.  When somebody who does not know much
about Git tries to commit without configuring anything and hits the
error, what is a more appropriate message to guide who does not know
what he or she does not know?

The user claims that "committer identity unknown, please tell me who
you are" were not helpful enough.  Would it make it more helpful if
we append how to "tell who they are" after that message, perhaps
with "git config" on user.email and user.name variables, or
something?

Or do we need three-way switch that does

	if (neither is known) {
        	printf("neither author or committer is known");
	} else if (author is known but committer is not known) {
		printf("author is known but committer is not"):
	} else if (author is not known but committer is known) {
		printf("committer is known but author is not"):
	} else {
		return happy;
	}

	printf("please tell us who you are...");

perhaps?

Re: Bug: Commit fails when no global email address is set even though --author is used

[Kristoffer Haugsbakk]

On Fri, Feb 9, 2024, at 18:30, Junio C Hamano wrote:
> So, now, let's be productive.  When somebody who does not know much
> about Git tries to commit without configuring anything and hits the
> error, what is a more appropriate message to guide who does not know
> what he or she does not know?
>
> The user claims that "committer identity unknown, please tell me who
> you are" were not helpful enough.  Would it make it more helpful if
> we append how to "tell who they are" after that message, perhaps
> with "git config" on user.email and user.name variables, or
> something?
>
> Or do we need three-way switch that does
>
> 	if (neither is known) {
>         	printf("neither author or committer is known");
> 	} else if (author is known but committer is not known) {
> 		printf("author is known but committer is not"):
> 	} else if (author is not known but committer is known) {
> 		printf("committer is known but author is not"):
> 	} else {
> 		return happy;
> 	}
>
> 	printf("please tell us who you are...");
>
> perhaps?

I think a three-way switch looks good. With the amendment that it steers
you towards `user.*` instead of setting both `author.*` and
`committer.*`.

Something like

• Author is set, not committer
  • Message: author is set but not committer: you might want to set
    *user* instead (prints suggested config)

I can try to make a patch later.

-- 
Kristoffer Haugsbakk

Re: Bug: Commit fails when no global email address is set even though --author is used

[Junio C Hamano]

"Kristoffer Haugsbakk" <code@khaugsbakk.name> writes:

> On Fri, Feb 9, 2024, at 18:30, Junio C Hamano wrote:
>> So, now, let's be productive.  When somebody who does not know much
>> about Git tries to commit without configuring anything and hits the
>> error, what is a more appropriate message to guide who does not know
>> what he or she does not know?
>>
>> The user claims that "committer identity unknown, please tell me who
>> you are" were not helpful enough.  Would it make it more helpful if
>> we append how to "tell who they are" after that message, perhaps
>> with "git config" on user.email and user.name variables, or
>> something?
>>
>> Or do we need three-way switch that does
>>
>> 	if (neither is known) {
>>         	printf("neither author or committer is known");
>> 	} else if (author is known but committer is not known) {
>> 		printf("author is known but committer is not"):
>> 	} else if (author is not known but committer is known) {
>> 		printf("committer is known but author is not"):
>> 	} else {
>> 		return happy;
>> 	}
>>
>> 	printf("please tell us who you are...");
>>
>> perhaps?
>
> I think a three-way switch looks good. With the amendment that it steers
> you towards `user.*` instead of setting both `author.*` and
> `committer.*`.
>
> Something like
>
> • Author is set, not committer
>   • Message: author is set but not committer: you might want to set
>     *user* instead (prints suggested config)
>
> I can try to make a patch later.

Wait. I didn't realize this when I wrote the message you are
responding to, but we *do* already suggest settig user.* variables.

If the user chose to ignore that, then there isn't much we can do to
help, is there?

Puzzled, but I'll stop here.

Re: Bug: Commit fails when no global email address is set even though --author is used

[Marcus Tillmanns]

For me the issue was that I knew already about the global config setting, but in my case I specifically did not want to set that.

What pulled my attention away from the important part was the huge wall of text about how to setup global stuff etc, making the single word in the top left basically invisible to my eye. I was thinking "I already knew all that, I’ve set the author, so you complaining still with the same error message must be a bug”.

I think, if a user tries to commit with “—author”, and git fails to figure out the comitter, it should have a specific error message about “hey, you’ve set the author, but we still have to figure out whom to set as the committer, or you can use “—authorAndComitter” if they should be the same”.

That would make it obvious to the user what’s going on.

Cheers,
Marcus


> On 9. Feb 2024, at 20:56, Junio C Hamano <gitster@pobox.com> wrote:
> 
> [You don't often get email from gitster@pobox.com. Learn why this is important at https://aka.ms/LearnAboutSenderIdentification ]
> 
> "Kristoffer Haugsbakk" <code@khaugsbakk.name> writes:
> 
>> On Fri, Feb 9, 2024, at 18:30, Junio C Hamano wrote:
>>> So, now, let's be productive.  When somebody who does not know much
>>> about Git tries to commit without configuring anything and hits the
>>> error, what is a more appropriate message to guide who does not know
>>> what he or she does not know?
>>> 
>>> The user claims that "committer identity unknown, please tell me who
>>> you are" were not helpful enough.  Would it make it more helpful if
>>> we append how to "tell who they are" after that message, perhaps
>>> with "git config" on user.email and user.name variables, or
>>> something?
>>> 
>>> Or do we need three-way switch that does
>>> 
>>>     if (neither is known) {
>>>             printf("neither author or committer is known");
>>>     } else if (author is known but committer is not known) {
>>>             printf("author is known but committer is not"):
>>>     } else if (author is not known but committer is known) {
>>>             printf("committer is known but author is not"):
>>>     } else {
>>>             return happy;
>>>     }
>>> 
>>>     printf("please tell us who you are...");
>>> 
>>> perhaps?
>> 
>> I think a three-way switch looks good. With the amendment that it steers
>> you towards `user.*` instead of setting both `author.*` and
>> `committer.*`.
>> 
>> Something like
>> 
>> • Author is set, not committer
>>  • Message: author is set but not committer: you might want to set
>>    *user* instead (prints suggested config)
>> 
>> I can try to make a patch later.
> 
> Wait. I didn't realize this when I wrote the message you are
> responding to, but we *do* already suggest settig user.* variables.
> 
> If the user chose to ignore that, then there isn't much we can do to
> help, is there?
> 
> Puzzled, but I'll stop here.
> 

Re: Bug: Commit fails when no global email address is set even though --author is used

[Junio C Hamano]

Marcus Tillmanns <Marcus.Tillmanns@qt.io> writes:

> For me the issue was that I knew already about the global config
> setting, but in my case I specifically did not want to set that.
>
> What pulled my attention away from the important part was the huge
> wall of text about how to setup global stuff etc, making the
> single word in the top left basically invisible to my eye. I was
> thinking "I already knew all that, I’ve set the author, so you
> complaining still with the same error message must be a bug”.
>
> I think, if a user tries to commit with “—author”, and git
> fails to figure out the comitter, it should have a specific error
> message about “hey, you’ve set the author, but we still have to
> figure out whom to set as the committer, or you can use
> “—authorAndComitter” if they should be the same”.
>
> That would make it obvious to the user what’s going on.

I guess so, provided that such an additional text would not trigger
the "I missed due to the huge wall of text" again.  So in short, you
are saying that the three-way message would help?

Can I get you intereseted enough to care to come up with a patch
;-)?

Thanks.

>>>> Or do we need three-way switch that does
>>>> 
>>>>     if (neither is known) {
>>>>             printf("neither author or committer is known");
>>>>     } else if (author is known but committer is not known) {
>>>>             printf("author is known but committer is not"):
>>>>     } else if (author is not known but committer is known) {
>>>>             printf("committer is known but author is not"):
>>>>     } else {
>>>>             return happy;
>>>>     }
>>>> 
>>>>     printf("please tell us who you are...");
>>>> 
>>>> perhaps?

Re: Bug: Commit fails when no global email address is set even though --author is used

[Kristoffer Haugsbakk]

On Fri, Feb 9, 2024, at 20:56, Junio C Hamano wrote:
>> I think a three-way switch looks good. With the amendment that it steers
>> you towards `user.*` instead of setting both `author.*` and
>> `committer.*`.
>>
>> Something like
>>
>> • Author is set, not committer
>>   • Message: author is set but not committer: you might want to set
>>     *user* instead (prints suggested config)
>>
>> I can try to make a patch later.
>
> Wait. I didn't realize this when I wrote the message you are
> responding to, but we *do* already suggest settig user.* variables.
>
> If the user chose to ignore that, then there isn't much we can do to
> help, is there?
>
> Puzzled, but I'll stop here.

Aye, good point. Maybe I misremembered and/or didn’t look carefully
enough at the error message back when I set `author.*` instead of
`user.*`.

Maybe the error could say (back to the multi-way switch):

```
Author identity known, but committer identity unknown

*** Please tell me who you are.

Run

  git config --global user.email "code@khaugsbakk.name"
  git config --global user.name "Kristoffer Haugsbakk"

to set your account's default identity.
Omit --global to set the identity only in this repository.

fatal: empty ident name (for <>) not allowed
```

And then (see above) fill in the real information in the proposed config
commands if the other identity is known. Maybe that could be
accomplished with one more parameter to `ident_env_hint`:

```
static void ident_env_hint(enum want_ident whose_ident, char *other_ident)
```

(turned out to be a bit more tricky than I thought)

But that’s a slightly longer error message, which (again) can cause the
already-discussed glaze-over effect caused by the “wall of text”.

-- 
Kristoffer Haugsbakk