threads / patch / 32347

patchDocumentation: don't link to example mail addresses

Subject: [PATCH] Documentation: don't link to example mail addresses

## tl;dr

12 messages between Dec 15, 2012 and Dec 17, 2012. Diffs are folded; open one to read it.

replies: 11people: 3as markdown or json

John Keeping· Dec 15, 2012, 15:03 UTC · lore

Mail addresses in documentation are converted into mailto: hyperlinks in the HTML output and footnotes in man pages. This isn't desirable for cases where the address is used as an example and is not valid.

Particularly annoying is the example "jane@laptop.(none)" which appears in git-shortlog(1) as "jane@laptop[1].(none)", with note 1 saying:

	1. jane@laptop
	   mailto:jane@laptop

Fix this by quoting example mail addresses with "$$", preventing Asciidoc from processing them.

In the case of mailmap.txt, render the address monospaced so that it matches the block examples surrounding that paragraph.

Signed-off-by: John Keeping <john@keeping.me.uk>
---

I can't see any other uses of the "$$" quote in the documentation, so it's probably worth noting that I've tested this with Asciidoc 8.6.8, although I can't see anything in the changelog to indicate that Asciidoc's treatment of it has changed recently.

 Documentation/git-fast-import.txt | 2 +-
 Documentation/git-tag.txt         | 2 +-
 Documentation/mailmap.txt         | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)
Show changes to 3 files +3 −3

Documentation/git-fast-import.txt, Documentation/git-tag.txt, Documentation/mailmap.txt

diff --git a/Documentation/git-fast-import.txt b/Documentation/git-fast-import.txt
index d1844ea..05913cc 100644
--- a/Documentation/git-fast-import.txt
+++ b/Documentation/git-fast-import.txt
@@ -427,7 +427,7 @@ they made it.
 
 Here `<name>` is the person's display name (for example
 ``Com M Itter'') and `<email>` is the person's email address
-(``cm@example.com'').  `LT` and `GT` are the literal less-than (\x3c)
+(``$$cm@example.com$$'').  `LT` and `GT` are the literal less-than (\x3c)
 and greater-than (\x3e) symbols.  These are required to delimit
 the email address from the other fields in the line.  Note that
 `<name>` and `<email>` are free-form and may contain any sequence
diff --git a/Documentation/git-tag.txt b/Documentation/git-tag.txt
index 247534e..ed63edb 100644
--- a/Documentation/git-tag.txt
+++ b/Documentation/git-tag.txt
@@ -129,7 +129,7 @@ This option is only applicable when listing tags without annotation lines.
 CONFIGURATION
 -------------
 By default, 'git tag' in sign-with-default mode (-s) will use your
-committer identity (of the form "Your Name <your@email.address>") to
+committer identity (of the form "Your Name $$<your@email.address>$$") to
 find a key.  If you want to use a different default key, you can specify
 it in the repository configuration as follows:
 
diff --git a/Documentation/mailmap.txt b/Documentation/mailmap.txt
index 288f04e..dd89fca 100644
--- a/Documentation/mailmap.txt
+++ b/Documentation/mailmap.txt
@@ -46,7 +46,7 @@ Jane Doe         <jane@desktop.(none)>
 Joe R. Developer <joe@example.com>
 ------------
 
-Note how there is no need for an entry for <jane@laptop.(none)>, because the
+Note how there is no need for an entry for `<jane@laptop.(none)>`, because the
 real name of that author is already correct.
 
 Example 2: Your repository contains commits from the following

-- 
1.8.0
Jeff King· Dec 15, 2012, 17:20 UTC · re: John Keeping · lore

Re: [PATCH] Documentation: don't link to example mail addresses

On Sat, Dec 15, 2012 at 03:03:15PM +0000, John Keeping wrote:
Show 9 quoted lines
> Mail addresses in documentation are converted into mailto: hyperlinks in
> the HTML output and footnotes in man pages.  This isn't desirable for
> cases where the address is used as an example and is not valid.
> 
> Particularly annoying is the example "jane@laptop.(none)" which appears
> in git-shortlog(1) as "jane@laptop[1].(none)", with note 1 saying:
> 
> 	1. jane@laptop
> 	   mailto:jane@laptop
Thanks, this is definitely worth fixing.
Show 5 quoted lines
> Fix this by quoting example mail addresses with "$$", preventing
> Asciidoc from processing them.
> 
> In the case of mailmap.txt, render the address monospaced so that it
> matches the block examples surrounding that paragraph.

I think I'd just render them monospace everywhere. We are very inconsistent about which form of quotes we use in the documentation (I think because most of the developers read the source directly and not the rendered asciidoc). And then we don't have to worry about the "$$" construct (and IMHO it makes the source much more readable, and marking the address as a literal looks good in the output, too).

-Peff
Junio C Hamano· Dec 15, 2012, 18:00 UTC · re: Jeff King · lore

Re: [PATCH] Documentation: don't link to example mail addresses

Jeff King <peff@peff.net> writes:
Show 6 quoted lines
> I think I'd just render them monospace everywhere. We are very
> inconsistent about which form of quotes we use in the documentation (I
> think because most of the developers read the source directly and not
> the rendered asciidoc). And then we don't have to worry about the "$$"
> construct (and IMHO it makes the source much more readable, and marking
> the address as a literal looks good in the output, too).

Agreed---thanks for spotting, John, and thanks for showing the overall direction, Peff.

John Keeping· Dec 15, 2012, 18:24 UTC · re: Jeff King · lore

Re: [PATCH] Documentation: don't link to example mail addresses

On Sat, Dec 15, 2012 at 12:20:18PM -0500, Jeff King wrote:
Show 25 quoted lines
> On Sat, Dec 15, 2012 at 03:03:15PM +0000, John Keeping wrote:
>> Mail addresses in documentation are converted into mailto: hyperlinks in
>> the HTML output and footnotes in man pages.  This isn't desirable for
>> cases where the address is used as an example and is not valid.
>> 
>> Particularly annoying is the example "jane@laptop.(none)" which appears
>> in git-shortlog(1) as "jane@laptop[1].(none)", with note 1 saying:
>> 
>> 	1. jane@laptop
>> 	   mailto:jane@laptop
> 
> Thanks, this is definitely worth fixing.
> 
>> Fix this by quoting example mail addresses with "$$", preventing
>> Asciidoc from processing them.
>> 
>> In the case of mailmap.txt, render the address monospaced so that it
>> matches the block examples surrounding that paragraph.
> 
> I think I'd just render them monospace everywhere. We are very
> inconsistent about which form of quotes we use in the documentation (I
> think because most of the developers read the source directly and not
> the rendered asciidoc). And then we don't have to worry about the "$$"
> construct (and IMHO it makes the source much more readable, and marking
> the address as a literal looks good in the output, too).

I agree that the source is more readable as monospaced, but I think we need to keep the quotes around the text in git-tag(1) and probably git-fast-import(1) so that it is differentiated from the surrounding text in the man page output.

I just tried this and got strange results. Taking the example in git-tag(1):

    ...of the form ``Your Name <your@email.address>''

I tried this to preserve the quotes and make the entire quoted text monospaced (I don't think we want to have just the email address monospaced):

    ...of the form ```Your Name <your@email.address>`''

which had the desired effect - smart quotes and everything inside rendered monospaced - BUT the email address is hyperlinked!

Switching the smart Asciidoc double quotes (``...'') for "normal" double quotes ("...") produces the desired result, but IMHO doesn't look quite as good (on the other hand it looks like there are many more use of "..." than ``...'' in the Git documentation).

[While searching, the only other example I could find is in git-commit(1) where `A U Thor <author@example.com>` is rendered monospaced without quotes, but I think this hurts readability in the man page output - expect a follow-up patch if we resolve this in the direction of having quotes and monospaced text.]

At this point I've exhausted my Asciidoc knowledge and I'm not entirely happy with any of these options. Suggestions?

John
Jeff King· Dec 16, 2012, 12:04 UTC · re: John Keeping · lore

Re: [PATCH] Documentation: don't link to example mail addresses

On Sat, Dec 15, 2012 at 06:24:09PM +0000, John Keeping wrote:
Show 11 quoted lines
> > I think I'd just render them monospace everywhere. We are very
> > inconsistent about which form of quotes we use in the documentation (I
> > think because most of the developers read the source directly and not
> > the rendered asciidoc). And then we don't have to worry about the "$$"
> > construct (and IMHO it makes the source much more readable, and marking
> > the address as a literal looks good in the output, too).
> 
> I agree that the source is more readable as monospaced, but I think we
> need to keep the quotes around the text in git-tag(1) and probably
> git-fast-import(1) so that it is differentiated from the surrounding
> text in the man page output.

Hmm, good point. I use MAN_BOLD_LITERAL, which serves that purpose, but I guess not everyone does (and it is not the default; I wonder if it should be). But if MAN_BOLD_LITERAL is not in use, then `` literals have zero typographical impact in the manpages.

So maybe we do need to do something else. I was just hoping we could get away with something more readable in the source.

By my reading of the asciidoc user guide, the mechanism that is hurting us is that mailto is an inline macro, and that foo@bar automatically triggers that inline macro. Furthermore, the right way to suppress expansion of macros is with a backslash escape. Doing this:

Show changes to Documentation/git-fast-import.txt +1 −2
diff --git a/Documentation/git-fast-import.txt b/Documentation/git-fast-import.txt
index d1844ea..68bca1a 100644
--- a/Documentation/git-fast-import.txt
+++ b/Documentation/git-fast-import.txt
@@ -427,7 +427,7 @@ Here `<name>` is the person's display name (for example
 
 Here `<name>` is the person's display name (for example
 ``Com M Itter'') and `<email>` is the person's email address
-(``cm@example.com'').  `LT` and `GT` are the literal less-than (\x3c)
+(``\cm@example.com'').  `LT` and `GT` are the literal less-than (\x3c)
 and greater-than (\x3e) symbols.  These are required to delimit
 the email address from the other fields in the line.  Note that
 `<name>` and `<email>` are free-form and may contain any sequence

seems to produce the output we want. I think it's a little less ugly
than the "$$" quoting, but not by much. No clue if one is accepted by
more asciidoc versions or not.

-Peff
John Keeping· Dec 16, 2012, 14:00 UTC · re: Jeff King · lore

[PATCH v2] Documentation: don't link to example mail addresses

Email addresses in documentation are converted into mailto: hyperlinks in the HTML output and footnotes in man pages. This isn't desirable for cases where the address is used as an example and is not valid.

Particularly annoying is the example "jane@laptop.(none)" which appears in git-shortlog(1) as "jane@laptop[1].(none)", with note 1 saying:

	1. jane@laptop
	   mailto:jane@laptop

Fix this by escaping these email addresses with a leading backslash, to prevent Asciidoc expanding them as inline macros.

In the case of mailmap.txt, render the address monospaced so that it matches the block examples surrounding that paragraph.

Helped-by: Jeff King <peff@peff.net>
Signed-off-by: John Keeping <john@keeping.me.uk>
---
On Sun, Dec 16, 2012 at 07:04:05AM -0500, Jeff King wrote:
> Furthermore, the right way to suppress
> expansion of macros is with a backslash escape.
[snipped an example]
> I think it's a little less ugly
> than the "$$" quoting, but not by much. No clue if one is accepted by
> more asciidoc versions or not.
>From a quick reading of the Asciidoc changelog, I think backslash

escaping should be supported just as well as "$$" quoting, which leaves the minimal patch looking like this.

 Documentation/git-fast-import.txt | 2 +-
 Documentation/git-tag.txt         | 2 +-
 Documentation/mailmap.txt         | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)
Show changes to 3 files +3 −3

Documentation/git-fast-import.txt, Documentation/git-tag.txt, Documentation/mailmap.txt

diff --git a/Documentation/git-fast-import.txt b/Documentation/git-fast-import.txt
index d1844ea..68bca1a 100644
--- a/Documentation/git-fast-import.txt
+++ b/Documentation/git-fast-import.txt
@@ -427,7 +427,7 @@ they made it.
 
 Here `<name>` is the person's display name (for example
 ``Com M Itter'') and `<email>` is the person's email address
-(``cm@example.com'').  `LT` and `GT` are the literal less-than (\x3c)
+(``\cm@example.com'').  `LT` and `GT` are the literal less-than (\x3c)
 and greater-than (\x3e) symbols.  These are required to delimit
 the email address from the other fields in the line.  Note that
 `<name>` and `<email>` are free-form and may contain any sequence
diff --git a/Documentation/git-tag.txt b/Documentation/git-tag.txt
index 247534e..2f1c0c3 100644
--- a/Documentation/git-tag.txt
+++ b/Documentation/git-tag.txt
@@ -129,7 +129,7 @@ This option is only applicable when listing tags without annotation lines.
 CONFIGURATION
 -------------
 By default, 'git tag' in sign-with-default mode (-s) will use your
-committer identity (of the form "Your Name <your@email.address>") to
+committer identity (of the form "Your Name <\your@email.address>") to
 find a key.  If you want to use a different default key, you can specify
 it in the repository configuration as follows:
 
diff --git a/Documentation/mailmap.txt b/Documentation/mailmap.txt
index 288f04e..dd89fca 100644
--- a/Documentation/mailmap.txt
+++ b/Documentation/mailmap.txt
@@ -46,7 +46,7 @@ Jane Doe         <jane@desktop.(none)>
 Joe R. Developer <joe@example.com>
 ------------
 
-Note how there is no need for an entry for <jane@laptop.(none)>, because the
+Note how there is no need for an entry for `<jane@laptop.(none)>`, because the
 real name of that author is already correct.
 
 Example 2: Your repository contains commits from the following

-- 
1.8.0.2
Junio C Hamano· Dec 16, 2012, 18:48 UTC · re: John Keeping · lore

Re: [PATCH v2] Documentation: don't link to example mail addresses

John Keeping <john@keeping.me.uk> writes:
Show 31 quoted lines
> Email addresses in documentation are converted into mailto: hyperlinks
> in the HTML output and footnotes in man pages.  This isn't desirable for
> cases where the address is used as an example and is not valid.
>
> Particularly annoying is the example "jane@laptop.(none)" which appears
> in git-shortlog(1) as "jane@laptop[1].(none)", with note 1 saying:
>
> 	1. jane@laptop
> 	   mailto:jane@laptop
>
> Fix this by escaping these email addresses with a leading backslash, to
> prevent Asciidoc expanding them as inline macros.
>
> In the case of mailmap.txt, render the address monospaced so that it
> matches the block examples surrounding that paragraph.
>
> Helped-by: Jeff King <peff@peff.net>
> Signed-off-by: John Keeping <john@keeping.me.uk>
> ---
>
> On Sun, Dec 16, 2012 at 07:04:05AM -0500, Jeff King wrote:
>> Furthermore, the right way to suppress
>> expansion of macros is with a backslash escape.
> [snipped an example]
>> I think it's a little less ugly
>> than the "$$" quoting, but not by much. No clue if one is accepted by
>> more asciidoc versions or not.
>
> From a quick reading of the Asciidoc changelog, I think backslash
> escaping should be supported just as well as "$$" quoting, which leaves
> the minimal patch looking like this.

The patch looks reasonable to me, too. We were bitten by relying on the description in AsciiDoc documentation (which shows the state of their latest software) before; between the constructs that work, it is safer to use the older, more vanilla and more common one.

Thanks.
Junio C Hamano· Dec 17, 2012, 02:24 UTC · re: John Keeping · lore

Re: [PATCH v2] Documentation: don't link to example mail addresses

John Keeping <john@keeping.me.uk> writes:
Show 10 quoted lines
> diff --git a/Documentation/git-fast-import.txt b/Documentation/git-fast-import.txt
> index d1844ea..68bca1a 100644
> --- a/Documentation/git-fast-import.txt
> +++ b/Documentation/git-fast-import.txt
> @@ -427,7 +427,7 @@ they made it.
>  
>  Here `<name>` is the person's display name (for example
>  ``Com M Itter'') and `<email>` is the person's email address
> -(``cm@example.com'').  `LT` and `GT` are the literal less-than (\x3c)
> +(``\cm@example.com'').  `LT` and `GT` are the literal less-than (\x3c)
I seem to be getting
(<tt>`\cm@example.com'').  `LT</tt> and <tt>GT</tt> are the literal less-than (\x3c)

out of this part in the resulting HTML output, which is probably not what you wanted to see.

I have a feeling that it might be a better solution to stop using these pretty quotes. It's not like we use them a lot and a quick scanning of "git grep '``'" output seems to tell me that many of them should be `monospace output`, and the rest (mostly references to section headers) can be "Section".

Jeff King· Dec 17, 2012, 12:02 UTC · re: Junio C Hamano · lore

Re: [PATCH v2] Documentation: don't link to example mail addresses

On Sun, Dec 16, 2012 at 06:24:58PM -0800, Junio C Hamano wrote:
Show 12 quoted lines
> I seem to be getting
> 
> (<tt>`\cm@example.com'').  `LT</tt> and <tt>GT</tt> are the literal less-than (\x3c)
> 
> out of this part in the resulting HTML output, which is probably not
> what you wanted to see.
> 
> I have a feeling that it might be a better solution to stop using
> these pretty quotes.  It's not like we use them a lot and a quick
> scanning of "git grep '``'" output seems to tell me that many of
> them should be `monospace output`, and the rest (mostly references
> to section headers) can be "Section".

Typographically speaking, I would also be just as happy to use regular double-quotes throughout. But here's an example where they also caused a problem (which was fixed by moving to smart-quotes in f34e9ed):

  http://thread.gmane.org/gmane.comp.version-control.git/163067
-Peff
Junio C Hamano· Dec 17, 2012, 19:57 UTC · re: Jeff King · lore

Re: [PATCH v2] Documentation: don't link to example mail addresses

Jeff King <peff@peff.net> writes:
Show 20 quoted lines
> On Sun, Dec 16, 2012 at 06:24:58PM -0800, Junio C Hamano wrote:
>
>> I seem to be getting
>> 
>> (<tt>`\cm@example.com'').  `LT</tt> and <tt>GT</tt> are the literal less-than (\x3c)
>> 
>> out of this part in the resulting HTML output, which is probably not
>> what you wanted to see.
>> 
>> I have a feeling that it might be a better solution to stop using
>> these pretty quotes.  It's not like we use them a lot and a quick
>> scanning of "git grep '``'" output seems to tell me that many of
>> them should be `monospace output`, and the rest (mostly references
>> to section headers) can be "Section".
>
> Typographically speaking, I would also be just as happy to use regular
> double-quotes throughout.  But here's an example where they also caused
> a problem (which was fixed by moving to smart-quotes in f34e9ed):
>
>   http://thread.gmane.org/gmane.comp.version-control.git/163067
I agree with the typography argument.

Many monospaced fonts used in the windows where the programmers view the source of the docmentation, a 'normal' single quote does not slant by the same amount to the `opposite` quote, so ``a pair does not balance'' visually in the source. And the whole point of writing documentation in AsciiDoc is to keep the source readable. The use of pretty quote throws discards that primary benefit of choosing AsciiDoc in the first place.

Junio C Hamano· Dec 17, 2012, 05:29 UTC · re: John Keeping · lore

Re: [PATCH] Documentation: don't link to example mail addresses

John Keeping <john@keeping.me.uk> writes:
> I can't see any other uses of the "$$" quote in the documentation, so
> it's probably worth noting that I've tested this with Asciidoc 8.6.8,
> although I can't see anything in the changelog to indicate that
> Asciidoc's treatment of it has changed recently.
Thanks.
Show 18 quoted lines
>  Documentation/git-fast-import.txt | 2 +-
>  Documentation/git-tag.txt         | 2 +-
>  Documentation/mailmap.txt         | 2 +-
>  3 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/Documentation/git-fast-import.txt b/Documentation/git-fast-import.txt
> index d1844ea..05913cc 100644
> --- a/Documentation/git-fast-import.txt
> +++ b/Documentation/git-fast-import.txt
> @@ -427,7 +427,7 @@ they made it.
>  
>  Here `<name>` is the person's display name (for example
>  ``Com M Itter'') and `<email>` is the person's email address
> -(``cm@example.com'').  `LT` and `GT` are the literal less-than (\x3c)
> +(``$$cm@example.com$$'').  `LT` and `GT` are the literal less-than (\x3c)
>  and greater-than (\x3e) symbols.  These are required to delimit
>  the email address from the other fields in the line.  Note that
>  `<name>` and `<email>` are free-form and may contain any sequence
AsciiDoc 8.5.2 gives this:   
    asciidoc: WARNING: git-fast-import.txt: line 434: nested inline passthrough
Also in git-fast-import.1, there is this line:
    ... person\(cqs display name (for example `Com M Itter\(aq\(aq) and `<ema...
(notice two leftover `s); not that this is a new issue with this patch.
Which is puzzling, because the next input line formats better:
    ...  email address (\(lqcm@example\&.com\(rq)\&. ...

Other two hunks for git-tag and mailmap seem to give us a better results.

John Keeping· Dec 17, 2012, 09:44 UTC · re: Junio C Hamano · lore

Re: [PATCH] Documentation: don't link to example mail addresses

On Sun, Dec 16, 2012 at 09:29:24PM -0800, Junio C Hamano wrote:
Show 22 quoted lines
> John Keeping <john@keeping.me.uk> writes:
>>  Here `<name>` is the person's display name (for example
>>  ``Com M Itter'') and `<email>` is the person's email address
>> -(``cm@example.com'').  `LT` and `GT` are the literal less-than (\x3c)
>> +(``$$cm@example.com$$'').  `LT` and `GT` are the literal less-than (\x3c)
>>  and greater-than (\x3e) symbols.  These are required to delimit
>>  the email address from the other fields in the line.  Note that
>>  `<name>` and `<email>` are free-form and may contain any sequence
> 
> AsciiDoc 8.5.2 gives this:   
> 
>     asciidoc: WARNING: git-fast-import.txt: line 434: nested inline passthrough
> 
> Also in git-fast-import.1, there is this line:
> 
>     ... person\(cqs display name (for example `Com M Itter\(aq\(aq) and `<ema...
> 
> (notice two leftover `s); not that this is a new issue with this patch.
> 
> Which is puzzling, because the next input line formats better:
> 
>     ...  email address (\(lqcm@example\&.com\(rq)\&. ...

I wonder if you're hitting this bug, which is listed as fixed in AsciiDoc 8.6.0 [1]:

    FIXED: Sometimes multiple double quoted text elements in the same
    paragraph were mistakenly seen as starting with an inline literal.

As you said in your other message, it may be better to avoid using the pretty quotes, although I don't think just changing them to `monospace output` is sufficient as monospace output is not distinguished in the default man page output. IMHO that's necessary for any examples that contain whitespace. Perhaps a patch series turning MAN_BOLD_LITERAL on by default and then changing these examples to `monospace output`?

John
[1] http://www.methods.co.nz/asciidoc/CHANGELOG.html#_version_8_6_0_2010_08_16

← back to recent threads