Đoàn Trần Công Danh <congdanhqx@xxxxxxxxx> writes:
> Subject: Re: [PATCH v2 3/5] mailinfo: skip quoted CR on user's wish
Nothing wrong per-se, but "on user's wish" feel somewhat bizarre.
Perhaps
mailinfo: allow skipping quoted CR
or something along that line?
> In previous change, we've turned on warning for quoted CR in base64
> encoded email. Despite those warnings are usually helpful for our users,
> they may expect quoted CR in their emails.
>
> Let's give them an option to turn off the warning completely.
>
> Signed-off-by: Đoàn Trần Công Danh <congdanhqx@xxxxxxxxx>
> ---
> Documentation/git-mailinfo.txt | 18 +++++++++++++++++-
> builtin/mailinfo.c | 8 ++++++--
> mailinfo.c | 21 ++++++++++++++++++++-
> mailinfo.h | 8 ++++++++
> t/t5100-mailinfo.sh | 6 ++++--
> 5 files changed, 55 insertions(+), 6 deletions(-)
>
> diff --git a/Documentation/git-mailinfo.txt b/Documentation/git-mailinfo.txt
> index d343f040f5..c776b27515 100644
> --- a/Documentation/git-mailinfo.txt
> +++ b/Documentation/git-mailinfo.txt
> @@ -9,7 +9,7 @@ git-mailinfo - Extracts patch and authorship from a single e-mail message
> SYNOPSIS
> --------
> [verse]
> -'git mailinfo' [-k|-b] [-u | --encoding=<encoding> | -n] [--[no-]scissors] <msg> <patch>
> +'git mailinfo' [-k|-b] [-u | --encoding=<encoding> | -n] [--[no-]scissors] [--quoted-cr=<action>] <msg> <patch>
This line is getting really crowded. Perhaps it is time to do
'git mailinfo' [<options>] <msg> <patch>
like other Git subcommands with too many options? Certainly it can
be done after the dust settles from this entire series as a follow up
clean-up patch.
> @@ -89,6 +89,22 @@ This can be enabled by default with the configuration option mailinfo.scissors.
> --no-scissors::
> Ignore scissors lines. Useful for overriding mailinfo.scissors settings.
>
> +--quoted-cr=<action>::
> + Action when processes email messages sent with base64 or
> + quoted-printable encoding, and the decoded lines end with CR-LF
s/with CR-LF/with a CRLF/
> + instead of a simple LF.
> ++
> +The valid actions are:
> ++
> +--
> +* `nowarn`: Git will do nothing with this action.
s/with this action./when such a CRLF is seen./ perhaps?
> +* `warn`: Git will issue a warning for each message if such CR-LF is
s/such CR-LF/such a CRLF/
> diff --git a/builtin/mailinfo.c b/builtin/mailinfo.c
> index b309badce5..1d600263cb 100644
> --- a/builtin/mailinfo.c
> +++ b/builtin/mailinfo.c
> @@ -9,7 +9,7 @@
> #include "mailinfo.h"
>
> static const char mailinfo_usage[] =
> - "git mailinfo [-k | -b] [-m | --message-id] [-u | --encoding=<encoding> | -n] [--scissors | --no-scissors] <msg> <patch> < mail >info";
> + "git mailinfo [-k | -b] [-m | --message-id] [-u | --encoding=<encoding> | -n] [--scissors | --no-scissors] [--quoted-cr=<action>] <msg> <patch> < mail >info";
It is surprising that we haven't switched this to parse_options().
It of course is outside the scope of this series, but from a cursory
look of its option parsing loop, it looks like a trivial improvement
to make.
> @@ -43,7 +43,11 @@ int cmd_mailinfo(int argc, const char **argv, const char *prefix)
> mi.use_scissors = 0;
> else if (!strcmp(argv[1], "--no-inbody-headers"))
> mi.use_inbody_headers = 0;
> - else
> + else if (skip_prefix(argv[1], "--quoted-cr=", &str)) {
> + mi.quoted_cr = mailinfo_parse_quoted_cr_action(str);
> + if (mi.quoted_cr == quoted_cr_invalid_action)
> + usage(mailinfo_usage);
This is not all that helpful, given that mailinfo_usage[] only says
<action> without saying what the supported values are, and the
message does not make it clear it was issued while looking at the
--quoted-cr option.
At least, something like
if (mi.quoted_cr == quoted_cr_invalid_action)
die("--quoted-cr=%s: invalid action", str);
would be more palatable, but I wonder if mailinfo_parse_quoted_cr_action()
should have an option to die with the list of actions it knows about
in a message.
> diff --git a/mailinfo.c b/mailinfo.c
> index 713567f84b..fe7ffd01d0 100644
> --- a/mailinfo.c
> +++ b/mailinfo.c
> @@ -1040,7 +1040,8 @@ static void handle_filter_flowed(struct mailinfo *mi, struct strbuf *line,
>
> static void summarize_quoted_cr(struct mailinfo *mi, int have_quoted_cr)
> {
> - if (have_quoted_cr)
> + if (have_quoted_cr
> + && mi->quoted_cr == quoted_cr_warn)
Existing code in this file prefers to split a multi-line statement
after sequence point like &&, ||, etc., not before.
> warning("quoted CR detected");
> }
>
> @@ -1221,9 +1222,19 @@ int mailinfo(struct mailinfo *mi, const char *msg, const char *patch)
> return mi->input_error;
> }
>
> +enum quoted_cr_action mailinfo_parse_quoted_cr_action(const char *action)
> +{
> + if (!strcmp(action, "nowarn"))
> + return quoted_cr_nowarn;
> + else if (!strcmp(action, "warn"))
> + return quoted_cr_warn;
> + return quoted_cr_invalid_action;
> +}
OK.
> static int git_mailinfo_config(const char *var, const char *value, void *mi_)
> {
> struct mailinfo *mi = mi_;
> + const char *str;
>
> if (!starts_with(var, "mailinfo."))
> return git_default_config(var, value, NULL);
> @@ -1231,6 +1242,13 @@ static int git_mailinfo_config(const char *var, const char *value, void *mi_)
> mi->use_scissors = git_config_bool(var, value);
> return 0;
> }
> + if (!strcmp(var, "mailinfo.quotedcr")) {
> + git_config_string(&str, var, value);
> + mi->quoted_cr = mailinfo_parse_quoted_cr_action(str);
> + if (mi->quoted_cr == quoted_cr_invalid_action)
> + die(_("bad action '%s' for '%s'"), str, var);
> + free((void *)str);
> + }
Here, it is more reasonable. It still does not say what actions are
accepted, but at least the user learns where our displeasure comes
from.
> /* perhaps others here */
> return 0;
> }
