App-PFT

 view release on metacpan or  search on metacpan

bin/pft-edit  view on Meta::CPAN

#!/usr/bin/perl
#
# Copyright 2014-2019 - Giovanni Simoni
#
# This file is part of PFT.
#
# PFT is free software: you can redistribute it and/or modify it under the
# terms of the GNU General Public License as published by the Free
# Software Foundation, either version 3 of the License, or (at your
# option) any later version.
#
# PFT is distributed in the hope that it will be useful, but WITHOUT ANY
# WARRANTY; without even the implied warranty of MERCHANTABILITY or
# FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
# for more details.
#
# You should have received a copy of the GNU General Public License along
# with PFT.  If not, see <http://www.gnu.org/licenses/>.

=encoding utf8

=head1 NAME

pft edit - Edit an entry

=head1 SYNOPSIS

    B<pft edit -P> [options] title of your page
    B<pft edit -B> [options] title of your blog page
    B<pft edit -M> [options] title of your month page
    B<pft edit -T> [options] title of your tag page

=head1 DESCRIPTION

The L<pft-edit(1)> command allows to conveniently edit a text entry in a
I<PFT> site. It takes care of creating a header for the new entries, to
position files correctly within your I<PFT> filesystem tree, and to maintain
the filesystem location consistent depending on the type of content.

All entries are stored under the C<ROOT/content> directory, where C<ROOT>
is the base path of your I<PFT> site. Each entry has the same format:
a text file encoded as by locale, composed by a I<YAML> header concatenated
with Markdown text (more details later in this document).

There can be different kind of entries:

=over

=item Regular Pages

Regular pages do not declare a date in their header. They get positioned
in C<ROOT/content/pages>. Their file name depends on the page title.

For example, a page entitled I<Hello World> is stored as:

    ROOT/content/pages/hello-world

=item Blog pages

Blog pages declare a full date (year, month and day) in their header. They
are positioned in the C<ROOT/content/blog/YYYY-MM> directory (where
C<YYYY> and C<MM> correspond to the declared year and month, respectively).
The file name of a blog page starts with a C<DD-> prefix, where C<DD>
corresponds to the declared day. The remaining part of the file name depends
on the entry title.

For example, an entry entitled I<Hello World> edited the 12th of September 2015
is stored as:

    ROOT/content/blog/2015-09/12-hello-world

=item Month pages

Month pages are meant as I<entries summarizing a whole month>. They are
like blog pages, but the date defined in their headers lack of the day
part (it is replaced by C<*> character). They are stored as
C<ROOT/content/blog/YYYY-MM.month> (where C<YYYY> and C<MM> represent the
declared year and month).

For example, the month page for September 2015 is stored as:

    ROOT/content/blog/2016-09.month

=item Tag pages

All pages can optionally declare one or more tags. A tag in turn can be
optionally associated to a tag page.

For example, if your site has have a number of entries about cuisine, some
of them might be tagged with the C<chicken> tag. You may want to have a tag
page entitled C<chicken>, where you give a description of what I<chicken> is
(just in case someone does not know).

Tag pages are stored in C<ROOT/content/tags>, and their file name depends
on the tag title.

For example, a page containing a description of the tag I<chicken> is stored
as:

    ROOT/content/tags/chicken

=back

bin/pft-edit  view on Meta::CPAN

=item B<--author>=I<name> | B<-a> I<name>

Defines the author for the entry.

This flag gets honored only when generating a header for the entry, that
is if the file does not exist yet.

=item B<--tag>=I<tag> | B<-t> I<tag>

This flag gets honored only when generating a header for the entry, that
is if the file does not exist yet.

=item B<--resume> | B<-r>

Resume editing of a blog entry.

This option implies B<-B>. The blog entry resumed for editing corresponds by
default to the latest in chronological order.

A date can be specified via the B<-y>, B<-m> and B<-d> options. If the date
is only partially specified, that is not all of B<-y>, B<-m> and B<-d> are
provided, the missing bits get completed with the current date.  If no date
is specified, B<-r> is equivalent to B<--back 0>.

If more then one entry exists for the specified date, the command fails and
the user is explicitly required to specify a selection via the B<--select>
option.

=item B<--back>=I<count> | B<-b> I<count>

Resume editing an old entry. The supplied parameter defines how many days
we should go back in history since the last day we have an entry for (0
means the most recent day, 1 means the second to last).

=item B<--editor> I<command>

Specify an editor to use.

The editor can be specified by name (e.g. C<vim>) or as a shell command,
where C<%s> is replaced with the file name (e.g. C<vim [options] %s>).

This flag overrides the C<$EDITOR> environment variable and the
C<system.editor> setting in the main configuration file.

This option is ignored if B<--stdin> is used.

=item B<--stdin>

Retrieve content from stdin.

The content file is selected according to the given options, and the content is
retrieved from standard input.  If the file does exist it gets created with a
reasonable header. If the file exists but the header is corrupted, the whole
file gets replaced.  If the file exists and the header is well formed, the
header is maintained and the remaining part of the file is replaced with content
retrieved from standard input.

If this option is in use the B<--editor> is ignored.

This editing mode handles input and output encoding according to the current
locale, under the assumption that your content is text data.

=item B<--append>

As for B<--stdin>, but append content instead of replacing it.

If the content file does not exist or is empty, and header will be
automatically defined depending on the given command options.

If the header exists but is corrupted, the command will issue a warning.
No corrective action will be taken.

=item B<--select>=I<id>

Select a certain entry by index, in case of ambiguity.

This option applies when multiple files are matching the parameters for file
selection, and a choice is required.  In these situations, if B<--select> is
not provided the L<pft-edit(1)> command will fail, and the use of
B<--select> will be explicitly requested by an error message.

=item B<--raw>

This option is useful only if B<--stdin> or B<--append> are specified.

If the B<--raw> option is used the header is not maintained: C<pft edit
--stdin --raw> command will behave like C<cat E<gt> $content_file>, while
C<pft edit --append --raw> will behave like C<cat E<gt>E<gt> $content_file>.
The path of C<$content_file> is determined by the previously described file
naming rules.

=item B<--help> | B<-h>

Shows this help.

=back

=head2 Editing

Content entries are in encoded text format.

The expected encoding for the file corresponds to the C<locale>.  On modern
systems I<UTF-8> is the more popular encoding, but you can use the
L<locale(1)> command to figure out.

Each file starts with a header in I<YAML> format. The header is followed
by a line with three dashes (C<--->) which marks the beginning of the
actual content. The content will be parsed as I<Markdown> when the site is
compiled.

The header of a content entry is created automatically by L<pft-edit(1)>
when the accessed entry does not exist. The file gets then opened with a
text editor. The C<$EDITOR> environment variable will be honored unless an
editor is defined in the C<pft.yaml> configuration file (see
L<pft-init(1)>). You may also specify a different editor using the
B<--editor> command line option.

When the editor is terminated the file content is evaluated.  A warning will
be issued if the header is invalid.  If the file completely empty (zero
bytes) it will be removed from the filesystem.  If the header is valid but
not consistent with the position in the filesystem (e.g. the date was
manually changed) the file position is updated according to the header.

=head2 The content header

Each entry starts with a header in I<YAML> format. The header defines the
properties of the entries. Valid properties with their descriptions follow
in this section.

=over

=item I<Author>

The author of the entry. Any string can be entered as value. The default
value of it is taken from the C<pft.yaml> configuration file.

=item I<Date>

The date associated with the entry. This field is required only for blog
entries. Changing the value while editing with L<pft-edit(1)> will result to
the position in the filesystem to be adapted consistently.

=item I<Options>

Some optional properties of the entry:

=over

=item I<hide>

I<This feature is currently broken>.

If set to 1 the entry is not going to be built by L<pft-make(1)>.

The intention is to temporarily hide an entry (e.g. if it is incomplete).

=item I<template>

Each page can specify a different template to use by means of this field.
It is by default defined as C<~> (in I<YAML> that is I<null>) meaning that
the page will use the global template as defined in the C<pft.yaml>
configuration file.

=back

bin/pft-edit  view on Meta::CPAN

                date => PFT::Date->from_spec(%datespec)->derive(d => undef),
            )
        } elsif ($opts{B}) {
            $hdr = PFT::Header->new(
                title => join(' ', @ARGV) || 'Today',
                author => $conf->{site}{author},
                tags => $opts{tags} || [],
                date => eval{ PFT::Date->from_spec(%datespec) } || do {
                    say STDERR 'Invalid date: ', $@ =~ s/ at.*$//rs;
                },
            )
        } elsif ($opts{P}) {
            die 'mandatory title' unless @ARGV;
            $hdr = PFT::Header->new(
                title => join(' ', @ARGV),
                author => $conf->{site}{author},
                tags => $opts{tags} || [],
            )
        } else { die "unhandled case? This is a bug" }

        $entry = $tree->content->new_entry($hdr)
    }
    1;  # can be 1 only if there was no error.
}
or $@ && do {
    say STDERR 'Editing entry: ', $@ =~ s/at .*$//rs;
    exit 6
};

sub edit_file {
    my $path = $entry->path;
    my $editor = $opts{editor}
              || $conf->{system}{editor}
              || $ENV{EDITOR}
              || do {
        say STDERR "Cannot infer editor. Try setting env EDITOR or to";
        say STDERR "define it in configuration file (system -> editor)";
        exit 5
    };
    if ($editor =~ s/(?<!%)%s/$path/g) {
        system($editor)
    }
    else {
        system($editor, $path)
    }
}

sub feed_file {
    my $mode = '>';
    my $skip_header = $opts{raw} || !defined $hdr;
    if ($opts{stdin}) {
        die "Supported --stdin or --append, not both" if $opts{append};
    }
    elsif ($opts{append}) {
        if (-f $entry->path) {
            $mode .= '>';
            $skip_header = 1;
        }
    }

    open my $out, "$mode:encoding(locale)", $entry->path
        or die 'Cannot open ', $entry->path, ": $!";
    $hdr->dump($out) unless $skip_header;
    print $out <STDIN>;
    close $out;
}

eval {
    if ($opts{stdin} or $opts{append}) {
        feed_file
    }
    else {
        edit_file
    }

    if ($entry->exists) {
        if ($entry->void) {
            say STDERR "Removing empty entry at ", $entry->path;
            $entry->unlink;
        }
        else {
            $entry->make_consistent;
        }
    }
} or $@ && do {
    say STDERR "After editing: ", $@ =~ s/at .*$//sr;
    exit 7
}