todo.txt

- cleanup hard-coded names: "web"

- cleanup: Comic::not_yet_published should not consider locations.

- cleanup: always use lower case language names in templates; accept any
  case in comic metadata for backwards compatibility?

- feature: reject slashes and other funny characters (backslashes? colons?
  dotdot?) in all folder names in settings (File::Path::makepath CAN handle
  slashes and makes a tree, as expected), but certain characters may or may
  not work if local OS and web server OS differ, and some code assumes only
  one level of directories (e.g., all template processing that defines the
  root variable).

- feature: more email customization: templates (per language), subject
  prefix, EHLO host name, max recipients per message

- defect: Mastodon 500 error when posting images

- cleanup: rework variables available for templates: standardize names and
  case, pick decent names, DRY up them up, document them

- feature: (output) modules dependency graph, to easily add new output
  modules, and to decide to skip output (and check) modules based on their
  dependencies being up to date

- should the names in the JSON configuration file be case-insensitive? Could
  be confusing when case actually matters, e.g., for Perl modules.
  If case matters, make sure it's used consistently. Language names are
  still used inconsistently.

- refactor: Perl: use feature qw(signatures); no warnings
  qw(experimental::signatures); or Function::Parameters, Method::Signatures

- defect: twitter with @someone does not get tweeted for some reason

- feature: list comics (svg files) under series headings (hideable?) in
  backlog, to see which comics are in a series; extend to tags and who

- feature: put unpublished comics in the right position in their series

- feature: define comic order by refering the previous or next comic,
  instead of having a date, so that inserting one for current events doesn't
  ripple through the backlog (and still maintaining a regular / weekly
  schedule); or provide a script that inserts a comic by moving all others
  (complicated, needs to know on which dates comics are published, and would
  still mess up comics that need to stay in place for e.g., holidays)

- feature: allow adding (invisible) watermarks

- feature: transcript needs to check all frames for ordering, not just rows;
  that should get rid of space padding in texts

- feature: handle more frame varieties, e.g., long frame over 2 rows on the
  side

- feature: allow comic to be in more than one series

- feature: export svg instead of png (for a truly fluent / responsive web
  design), but would also need to offer the font for download, I guess? or
  chop the svgs or pngs up per frame (ugly, relies on either table-based
  HTML layout, CSS magic beyond my skills, or chopping comics by rows only)

- feature: minimize svg file sizes (e.g., with svgo), if svg export is supported

- feature: generate latex code for pdf, or generate pdf right away (this may
  be just a templating issue)

- feature: liner notes with comments in pdf, suck in other interesting
  documentation. This can pull stats per comics from git: how long between
  creation and publication, how long was the backlog each Fri, how big was
  the ideas file each week, how often was the comic changed (after being
  published).

- feature: sizemap should have colors depending on where the comic was
  published, not published vs unpublished

- feature: make more sizemap svg attributes configurable

- feature: add a thumbnail / preview for really large images (posters),
  especially in RSS

- check for date holes (why? are those bad? only in the future? but I could
  schedule a xmas or Fri 13 comic ahead of time and then have holes and
  that's ok); warn in backlog if no comic for next Friday

- feature: more social media: Openbook, Ello, Diaspora, inrupt.com
  (via Zapier?)

- feature: use inkscape shell to export in bulk (could be faster than
  creating multiple inkscape processes), would be great to avoid writing
  temp svg files if inkscape accepts the svg on stdin

- publish on CPAN, add badge to README.md like this for Minilla:
  [![MetaCPAN Release](https://badge.fury.io/pl/Minilla.svg)](https://metacpan.org/release/Minilla)
  https://badge.fury.io/for/pl

- feature: check that speaker names all show up in the "who" metadata (but
  be careful about additional text, like "Paul, opening a beer:")

- performance: write qr code file only when writing the pngs, i.e., when the
  source has changed (really only need this if the generated title / qr code
  file name has changed, but this is harder to detect)

- performance: speed up with multiple threads for checking and converting
  svg files (1 thread per comic, then wait till all are done; reuse threads
  for writing per-comic pages, when all next/prev links are known)

- feature: verbose and debug flags to print what's going on (helpful for
  configurable checks and actions)

- refactor: Twitter module uses outdated Twitter 1.0 API; switch to Twitter::API
  or do the http calls manually; rename to X.

- feature: check 2FA for Reddit and Twitter

- performance: when modifying only the latest comic (or an unpublished
  comic), don't recreate all html pages (only if the latest comic file name
  (title) changes, links need to get updated). This should speed up final
  touches on the latest comic

- defect: frame detection for text outside of a single frame (comic title)
  does not seem to treat that text as after anything within the frame. The
  whole frame handling code needs a close look.

- feature: should frame detection at the top of the comic use frames of the
  next (first) row? that would allow having a title above the frame, like
  "yesterday" and "today".

- defect: does it treat text outside of a meta layer as speaker label just
  because it ends in a colon (motivation.svg)? it probably should not.

- feature (defect?): transcript check needs to not just grab all texts in
  order but also respect layers: don't treat regular text as speaker
  indicator just because it ends in a colon (could be a before / after style
  comic).

- check if meta text is not in a meta layer? by checking for colons? but
  trailig colons should be ok.

- feature: upload to e.g., imgur.com and then post a link to that on reddit.
  This may be interesting for people not running their own website, and for
  subreddits that only accept file hosting on certain sites (like imgur).

- feature: a rss / atom feed module that doesn't require the user to provide
  a template and hence know the feed format.

- feature: add a simple tag cloud by default if the user doesn't provide a
  tag cloud index template.

- feature: sitemap should work without people having to define a template

- feature: if we keep the sitemap via template, filter the list passed to
  each template instead of passing the filter notFor function for easier use

- feature: post "classics" to social media if there is no new comic

- feature: complain about unknown settings to catch typos (done for Generators,
  missing in Checks)

- clean out backlog: when a comic goes from backlog to published, remove its
  html file and qr code image from the backlog directory

- feature: check text positions as double, so that you can move texts
  minimally to control the transcript without visual impact

- performance: does it make a difference to remove unneeded top level layers
  (other languages, turned invisible in Inkscape) instead of switching them
  to hidden? Files would be smaller (is disk space an issue anyway?), but
  would converting them be faster?

- replace vars{'tags'} and vars{'tagsOrder'} in backlog with an ordered hash
  if Template supports this nicely

- feature: backlog currently appends language name to tags and series, it
  could instead return hashes with languages as keys. On the one hand the
  template gets more complicated (one more nested loop), on the other hand
  it's cleaner and would allow to show / hide tags or series per language

- refactor: break up Comic::Out::Backlog into the actual backlog part and
  metadata (e.g., tags) aggregation overview

- feature: support aggregating metadata that's not per-language in the
  backlog (author, contrib, published when / where)

- feature: default subreddits should be per language hash, both in config
  and in comic metadata (support meta data single value for backward
  compatibility)

- feature: Comic::Social::Twitter needs a svg mode

- performance: write everything needed to publish (png dimensions,
  transcript), to a cache file (database?) so that publishing code doesn´t
  need to scan all comics again (may be tricky with generating pages for
  archive, tags, etc)

- feature: see metadata should try to resolve path relative to current
  comic

- feature: allow copying static files and pipe them through a Template to
  modify bits in mostly static files, like a copyright year in an index.html
  Maybe this could be done in server-side includes.

- feature: deduct output path from fragments, like outdir / base dir plus
  language placeholders, for less repetitive configuration files

- feature: warnings as errors

- feature: warn about spellcheck ignore words defined in the comic but not
  used

- feature: specify qr code image size (e.g., in pixels)

- feature: allow picking up Dublin core entities from Inkscape's document
  properties: Creator = Author and Artist, Rights = Copyright; use for PNG
  metadata and make available in templates

- feature: pick up Contributors from Inkscape document properties, make
  available in templates (merge with current contributors JSON metadata);
  but then again this could be confusing if metadata is spread over Inkscape
  regular properties and metadata JSON.

- feature: use relative links for HtmlLink for easy offline use

- doc: add example comic project with comics, templates, cron script,
  settings.json

- feature: disable one or all checks on the command line... sounds hackish,
  after all, the configuration controls features like checks, or the entry
  point (Comics::generate vs Comics::upload)

- feature: allow putting a ssh key for rsync in the config, so that a dumb scp
  doesn't skip a .ssh directory or .ssh-key.pem file (when switching between
  computers)

- feature: decouple archive comic pages from index.html; this could make for
  easier templates if the main page differs a lot from each comic's page

- feature: on the publish day: warn (email, ...) if there is no comic
  scheduled for today (maybe only if there are comics in the backlog);
  probably warn the day before already

- feature: caldav support to show planned comics in a calendar

- feature: calendar widget in backlog to show for which days a comic is
  scheduled and where holes are

- feature: cli tool to add a new language to an existing comic: language
  layers as in another language, empty JSON attributes for metadata

- feature: multiple pages comics (for print or ebooks)

- feature: patreon integration: designate patreon only-comics, or upload to
  patreon earlier than to website

- feature: substack integration

- feature: blogger.com integration

- feature: medium.com integration

- feature: Wordpress integration

- feature: Lemmy integration

- feature: Threads integration

- feature: git integration: host your comic on github? Publishing module could
  commit to a (separate?) repository, push changes, and the web server checks
  out from git.

- feature: warn about missing or bad metadata
  description: should not be empty
  published where: not empty? certain well-known locations? (would only
  make sense it these locations are configured somewhere, but most don't
  need configuration)

- DRY up root vs settings->Paths->siteComics vs indexAdjust in templates

- DRY up backlogPath (has the language) vs Settings->Paths->backlog

- in Comic.pm:
        $self->{urlEncoded}{$language} = 'https://' .
            Net::IDN::Encode::domain_to_ascii($domain) .
            "/$self->{siteComicsPath}" .
            uri_escape_utf8($html_file);
  why not uri_escape siteComicsPath? Could be non-ascii.

- feature: support a single domain setup with top-level language directories