ioccc/bin

bin

The bin directory holds tools that build files, such as HTML content, for the official IOCCC web site.

For HTML content, the bin directory tools make use of HTML fragments from the inc directory as well as various JSON files and other content from the IOCCC GitHub repo.

bin/ tools

all-run.sh

Run a command on all winning entries.

For example:

bin/all-run.sh -v 3 bin/quick-readme2index.sh -v 1

Or for example:

bin/all-run.sh -v 3 bin/readme2index.sh -v 1

all-years.sh

Run a command on all IOCCC years.

For example:

bin/all-years.sh -v 1 bin/gen-year-index.sh -v 1

Or for example:

bin/all-years.sh -v 1 bin/chk-entry.sh

chk-entry.sh

Check that files required by an entry match the entry's manifest as found in .entry.json.

For example:

bin/chk-entry.sh 2020/ferguson1

filelist.entry.json.awk

Generate the list of files in the entry's manifest, .entry.json.

For example:

awk -f bin/filelist.entry.json.awk 2020/ferguson1/.entry.json

gen-authors.sh

Generate the top level /authors.html page.

Usage:

bin/gen-authors.sh -v 1

gen-location.sh

Generate the top level /location.html page.

Usage:

bin/gen-location.sh -v 1

gen-other-html.sh

Generate entry HTML files (other than README.md to index.html) from markdown files.

Usage:

bin/gen-other-html.sh -v 1

gen-sitemap.sh

Generate an XML sitemap for the IOCCC web site.

Usage:

bin/gen-sitemap.sh -v 1

gen-status.sh

Generate status.json according to the modification dates of status.json and news.html.

Without argument, the contest_status is unchanged.

Usage:

bin/gen-status.sh -v 1

To force the contest_status to be closed:

bin/gen-status.sh -v 1 closed

To force the contest_status to be open:

bin/gen-status.sh -v 1 open

To see other valid statuses:

bin/gen-status -h

gen-top-html.sh

Generate a number of top level HTML pages for the IOCCC web sites.

Usage:

bin/gen-top-html.sh -v 1

Examples of top level HTML pages built by this tool include:

• CODE_OF_CONDUCT.html
• archive/historic/index.html
• bin/index.html
• bugs.html
• contact.html
• faq.html
• inc/index.html
• index.html
• thanks-for-help.html
• judges.html
• news.html
• thanks-for-help.html

gen-year-index.sh

Generate an index.html page for an given IOCCC year.

Usage:

bin/gen-year-index.sh -v 1 2020

gen-years.sh

Generate the top level /years.html page.

Usage:

bin/gen-years.sh -v 1

md2html.sh

This is the primary tool that forms IOCCC generated HTML content from markdown files (permanent markdown files or temporarily generated markdown files) and HTML fragments from the inc directory.

The ../inc/md2html.cfg configuration file is used by md2html.sh to drive the generation process.

output-index-author.sh

Output author's or authors' related HTML details for an entry's index.html page.

output-index-inventory.sh

Output the inventory in HTML form for an entry's index.html page.

output-year-index.sh

Output markdown of wining entry links for a given year.

pandoc-wrapper.sh

Wrapper tool to run pandoc.

quick-readme2index.sh

Runs readme2index.sh if the entry directory does not have a non-empty index.hmtl file, or if either .entry.json or index.html is newer than the index.hmtl file.

This is useful when only a few entries have been modified (resulting in an updated .entry.json file) or if the index.html of a few entries has been changed.

While the readme2index.sh take a few seconds to run, when applied to 300+ entries, the extra time can add up.

If only a few index.hmtl files need updating, then this command will only briefly pause while the slightly longer readme2index.sh is run:

bin/all-run.sh -v 3 bin/quick-readme2index.sh -v 1

NOTE: This command assumes that the relative modification times for index.hmtl, .entry.json, and index.html are correct. If in doubt, use:

bin/all-run.sh -v 3 bin/readme2index.sh -v 1

readme2index.sh

Convert an entry README.md into entry directory index.html.

status2html.sh

Convert status.json into HTML.

This tool is a 'before tool' (-b tool) that is intended to be used by bin/gen-status.sh.

sgi.sh

Sort .gitignore content from stdin to stdout.

We sort with lines starting with # first. We sort with lines starting with * second. We sort with lines that do not start with [#!*] third. We sort with lines starting with ! fourth.

This tool is used by sort.gitignore.sh.

sort.gitignore.sh

Sort a .gitignore in a entry directory.

Usage:

bin/sort.gitignore.sh -v 1 YYYY/dir

Suggested usage:

bin/all-run.sh bin/sort.gitignore.sh -v 1

subst.default.sh

Print default substitutions.

subst.entry-index.sh

Print substitutions for an entry index.html.

subst.entry-navbar.awk

Output substitutions for navbar on behalf of an entry.

subst.year-index.sh

Print substitutions for a year level index.html.

subst.year-navbar.awk

Output substitutions for navbar on behalf of a year level index.html.

tar-entry.sh

Form a compressed tarball for an entry.

Usage:

bin/tar-entry.sh -v 1 YYYY/dir

Suggested usage:

bin/all-run.sh -v 3 bin/tar-entry.sh -v 1 -W

tar-all.sh

Form the compressed tarball for all IOCCC years and all IOCCC entries.

Usage:

bin/tar-all.sh -v 1

tar-year.sh

Form a compressed tarball for an IOCCC year.

Usage:

bin/tar-year.sh -v 1 YYYY

Suggested usage:

bin/all-years.sh -v 3 bin/tar-year.sh -v 1 -W

How IOCCC HTML content is built

The md2html.sh tool is the primary tool that is used to form all IOCCC related HTML pages for the official IOCCC web site.

Nearly all IOCCC related HTML pages are built from markdown files, from either permanent markdown files or temporary generated markdown files, as well as HTML fragments from the inc directory.

Most HTML content is built from permanent markdown files, such as a README.md markdown file found in each entry directory. Some HTML content are generated from temporary markdown files. These temporary markdown files are produced by tools in the bin directory and exist only while the tool is running.

In addition to converting markdown to HTML, the canonical way that HTML content is built uses, by default, files from the inc directory, of the file name form:

*.default.html

By using command line options of the form:

-H phase=name

where phase is the name of an HTML phase, a non-default file may be used. For example:

-H topnav=up2index

will cause inc/topnav.up2index.html instead of inc/topnav.default.html to be used during the topnav HTML phase.

If name is dot (i.e., .), then the given HTML phase is skipped. For example:

-H footer=.

will cause no HTML content to be produced during the footer HTML phase.

HTML phases

The following HTML phase files are used to build HTML content:

0. inc/top.default.html
1. inc/head.default.html
2. inc/body.default.html
3. inc/header.default.html
4. inc/topnav.default.html
5. inc/begin-row.default.html
6. inc/begin-leftcolumn.default.html
7. inc/sidenav.default.html
8. inc/end-leftcolumn.default.html
9. inc/begin-rightcolumn.default.html
10. inc/before-content.default.html

Phases 11-19 are reserved for future use.

20. ((insert 'before tool' output))
21. ((insert 'pandoc wrapper tool' output))
22. ((insert 'after tool' output))

Phases 23-29 are reserved for future use.

30. inc/after-content.default.html
31. inc/end-rightcolumn.default.html
32. inc/end-row.default.html
33. inc/footer.default.html
34. inc/bottom.default.html

Phases 35-39 are reserved for future use.

In the above HTML phase numbers, except during HTML phase numbers 20-29, symbols of the form %%TOKEN%% are substituted.

Any %%TOKEN%% printed by the 'before tool' (-b tool), the 'pandoc wrapper tool' (-p tool), and the 'after tool' (-a tool) are ignored (i.e., not substituted).

See the tool readme2index.sh for an example of how HTML phases are implemented.

getopt phase processing of the command line

The command options are evaluated in the following getopt phases:

getopt phase 0

In getopt phase 0 we parse command line options and save all arguments for the end of the final command line.

Depending on the program, the argument at the end is directory under topdir, or it is a file under topdir.

getopt phase 0 is the only getopt phase where -d topdir and -c md2html.cfg may be used.

getopt phase 1

In getopt phase 1 we execute each -o "output tool" from getopt phase 0, parsing the output as a command line.

The -o "output tool" may be given optstr from "-O tool=optstr", followed by the phase 0 argument.

The -o "output tool" is not allowed to output another "-o tool" or a "-O tool=optstr". Instead a -o "output tool" may execute another -o "output tool" and merge the output into its own.

getopt phase 2

In getopt phase 2 we parse the cfg_options from the first md2html.cfg line matched by a saved argument.

The match is made with the (possibly modified) phase 0 argument.

getopt phase 3

In getopt phase 3 we execute each -o "output tool" from getopt phase 2, parsing the output as a command line.

The -o "output tool" may be given optstr from "-O tool=optstr", followed by the phase 0 argument.

The -o "output tool" is not allowed to output another "-o tool" or a "-O tool=optstr". Instead a -o "output tool" may execute another -o "output tool" and merge the output into its own.

Command line option order

Most command line options override earlier copies of the same option. However in the case of -H phase=name, a later -H phase=name only overrides an earlier use of -H for the same phase only. Also in the case of -s token=value, a later -s token=value only overrides an earlier use of -s for the same token only.

Output tools

An "output tool" may be used to add certain additional command line options to be processed.

IMPORTANT: An "output tool" will print each command line option / argument on a separate line. So for example, if an "output tool" wishes to convey:

-s TITLE='IOCCC entry locations' -H topnav=top-of-site

The "output tool" would output the following 4 lines:

-s
TITLE='IOCCC entry locations'
-H
topnav=top-of-site

The command line options printed by an "output tool" are processed after all command line options, and all options from a matching line from the md2html.cfg file, and before any filename arguments on the command line.

For example:

command [command_options ..] [cfg_options ..] [output_tool_options ..] [--] [filename_arg ..]

Here the command first processes any command line options, then it processes cfg_options obtained from this file from a first match of a given file_glob, then it processes output tool (i.e., -o tool) options found in command_options, and cfg_options, then an optional -- (end of all options), then zero of more filename_arg filename arguments.

The -o tool must NOT output any -o tool options as -o tool is NOT RECURSIVE. A -o tool is free, however, to execute other -o tool output tools and merge the output from those tools into its own.

Special -E exitcode option

When -E exitcode is evaluated, the application will exit with the exitcode value. If one wishes to also output a message to stderr, the -e string must come BEFORE any -E exitcode in the command line.

Substitution tokens

All tokens (i.e., strings of the form %%token%%) MUST be substituted (by some -s token=value) in all HTML output, except during phase numbers 20-29 (i.e., except before tool output, pandoc wrapper output, and after tool output), or the command will exit non-zero, unless -S is given. If -S is given, only a warning about non-substituted tokens will be written to stderr.

The command line of tools in the bin directory, and perhaps modified via the md2html config file may change to using a different filename for a given phase.

For example when forming the HTML from 2020/ferguson1/chocolate-cake.md, a different topnav navigation bar is needed. So instead of the usual top navigation bar that normally directs people to the previous entry for the year, or go up to the year page, or to the next entry for the year, a top navigation bar to just go up to the entry's main page is needed. A line in the md2html config file that refers to 2020/ferguson1/chocolate-cake.md may specify use of topnav.up2index.html (as topnav.up2index)instead of using the topnav.default.html (topnav.default.html) default.

The HTML phase may be skipped resulting in no HTML output during a given phase and furthermore, forming no HTML content from a given markdown file altogether.

See comments in the md2html config file for details. See also, the tool readme2index.sh for an example of how such command lines are used.

Use CAUTION when modifying inc files

Some of the files under this directory are used to form MOST of the HTML content on the official IOCCC web site.

These files are used to form MOST of the HTML content on the experimental web site.

... and in particular files under inc that are of the form (called default HTML files) ...:

*.default.html

... contain default content used to form IOCCC HTML / IOCCC web pages.

Instead of editing the default HTML files in order to fix a special web page, consider making a copy of the default file and modifying the md2html config file to refer to the copy instead. That way your special case situation will not impact MOST of the HTML content.

Why we do not use certain well-known HTML technologies

You may wonder why we need these files in the first place. You may wonder why we even need the tools in the bin directory when there other solutions available to form web pages.

Static web pages only

Here are some reasons why we are using these files and special tools to create HTML content / IOCCC web pages:

We host official IOCCC web site via GitHub pages. As of the time this file written, only static web pages are supported.

We cannot use server side include

We use static web pages, so use of "server side include" is not available to the IOCCC.

For example, Apache SSI "#include" does not work on GitHub pages.

We cannot use a back-end database

We use static web pages, so use of a "back-end database" is not available to the IOCCC.

We cannot use non-GitHub web servers

The GitHub pages have the distributed server capacity needed to handle the huge download volume or Slashdot effect that happens when a IOCCC entries are published.

Most so-called low cost web hosting sites have a somewhat hidden excessive bandwidth charge and/or cap bandwidth when the volume gets too high, and/or do not have the infrastructure that can handle the Slashdot effect.

The IOCCC had experience with people offering web mirrors and even free hosting only to find that the generous benefactor moves on, or their service is purchased by someone else who is not so generous. While it is possible that GitHub might someday suffer a similar fate, for the time being we are betting that GitHub will remain willing to host the IOCCC.

The official IOCCC web site is, after all, primarily C source code with some supporting documentation (sometimes :-) ). As such it is a natural fit for GitHub and GitHub pages.

We cannot use the HTML <object> element

The <object> HTML element does not work for our needs.

HTML elements do not extend into the content that they include. For example, menu bars ioccc.css will not operate under an HTML element.

We cannot use the HTML <embed> element

The <embed> HTML element does not work for our needs.

This element wants one to specify the width and height in pixels. Use of a percentage is not officially supported even if some browsers might do so. Our Responsive Web Design in the ioccc.css needs to be responsive to small-sized cell phone-like screens, mid-sized table-like screens, as well as large-sized desktop-like screens. Specifying a width and height in pixels will not work well in all of those screen size contexts.

We cannot use the HTML <iframe> element

The <iframe> HTML element does not work for our needs.

This element wants one to specify the width and height in pixels. Use of a percentage is not officially supported even if some browsers might do so. Our Responsive Web Design in the ioccc.css needs to be responsive to small-sized cell phone-like screens, mid-sized table-like screens, as well as large-sized desktop-like screens. Specifying a width and height in pixels will not work well in all of those screen size contexts.

We cannot use JavaScript to include content

We do not use JavaScript to include HTML content.

While the IOCCC may use JavaScript in the future to directly render things like C source code, we will do so in such a way that someone will be able to view official IOCCC web site content with JavaScript disabled.

The IOCCC will NOT MANDATE USE OF JavaScript to view official IOCCC web site.

For this reason, we cannot use JavaScript include HTML content.

IOCCC terms

The following IOCCC terms apply to tools, JSON files, and the directory structure of this repo.

author

An individual who was an author of at least one winning IOCCC entry.

Some authors have submitted more than one IOCCC entry that won. Some winning IOCCC entries have more than one author. In that case we might use the word authors.

author_handle

An author_handle is string that refers to a given author and is unique to the IOCCC. Each author has EXACTLY ONE author_handle.

For each author_handle, there will be a JSON file located at:

author/author_handle.json

where author_handle is the author's author_handle.

Because the author_handle is used to form a JSON filename, the string must be a POSIX safe string with the addition of + (for technical reasons beyond this document). In particular, the author_handle must be an ASCII string that matches this regexp:

^[0-9A-Za-z][0-9A-Za-z._+-]*$"

Default author_handle's do not have multiple consecutive _ (underscore) characters. Nevertheless, multiple consecutive _ (underscore) characters are allowed. Contest submitters who wish to override the author_handle may do so.

The author_handle is derived from the name of the author. While there is a algorithm that maps the name of the author (which can contain any UTF-8 characters) into a default author_handle string, those who submit an entry to the IOCCC are free to choose a different author_handle string if they so desire.

An author who has won a previous IOCCC is encouraged to reuse their author_handle so that new winning entries can be associated with the same author.

For an anonymous author, their handle is one of these forms:

Anonymous_year

or:

Anonymous_year.digits

The latter form is in case there are more than one anonymous author in a given year.

NOTE: even if the directory name is not anonymous the above rules apply as in in the case of 2005/anon.

Anonymous author_handle's match this regexp:

Anonymous_[0-9][0-9][0-9][0-9][.0-9]*$

entry

An IOCCC entry that won an award for a given IOCCC.

An entry has one more more authors.

Each entry is contained under its own directory. The parent of the entry directory is the year's directory.

While in most cases an entry consists of files under an entry's directory, there are a few cases where an entry directory contains subdirectories.

Use of subdirectories under an entry directory is discouraged and may be limited to previous entrys that used them.

NOTE: In the past, the term winner was used in instead of today's term of entry.

year

A year is a 4 character string. Most years are 4-digit strings that represent the year. Some special year strings are possible, such as mock. Non-numeric year strings are lower case.

A year string matches this regexp:

[0-9a-z][0-9a-z][0-9a-z][0-9a-z]

The year directories reside directly below the top level directory.

.year

Each year directory will have a file under it named:

.year

The contents of the .year file is the year string of the directory. For instance, 2020/.year has the string: 2020.

dir

A dir is a POSIX safe string that holds an entry.

A dir is a string that matches this regexp:

^[a-z][0-9a-z.-]*$

Each entry is under its own individual directory. This directory is directly under a year directory.

entry_id

A string that identifies the winning entry. The string is of the form:

year_dir