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.
Example usage in vim with a keyboard macro <escape>f:
```vim
:map ^[f 1G!G../../bin/sgi^M
```
or via .vimrc:
```vimrc
" sort the entire .gitignore file using ../../bin/sgi
:map ^[f 1G!G../../bin/sgi^M
```
NOTE: The ^[ above is ((Escape)) and ^M is ((Control-M)).
Example usage in shell:
```sh
rm -f t && ../../bin/sgi < .gitignore > t && mv -f t .gitignore
```
Suggested usage:
```sh
bin/all-run.sh bin/sort.gitignore -v 1
```
Sorted all winner .gitignore according to this method.
Sorted all year level .gitignore according to this method.
Sorted top level .gitignore according to this method.
The man page is in /man/man1/. It is not clear to me if man/man1 should
be under another directory but that can be decided later and if
necessary it can be moved over. The man page has a few examples. More
could be contrived if necessary. It has no SEE ALSO section because it
is not clear to me what should be linked to (perhaps the IOCCC
website?). I felt like I should write this man page first, in
particular, because I wrote the script too and the benefit is I
discovered some bugs in the script as well that are now fixed.
The script only checked if the status.json file was readable but it did
not check if it is writable. It should check for both and it now does.
The exit code is the same as if the file does not exist, is not a
regular file or not readable: 1. It also now explicitly exits 0 at the
end of the script. The version was updated to
IOCCC_STATUS_SCRIPT_VERSION="0.0.5-0 2024-01-22" # major.minor.release-patch YYYY-MM-DD
When fixing the script some things crossed my mind that might be worth
considering. First of all if dry-mode is specified (-N) does it really
need to have a status.json file arg specified? It might or might not.
Second should there be an option to create a new status.json file (if
the one does not exist)? These are perhaps worth considering and
wouldn't be hard to implement if these are desirable.
It should be != 1 rather than -gt 1. Not having this meant that if no
arg was specified the usage message was not printed and instead the arg
of the status.json file (which was not given) was shown to not exist
(because it was empty so not a file).
Improved `all-run.sh` and `readme2index.sh` so that
certain command line options (e.g., -p, -P, -u, -U ...)
are only passed along to other tools if they are
given on the command line.
all-run.sh VERSION changed from "1.1.1 2024-01-19"
to "1.1.2 2024-01-20",
readme2index.sh VERSION changed from "1.1 2024-01-16"
to "1.1.1 2024-01-20".
Added `quick-readme2index.sh` that only calls
`readme2index.sh` if the `index.html` is not
a non-empty file, or of either `.winner.json`
and/or `README.md` is newer than `index.html`.
Updated all `index.html` files because the
the file contains a comment that includes
the newly modified readme2index.sh VERSION.
Updated manifest as per recent file changes.
Updated winner `.winner.json` to reflect the manifest changes.
Added [bin/all_run.sh](bin/all_run.sh) tool to run a command over
all winning entries.
These tools had slight improvements in the help usage message:
* bin/gen-author.sh VERSION="1.3.1 2024-01-16"
* bin/gen-inventory.sh VERSION="1.0.1 2024-01-16"
* bin/pandoc_wrapper.sh VERSION="1.3.1 2024-01-16"
* bin/subst.default.sh VERSION="1.1.1 2024-01-16"
* bin/subst.winner-index.sh VERSION="1.1.1 2024-01-16"
The `bin/readme2index.sh` tool was improved so that
if the `index.html` file exists and the tool forms
the exact same content, then the file `index.html`
file is NOT touched.
* bin/readme2index.sh VERSION="1.1 2024-01-16"
We used the following command to generate `index.html` files
for all winning entries:
```sh
bin/all_run.sh -v 3 bin/readme2index.sh
```
NOTE: We have not (yet) built HTML files above the winner directory
level (e.g., no year level nor top level HTML), nor have be built
non-index HTML files for winners.
NOTE: By default we are running `pandoc(1)` with:
```
--fail-if-warnings=true
```
This means that any problems with markdown to HTML
generates an error. This is done to be sure that
the HTML conversion is clean.
We had to fix the following `README.md` files:
* `2000/rince/README.md`
The reference [^2] was never used. We made a guess that [^2] applies
to the same spot as [^1]. Perhaps there is a better spot where
[^2] applies?
NOTE: For every reference, there MUST be some place where that
reference is used.
* 2012/vik/README.md
The string "$...$" is assumed to be a TeX encoded string.
NOTE: Every "$..$" needs to be backslashed as in "\$..\$".
* 2018/burton1/README.md
The reference [2] was duplicated. We changed the
second [2] to [3].
NOTE: Every reference needs to have a unique number.
* 2020/carlini/README.md
The string "$...$" is assumed to be a TeX encoded string.
NOTE: Every "$..$" needs to be backslashed as in "\$..\$".
We also found 2 errors in the `author/author_handle.json` file set:
* author/Anonymous_2001.json
The `author/Anonymous_2001.json` file was missing.
* author/Jari_Arkko.json
The winner ID for `1989_jar.1` had a typo.
The initial release of `bin/readme2index.sh` tool that converts
`READMD.md` and various JSON files such as `.winner.json`
as well as "author/author_handle.json" files.
Applied the `readme2index.sh` tool to generate `2020/ferguson1/index.html`.
We generated the `2020/ferguson1/index.html` by executing:
```sh
./bin/readme2index.sh -v 3 2020/ferguson1
```
NOTE: This above tool requires **pandoc to be installed**.
See the [pandoc web site](https://pandoc.org) for how
to install. We used _HomeBrew_ to install it,
but there is also a _macports_ process as well.
NOTE: While the above web page does have formatting issues, it is
a first attempt data driven programmatic web page generation.
Some adjustments can be made by mostly improving `ioccc.css`,
and some perhaps by modifying how `bin/pandoc_wrapper.sh`
invokes pandoc. But probably most of the improvements can
be made by fixing and improving `ioccc.css`.
TODO: Fix and improve the `2020/ferguson1/index.html`
web page by fixing and improving `ioccc.css`.
Moved scripts and tools needed to generate HTML pages
into a new `/bin` subdirectory.
TODO: Consider getting rid of the [build-ioccc
repo](https://github.com/ioccc-src/build-ioccc) and just use the
`/bin` subdirectory.
NOTE: Some links in the `2020/ferguson1/index.html`
web page do not (yet) exist.
TODO: Once the basic format of the `2020/ferguson1/index.html` web
page is somewhat reasonable, generate `index.html` pages for other
winners. Generate the year level `index.html' files.
Generate the top level HTML files, including the missing
`location.html` file.
Greatly simplified the `.winner.json` file format,
removing a number of fields that were not strictly needed
for generating an inventory on a `index.html` page.
Updated various versions:
* `bin/gen-author.sh` VERSION="1.3 2024-01-14"
* `bin/gen-inventory.sh` VERSION="1.0 2024-01-14"
* `bin/manifest.winner.json.awk` VERSION="1.0 2024-01-14"
* `bin/pandoc_wrapper.sh` VERSION="1.3 2024-01-14"
* `bin/readme2index.sh` VERSION="1.0 2024-01-14"
* `bin/subst.default.sh` VERSION="1.1 2024-01-14"
* `bin/subst.winner-index.sh` VERSION="1.1 2024-01-14"
* `tmp/gen_path_list.found.sh` VERSION="1.1 2024-01-14"
* `tmp/gen_winner_json.sh` VERSION="1.1 2024-01-14"
The `.winner.json` file format from winner_JSON_format_version
"0.5 2023-11-04" to "1.0 2024-01-14".
All `.winner.json` files have been updated to use the
new `.winner.json` file format.
The manifest (`tmp/manifest.numbers` and `tmp/manifest.csv`) has
updated to use the new `.winner.json` file format.
The following procedure was used to generate the
`tmp/manifest.csv` manifest file:
0. open [manifest.numbers](manifest.numbers) in numbers: modify if/as needed
NOTE: Please try to keep the [manifest.numbers](manifest.numbers) spreadsheet
in order order. Use Organize -> Sort -> ((Sort Now)) to sort.
NOTE: Empty fields in the [manifest.numbers](manifest.numbers) spreadsheet
are not allowed. By default, a cell that is empty is colored
as a warning.
1. File -> Export To -> CSV...
* 1a. `[ ]` Include table names (unset)
* 1b. Text Encoding: Unicode (UTF-8)
* 1c. Click ((Save..))
2. Save As: manifest.csv
* 2a. Select the `tmp` directory
3. Click ((Export))
* 3a. If needed click ((Replace))
4. Execute ./tmp/[fix_csv.sh](fix_csv.sh)
NOTE: You must run the [fix_csv.sh](fix_csv.sh) in the above
step in order to make the CSV file usable by other tools.
NOTE: If no changes are needed in the above step, the
original file content will be restored.
5. Execute ./tmp/[gen_path_list.found.sh](gen_path_list.found.sh)
NOTE: The above step silently performs a `make clobber` at the
top level, and so it takes a bit of time.
NOTE: If no changes are needed in the above step, the
original file content will be restored.
6. Execute ./tmp/[check_path_list.sh](check_path_list.sh)
NOTE: The above step prints nothing and for an exit 0, then
all is well. If there is a problem, such as files
found that are not part of the manifest, or files
in the manifest that are not found, then an error
message will be printed and the tool will exit non-zero.
If errors are reported, go back to step 0 and
correct the errors.
7. Execute ./tmp/[run_all.sh](run_all.sh) ./gen_winner_json.sh
NOTE: The above step will take a few seconds to run. It will rebuild
all `.winner.json` files for all winners.
NOTE: If no changes are needed in the above step, the
original file content will be restored.
8. Execute `git status` and inspect changes
NOTE: Check in changes if and as needed.
NOTE: In a number of shell scripts under `bin/*.sh`,
and in the `bin/manifest.winner.json.awk` script
you will find comments of the form:
```
```
This is a note that when tools such as `jnamval` (and `jval` and `jfmt`) are
ready, these gross hacks should be replaced with calls to those tools.
Until they are ready, some must be careful with the layout of JSON files.
A number of bugs and typos have been fixed.
NOTE: Fixing typos in the files touched by this comment is welcome.
This includes scripts, data files, text files, etc.
The _inventory_order_, _display_as_, _display_via_github_. and
_winners_text_ fields of the [manifest.numbers](manifest.numbers)
spreadsheet were just set to have some initial values.
As they impact the inventory displayed on `index.html` pages,
their content needs to be improved.
NOTE: Improving the above mentioned fields of the
[manifest.numbers](manifest.numbers) spreadsheet is welcome.
The src tree now contains source for tools that will be published
as part of the IOCCC winner web site. However, the compiled and/or
installed tools that install under bin, will NOT be published.
The bin directory is now excluded under .gitignore so that
those who use this repo will have to do:
```sh
cd src; make clobber all install
```
The src/Makefile in turn drives rules in sub-directory
Makefiles. Currently there are only two sub-directories:
winner_name - a completely re-written extract_winner_handles.sh
sgit - run sed on all files under git control
The check_awards.sh tool was moved under tmp
as this tool depends on CSV file, will not be needed soon,
and will go away when the tmp directory is cleaned
out towards the end of finishing issue #4.
This is the first step in making the new CSV file that identifies which
entries each author won. To get a list of all winning handles run
bin/extract_winner_handles.sh
...which is a very simple pipeline (that could probably be simplified
further) acting on winners.html.
This was done quite easily with a bit of thought. I generated a script
that could be modified in vim if necessary (as it turned out to be
though some of it could have been generated too but I did not think of
everything at once and it is a hack).
This is how I generated it:
echo '#!/bin/bash' > bin/check_awards.sh ; ( while read -r g ; do echo grep "$g"; done < <(sed -e 's,_,/,g' -e 's/"/\\"/g' tmp/year-auth-prize.csv | awk -F, '{print grep "\x22" $2 "\x22" " " $1 "/" "README.md"}' ); ) >> bin/check_awards.sh
Then I I modified the file (again this was a hack to make it easier and
quicker and better for being tired too) in vim to make the grep use -Hc
so I could pipe it to 'grep :0'. This found the files without the right
award. So the final command was:
sh bin/check_awards.sh |grep :0
though I did have to edit some files several times.
For reference the script has been added to the repo too but with +x
permissions.
This change also required updating the CSV file in tmp/.
Now supports backup extension.
Some warnings are noted in the usage (about backup extensions - one of
which goes for sed in general) and the README.md file.
Typo fixed the README.md file (some examples were in error for instance)
as well and add additional examples.
The new version corresponds to
https://github.com/xexyl/sgit/commit/9fccaa8cd680d3ed7cd91516710a5672630b5dff;
the README.md file had several new commits.
Add warning about use of -o n without -I depending on the sed command.
Updated version.
The README.md I'm not sure about but it feels like it might belong there
anyway. It can be removed if it does not seem useful or it is out of
context (and it might very well be). Nevertheless I think it's important
that the warning be there so it's good to update sgit here as well.
It did not iterate over all globs because the condition of the loop was
while [[ "$i" -le "$#" ]]; do
but in the loop i was incremented but the num args left was decremented
thus they went in opposite orders. Instead record how many globs are
there first and check i against that.
A problem existed where more than one command would not work (always or
in some conditions depending on the command: not sure which). This
prevented a fix in years.html which will be in the next commit. The sgit
repo has been updated with this fix as well.
The fix is to remove space from IFS and to change "${SED_COMMANDS[*]}"
to be "${SED_COMMANDS[@]}".
Now allows to not overwrite files.
Now allows sed options to be passed to sed.
Allows specifying path to sed.
This script now has its own repo at https://github.com/xexyl/sgit. I
don't expect many people to use it but it seemed like it was worthy of a
repo as it is a pretty handy script (though far from perfect).
New version of sgit.
Script now allows for spaces in sed commands. Thus one can do something
like:
sgit -v3 -x -e 's/web site/website/g' .
to set verbosity to level 3, turn on tracing (see below) and change
occurrences of 'web site' to be 'website'.
Also added -x option to turn on tracing.
Accidentally committed absolute path of sed.
Also it now shows the sed commands with high enough -v level. This is to
help with a strange problem I discovered (not fixed yet).
Script no longer requires GNU sed but works with other sed
implementations as well (only other one tested is macOS however).
Uses getopts to parse options. Usage is now:
usage: sgit [-h] [-V] [-v level] [-e command] <glob...>
-h print help and exit
-V print version and exit
-v level set verbosity level
-e command append sed command to execute on globs
which means that yes one can specify more than one glob. As long as one
sed command is specified (chose -e to match sed) and one arg is
specified after the last option then the command line is okay (whether
the sed command or file is okay is another matter entirely).
Verbose output to help test / debug features or if just interested in
what's being done.
Added `bin/sgit`.
This directory is NOT intended to become a huge software project
right now. For now it is just a convenient place to store tools.
Please avoid overthinking, over documenting tools under bin.
Later on there will be a tool that builds links in the web site
from JSON files, perhaps forms HTML files from markdown files, etc.
However we will have to make major progress on issue #1, issue #2,
issue #3 and issue #4 before such code can be written.
In some cases tools put under bin might go away later on
if they are no longer needed.
Cody Boone Ferguson