Due to the fact we now have .orig.c files and thus the possibility of
winner.orig or prog.orig being built I have updated all the .gitignore
files so that these files ignored. This was done with the following
shell commands, for reference:
for i in $(find 19* 200* 201[0123] -mindepth 2 -maxdepth 2 -name .gitignore) ; do echo $(dirname "$i"|sed -e 's,[0-9]\{4\}/,,g').orig >> "$i"; done
for i in $(find 201[4-9] 2020 -mindepth 2 -maxdepth 2 -name .gitignore) ; do echo prog.orig >> "$i"; done
and the two commands is because it was only in 2014 that the winning
entries were renamed to prog so prog.orig is the correct glob.
Used the following script to check for warnings
not silenced via ${CSILENCE}:
```sh
find ???? -mindepth 1 -maxdepth 1 -type d | while read dir; do
(cd "$dir"; echo =-=-= "$dir" =-=-= ; csilence -v 1 clobber all ; echo ; csilence -v 1 clobber alt)
done 2>&1 | less
```
Per discussion months ago with Landon the idea of rot13 will in general
be discouraged and the website will likely have a link to undo rot13
text. Also this was necessary to figure out how to format the
file.
The fix thank you's have been moved from `README.md` to
`thanks-for-fixes.md`.
**WARNING**: Some mistakes may have been made in doing this.
We presume later changes by others will correct any issues.
The `CONTRIBUTING.md` was renamed `how-to-enter.md`.
A 3rd line of links was added to top level html such as
`contact.html`, `index.html`, `judges.html`, `winners.html`,
and `years.html`.
Improved blank line use in a number (but not all) `README.md` files.
This to make the country code consistent with mkiocccentry (and I guess
international standard country codes).
This affects many README.md files and (for one line) winners.csv.
The typo of most importance (maybe the only typo) was that when I linked
to the archive tarball I left off /archive in the link.
Noted in the judges' comments that in the smaller version provided by
the author the exit() returns void problem is fixed but it still uses
gets(), not fgets(). I don't see the reason to do that as it'll compile
and it's really only an aside and not the entry.
About why it's useful. It's a problem to keep gets() but it might be a
problem to not keep it as well. Nevertheless with macOS the output of
the alternative code (which probably should be deleted but this can
happen another time) is confusing:
$ ./tbr.alt
$ warning: this program uses gets(), which is unsafe.
[]
The [] is the cursor on a blank line. The first line of output of the
program (which is not supposed to have a newline) instead has the
warning and then a newline appended to it.
As both Landon and I have been wondering if changing entries to use
fgets() is bordering on tampering with entries I decided to make the
original (sometimes bug fixed) versions (using gets()) alt code. I had
originally started to make the fgets() version of each entry the alt but
since the fgets() versions can handle input better and since some also
fixed bugs (see below) I decided to make the original version the alt
code (this also is how it is for other entries).
There were some exceptions here. If the change also fixed some crashes
or other bugs outside of the gets() to fgets() conversion I did not
bother to make an alt version. An example is 1992/dds. In some cases the
alt code does have bug fixes in which case they were added.
Updated the Makefiles and .gitignore files as well for the alt target
and the new files (compiled and in the case of 1992/adrian additional
source files in the Makefile alt rule).
This should solve the problem observed by Landon and me. It wouldn't be
a problem to change it so the alt version is the one that uses fgets()
but I think that might be the wrong choice.
In some entries I altogether avoided changing to fgets() say if the
author notes it uses gets() (in one case even providing a version that
uses fgets()). In some cases using fgets() introduces a bug so those are
also not changed.
As for 1990/tbr there's a strong reason for not having the fgets()
version being the alt. In macOS in particular (in linux - or maybe gcc,
not sure - it shows a warning at linker time) it will mess up the first
line of output in a potentially confusing way. For instance:
$ ./tbr.alt
$ warning: this program uses gets(), which is unsafe.
# nothing here, what to do?
And that touches on another issue of making the alt version the one that
uses gets(): the warnings that can be alarming during compilation,
linking and (in the case of macOS) runtime.
Not all entries should be or can be changed to fgets() but some can and
it's warranted for these reasons. More will possibly happen in the
future.
For Makefiles that have alt related rules, moved
the related rules under the alt rule section.
In some cases where there is no alt code,
removed alt rules and/or targets.
Simply the Makefile header.
Refer to the //www.ioccc.org/bugs.md URL (that will eventually exist).
Now using CC BY-SA 4.0 License to how you may share and adapt the
Makefile.
The segfault has been fixed when EOF is sent to the program (causing
fgets() to return NULL).
Since fgets() is far safer than gets() and since the gets() version will
intersperse the warning and the first line of output I have changed the
alt version to be the original (fixed to compile with modern compilers)
and the tbr is the one with fgets().
This is so that the warning about gets() is not interspersed with the
initial output of the program (see below). I do want to note that if one
sends EOF it will segfault so to exit one should do ctrl-c (for
instance) (exit never worked). I will look at fixing that fault later on
to match the original code but this is good for now. When fixed (low
priority though) I will consider making the gets() version the alt but
probably not (I guess time will tell).
As for the interspersed output it looks like:
$ warning: this program uses gets(), which is unsafe.
# cursor is here
At the cursor I can either hit enter to get another prompt or I can
input a command but this is not immediately obvious.
Updated .gitignore to have tbr.alt.
In some files changed the formatting was fixed to be a proper markdown
file.
There were some typos including that some README.md files that linked to
/winners.html#Cody_Boone_Ferguson actually linked to the non-existent
file /years which obviously would not work.
I also added links to the rest of the places I have fixed entries (but
see below) and I did the same for Yusuke Endoh's fixes.
As well apparently one entry I thought he fixed to compile under modern
systems does not seem to so that entry will have to be revisited another
time. That entry was 1993/jonth.
Those are the main changes here if not all of them.
Add thank you note to Cody in some README.md files, remove the warning
that 1990/westley/westley.c might not compile with modern compilers
(since I fixed it) and reworded a statement slightly. Yes I'm thanking
myself but it's actually from the judges.
Although I previously got them to all compile they did not all work. Now
they do except that jaw does not work properly in macOS (I have noticed
something odd in linux too but it appears to work mostly if not
entirely).
Yusuke Endoh pointed out for 1990/theorem that atof() nowadays needs
include stdlib.h so this was added to the file. As well I added the
inclusion of string.h for strcmp(). I also fixed some segfaults in this
entry (even though the judges pointed out that for NULL pointer one
needs the four '0 0 0 0' it still segfaulted under those conditions) so
that everything should work now under both linux and macOS. There were
iirc two parts to this: one the NULL pointer is no longer dereferenced
but also the type being passed to strcmp() was a int * and not a char *.
This prevented 1990/theorem/sorter from working at all (segfault) and as
I recall it there was another segfault I encountered as well.
For 1990/jaw Yusuke Endoh provided a ruby implementation of btoa as it's
not always available in modern systems. Thank you Yusuke!
In 1990/jaw shark.sh had to be modified as well. This had to do with '.'
not being in the default PATH.
As well I had previously made a mistake about btoa versus atob as both
were referenced in 1990/jaw/README.md. I believe I have this correct
now.
Many if not all the README files were modified with various changes
including formatting (like turning them into markdown) and section
titles amongst other things.
I seem to recall I'm missing something here but I can't recall what I
recall :-) so this will have to suffice this time round (maybe it was
the btoa/atob issue that I just remembered and added above but I'm not
sure).
My first 41st, 51st and 29th (all depending on the base) birthday
updates (possibly more will follow but this is not a certainty). All the
1990 entries now compile!
In one we used a patch from Yusuke Endoh (thank you Yusuke!) which we
credited him in the README and in another we used a tip from him (use
-ansi at the compiler line) to get the other to compile (again, thank
you Yusuke!). We fixed the others. In the case of 1990/tbr/tbr.c the
author had supplied a shorter version and we fixed that one too.
Added 1990/.gitignore which has the built files of 1990 so that after
doing `cd 1990 ; make all' running `git status .' will not show any
untracked files.
Added Makefiles for 1990 entries.
In a some cases the source code was modified.
Modified 1990 README.md files.
In some cases the program still cannot be compiled.
There is one exception:
$ find . -name '*.hint' -execdir git mv '{}' README.md \;
git statufatal: destination exists, source=1992/albert/albert.hint, destination=1992/albert/README.md
Looking at the two files they differ quite a bit so I've not changed it.
I'll bring this up on GitHub.
We we move 1988-2004 entries from all of a year's winning entries
in a single directory to each winner being in their own sub-directory
under the given year directory.
Cody Boone Ferguson