mirror/ioccc
1
0 Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge pull request #1315 from xexyl/1992-adrian-fixes

Browse Source
This commit is contained in:
Landon Curt Noll authored and GitHub committed 2023-11-08 12:45:37 -07:00
2 parents 322ee3ade8 f700aea9e1
commit 256c87cf11
5 files changed
+197 -79
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

1992/adrian/Makefile
+2 -2
View File
@@ -43,7 +43,7 @@ CSILENCE= -Wno-bitwise-op-parentheses -Wno-comment -Wno-deprecated-declarations
-Wno-logical-op-parentheses -Wno-macro-redefined -Wno-parentheses \
-Wno-return-type -Wno-uninitialized -Wno-unused-value -Wno-unused-variable \
-Wno-unused-but-set-variable -Wno-unsequenced -Wno-unused-parameter \
-Wno-empty-translation-unit
-Wno-empty-translation-unit -Wno-implicit-int
# Common C compiler warning flags
#
@@ -118,7 +118,7 @@ PROG= ${ENTRY}
#
OBJ= ${PROG}.o
DATA=
TARGET= ${PROG}
TARGET= ${PROG} adgrep
#
ALT_OBJ=
ALT_TARGET=
1992/adrian/README.md
+67 -49
View File
@@ -4,6 +4,17 @@
make all
```
### Bugs and (Mis)features:
The current status of this entry is:
```
STATUS: known bug - please help us fix
STATUS: INABIAF - please **DO NOT** fix
```
For more detailed information see [1992 adrian in bugs.md](/bugs.md#1992-adrian).
### Try:
@@ -17,7 +28,8 @@ For the slow minded, try:
./adsleep 32767
```
NOTE: if you do not specify an arg to the program it will segfault.
NOTE: if you do not specify an arg some of the programs such as `adsleep` it
will very likely segfault do something funny. This is supposed to happen.
## Judges' remarks:
@@ -32,9 +44,9 @@ NOTE: Some compilers have had trouble optimizing this entry.
### ADrian's GREP (adgrep)
For those confused by the complexity of full-blown egrep style regular
expressions, this program offers an alternative. It implements an
equivalent search, using a deterministic finite automaton.
For those confused by the complexity of full-blown `egrep(1)` style regular
expressions, this program offers an alternative. It implements an equivalent
search, using a deterministic finite automaton.
A deterministic finite automaton consists of a finite set of states,
along with transition rules to move from one state to another, an initial
@@ -48,25 +60,30 @@ accepting states, then the string is accepted.
The deterministic finite automaton is specified as a series of rules for
each state:
<state> chars1 <dest1> chars2 <dest2> ...
```
<state> chars1 <dest1> chars2 <dest2> ...
```
`chars1` is a list of characters (only the first 8 are significant) which
`chars1` is a list of characters (_only the first 8_ are significant) which
should trigger a transition to `<dest1>`. `<dest1>` is another state which
should have a similar specification somewhere. A state is accepting if
it is specified in square brackets: `[final]`, and state strings are
significant to only eight characters.
it is specified in square brackets: `[final]`; state strings are
significant to only eight characters like `chars1`.
### Example 1: matches `^abc$`
<q0> a <q1> The first state to appear is the start state
<q1> b <q2>
<q2> c [q3]
[q3]
```
<q0> a <q1> The first state to appear is the start state
<q1> b <q2>
<q2> c [q3]
[q3]
```
Technically, a deterministic finite automaton should have a rule for each
possible input character at each state. To simplify descriptions of the
automata, if no rule is present, the string will not be accepted. Also,
the `'.'` character matches any character if it occurs first in the
the `.` character matches any character if it occurs first in the
character list.
@@ -116,55 +133,55 @@ character list.
[q5]
```
With the automaton specification in 'filename', invoke the program by
With the automaton specification in `filename`, invoke the program by
typing
```sh
./adgrep 'filename'
./adrian filename
```
It will read stdin and print out all the lines which the automaton
It will read the file and print out all the lines which the automaton
accepts. If the file cannot be opened, a system error message will
be printed. If the input contains errors, then an error message along
with the number of the offending line will be printed to stderr. The\
number of rules for each state is limited to 17. If more than 17 rules\
are present, you get the error `too_many_rules`, and the state that was\
being processed is printed. Error `no_destination` occurs if you specify a\
with the number of the offending line will be printed to stderr. The
number of rules for each state is limited to 17. If more than 17 rules
are present, you get the error `too_many_rules`, and the state that was
being processed is printed. Error `no_destination` occurs if you specify a
set of characters, but no destination state, and error `too_many_states`
occurs if your automaton has more than 257 states.
Running:
```sh
./adgrep from < your_mailbox
./adrian from < your_mailbox
```
will perform a function similar to that of the unix from command.
will perform a function similar to that of the unix `from` command.
If no filename is specified on the command line, then `"adgrep.c"` is used
as the specification for the automaton. (This file has been renamed\
to `adrian.c` by the judges.) In this case, the program will search for\
If no filename is specified on the command line, then `__FILE__` is used
as the specification for the automaton. (Originally `"adgrep.c"` the file was renamed
to `adrian.c` by the judges but for more portability the arg was changed to
`__FILE__`.) In this case, the program will search for
matches to the regular expression:
```
^.[^|C][^w[Q]*(Q|[w[]c).*|^.[C|]$
```
I suggest using `adgrep.c` as input, and storing the output in `adwc.c`:
I suggest using `adrian.c` as input, and storing the output in `adwc.c`:
```sh
./adgrep < adgrep.c > adwc.c
./adrian < adrian.c > adwc.c
```
Compiling the new file, `mywc.c`, yields a clone of the unix wc command. It
runs on one file (defaulting to `"adgrep.c"` if no file is given) and
displays the number of lines, words, and bytes in the input file.
Compiling the new file, `adwc.c`, yields a clone of the unix `wc` command. It
runs on one file, defaulting to `"adrian.c"` if no file is given (again, this
was changed to be `__FILE__`) and displays the number of lines, words, and bytes
in the input file.
Another possibly interesting automaton can be created by slightly
adjusting the `adgrep.c` file. Change the first line to read
adjusting the `adrian.c` file. Change the first line to read
```c
/* . echo| . */
@@ -173,7 +190,7 @@ adjusting the `adgrep.c` file. Change the first line to read
and repeat the process above
```sh
./adgrep <adgrep.c > adecho.c
./adrian <adrian.c > adecho.c
```
The new file now contains all lines which match
@@ -182,7 +199,7 @@ The new file now contains all lines which match
^.[^5|m^]*[m^]([e=p,;]|[^e=+p,;].*)$
```
Compile and run. This is an echo clone. Note the efficient algorithm
Compile and run. This is an `echo(1)` clone. Note the efficient algorithm
employed.
@@ -199,10 +216,10 @@ you can search for matches to
^.[^W]*W..*$
```
By some freak happenstance, lines of adgrep.c which match this regular
expression form a unix head command. It prints the first ten lines of
the file specified on the command line (or adgrep.c if no file is
specified).
By some freak happenstance, lines of `adrian.c` which match this regular
expression form a unix `head(1)` command. It prints the first ten lines of
the file specified on the command line or adrian.c if no file is
specified (again this was changed to `__FILE__`).
By setting the first line to
@@ -210,14 +227,14 @@ By setting the first line to
/* . basename . */
```
a clone of the unix basename command can be unearthed. The automaton will
a clone of the unix `basename(1)` command can be unearthed. The automaton will
search for
```
^.[^j]*jr.*$
```
on standard input. And the program which results by running adgrep.c
on standard input. And the program which results by running `adrian.c`
through this filter requires two parameters. The first is meant to be a
filename, and the second, an extension. All leading pathname components
are removed from the filename, and the extension is removed if present.
@@ -235,11 +252,11 @@ you can search for
^.[^(~][^s]*sl.*$
```
Filtering adgrep.c through this search yields a clone of the sleep
command. Invoke with a single integer parameter, and it will pause
for that many seconds.
Filtering `adrian.c` through this search yields a clone of the `sleep(1)`
command. Invoke with a single integer parameter, and it will pause for that
many seconds.
If either adbasename or adsleep is invoked with too few parameters,
If either `adbasename` or `adsleep` is invoked with too few parameters,
the program will print the error message:
> Segmentation fault (core dumped)
@@ -248,15 +265,16 @@ the program will print the error message:
machine.) The four programs which read from stdin require lines
shorter than 999 characters.
The other info files are adrian.grep.[1-6] which contain the six
The other info files are `adrian.grep.[1-6]` which contain the six
examples that appear above, and from, which is used to emulate the
unix from command. For reasons of clarity, the name "from" should
probably not be changed if possible. I wouldn't want to be accused of\
unix `from` command. For reasons of clarity, the name `"from"` should
probably not be changed if possible. I wouldn't want to be accused of
confusing people by giving the input files weird names.
If you want to change the default input filename (line 80) you must be
careful to choose a name that doesn't match the wrong search patterns,
introducing extra lines into one of the programs.
introducing extra lines into one of the programs (again, this was changed to be
`__FILE__`).
The program will produce at least one warning and possible several
when compiled depending on the compiler.
1992/adrian/adrian.c
+2 -2
View File
@@ -23,7 +23,7 @@ int L=0,j= -28;
void E(int i, int m,char*c)
{
for(; i<43; i+=3)
putc("}|uutsrq`_^bji`[Zkediml[PO]a_M__]ISOYIRGTNR"[i]+i-9,stderr);
putc(i["}|uutsrq`_^bji`[Zkediml[PO]a_M__]ISOYIRGTNR"]+i-9,stderr);
fprintf(stderr,"(%d): %s\n" ,m,c);
exit(1);
}
@@ -77,7 +77,7 @@ main(int sl,char *j[]){
v(s = V[1]; if (*V=strrchr(s,'/'))s=*V+1; if(( !strncmp( s + (jr=strlen(s)) -
(q=strlen(V[2]) ),V[2],q))&&jr!=q) s[jr-q] = 0; puts(s); )
int e,p,C=0,Q ,basename;
W= fopen(wc>= 2 ? V[1] : "adgrep.c","rt");
W= fopen(wc>= 2 ? V[1] : __FILE__,"rt");if(!W)exit(1);
echo| m^ e| 5| (int) .8| echo|
wc |C ==o[o[C] .e] .
e| e==+p,p; s[o[C] .e ]
bugs.md
+53
View File
@@ -707,6 +707,59 @@ when you're cheating it ends up winning! Can you figure that out as well?
# 1992
## 1992 adrian
### STATUS: known bug - please help us fix
### Source code: [1992/adrian/westley.c](1992/westley/westley.c)
### Information: [1992/adrian/README.md](1992/westley/README.md)
The author states that:
```
If the input contains errors, then an error message along
with the number of the offending line will be printed to stderr. The
number of rules for each state is limited to 17. If more than 17 rules
are present, you get the error 'too_many_rules', and the state that was
being processed is printed. Error 'no_destination' occurs if you specify a
set of characters, but no destination state, and error 'too_many_states'
occurs if your automaton has more than 257 states.
```
but only `no_destination` appears to occur. It is possible that this is invalid
test cases but it's also possible it simply is a bug.
You are welcome to try and fix these problem(s).
### STATUS: INABIAF - please **DO NOT** fix
### Source code: [1992/adrian/westley.c](1992/westley/westley.c)
### Information: [1992/adrian/README.md](1992/westley/README.md)
The author stated that if the file cannot be opened then it will print a system
error but this is not the case unless it's showing a fault. However this was
instead fixed by Cody as part of another problem that was detected so that the
program just exits 1 if the file cannot be opened.
Also, if _some_ of the programs are not given an arg they will very likely crash
or do something funny. For instance `adsleep` and `adbasename`. This is very
simple to fix but the author explicitly noted that this will segfault and it is
that that is the error message.
Another thing is that the code:
```c
printf((W,Y));
```
appears to be wrong because `W` is a `FILE *` and `Y` is a `char[]` but if one
changes it to use `fprintf()` on the file (which, incidentally, is opened in
read only mode which is another reason) with the `%s` specifier it will make
`adhead` not work: it'll print nothing at all! So this code should not be
changed either even if it appears to be wrong. Notice too a curious thing: if
you did change it to fprintf, even if you have the right number of args, you'd
have to remove the outer `()` pair.
## 1992 gson
### STATUS: uses gets() - change to fgets() if possible
thanks-for-fixes.md
+73 -26
View File
@@ -12,22 +12,22 @@ contributed thousands, that we wish to thank.
We call out the extensive contributions of [Cody Boone
Ferguson](https://www.ioccc.org/winners.html#Cody_Boone_Ferguson) who is
responsible for most of the improvements and fixes including many **EXTREMELY
CHALLENGING bug fixes** like
HARD bug fixes** like
[1988/phillipps](/thanks-for-fixes.md#1988phillipps-readmemd),
[2001/anonymous](/thanks-for-fixes.md#2001anonymous-readmemd) and
[2004/burley](/thanks-for-fixes.md#2004burley-readmemd), making entries like
[1985/sicherman](/thanks-for-fixes.md#1985sicherman-readmemd) and
[1986/wall](/thanks-for-fixes.md#1986wall-readmemd) not need `-traditional-cpp`
(all **EXTREMELY CHALLENGING**), fixing entries to work with clang (some being
**EXTREMELY CHALLENGING** like
(all **EXTREMELY HARD**), fixing entries to work with clang (some being
**EXTREMELY HARD** like
[1991/dds](/thanks-for-fixes.md#1991dds-readmemd)) or as much as possible (like
[1989/westley](/thanks-for-fixes.md#1989westley-readmemd), a true masterpiece
that is **INCREDIBLY HARD, _MUCH, MUCH MORE SO_ than any other fix!**), porting
entries to macOS (some being **EXTREMELY CHALLENGING** like
entries to macOS (some being **EXTREMELY HARD** like
[1998/schweikh1](/thanks-for-fixes.md#1998schweikh1-readmemd)), fixing code like
[2001/herrmann2](/thanks-for-fixes.md#2001herrmann2-readmemd) to work in both
32-bit/64-bit which *can be* **EXTREMELY CHALLENGING**, providing alternate code
where useful/necessary, fixing where possible dead links or removing them,
32-bit/64-bit which *can be* **EXTREMELY HARD**, providing alternate code
where useful/necessary, fixing where possible/removing dead links,
typo/consistency fixes, improving **ALL _Makefiles_** and writing
[sgit](https://github.com/xexyl/sgit) that we installed locally to easily run
`sed` on files in the repo to help build the website. **THANK YOU VERY MUCH**
@@ -122,9 +122,15 @@ $ warning: this program uses gets(), which is unsafe.
whereas without the warning it's much easier to see that it's a prompt.
In some entries this change is not possible, in one-liners it might make them
too long and in some entries it's more complicated than others because of the
annoying fact that for '"compatibility" reasons' `fgets()` retains the newline
and `gets()` does not. Nevertheless some of the entries have been updated this
too long (though it's also been possible to do this in some cases) and in some
entries it's more complicated than others because of the annoying fact that for
'"compatibility" reasons' `fgets()` retains the newline and `gets()` does not.
Some entries like [1992/adrian](1992/adrian/README.md) are more complicated in
other ways due to the code generating other output; and because of how it works
it would generate code that could not be compiled, simply because of spaces
being added (a number of other fixes in that entry were also made)!
Nevertheless some of the entries have been updated this
way for the reasons described above and in the [FAQ](/faq.md).
@@ -143,7 +149,7 @@ the second line starts with `o, world!\n"`.
By request, the original code is provided as
[anonymous.alt.c](1984/anonymous/anonymous.alt.c) so that one can look at it and
the famous tattoo:
the famous tattoo which we also include here:
![1984-anonymous-tattoo.jpg](1984/anonymous/1984-anonymous-tattoo.jpg)
@@ -1299,27 +1305,68 @@ should have been removed.
## [1992/adrian](1992/adrian/adrian.c) ([README.md](1992/adrian/README.md]))
Cody changed the location that it used `gets()` to be `fgets()` instead to make
it safer and to prevent annoying warnings during compiling, linking or runtime
(interspersed with the program's output).
Cody fixed the code so that it will try opening the file the code was compiled
from (`__FILE__`), not `adgrep.c`, as
the latter does not exist: `adgrep` is simply a link to `adrian` as `adgrep` is
what the program was submitted as but the winner is `adrian`.
Not fixing this would cause the program to crash if no arg was specified as the
file did not exist. In doing this, at first the change to check for a NULL file
was added. Then it was noticed that the problem is that `adgrep.c` was an
incorrect reference that was never fixed in any of the files, not the code or
the documentation. A fun fact is that one can do:
```c
W= fopen(wc>= 2 ? V[1] : __FILE__,"rt");if(!W)exit(1);
```
but one _CANNOT_ do:
```c
W= fopen(wc>= 2 ? V[1] : __FILE__,"rt");if(!W)exit(1);
if (W==NULL)exit(1);
```
because `adwc.c` will be empty! The difference is it is on a newline, the check.
This is an example of how a simple change in code can break it and this is also
true of another change as further below.
Cody also restored a slightly more obscure line of code that had been changed:
```diff
- putc("}|uutsrq`_^bji`[Zkediml[PO]a_M__]ISOYIRGTNR"[i]+i-9,stderr);
+ putc(i["}|uutsrq`_^bji`[Zkediml[PO]a_M__]ISOYIRGTNR"]+i-9,stderr);
```
though it's questionable how much more (if at all) obscure that is.
Cody also changed the location that it used `gets()` to be `fgets()` instead to
make it safer and to prevent annoying warnings during compiling, linking or
runtime (interspersed with the program's output). This was complicated because
of how the other source files are generated, as noted above; simply changing the
code could cause invalid output in the program which made other files fail to
compile (for this example specifically, see below).
One might think that simply changing the `gets()` to `fgets()` (with `stdin`)
would work but it did not because `fgets()` stores the newline and `gets()` does
not. That is well known but this code was relying on not having this newline
(see also above).
With `fgets()` the code `if(A(Y)) puts(Y);` ended up printing an extra line
which made the generation of some files (like `adhead.c`) fail to compile. Why?
There was a blank line after a `\` at the end of the first line of a macro
definition! Thus the code now first trims off the last character of the buffer
read to get the same correct functionality but in a safe and non obnoxious way.
Later Cody improved the change to `fgets()` to make it slightly more like the
original. This still requires the additional stripping of the newline inside the
loop but now it uses what looks like before, just a call to `gets()`.
One might think that simply changing the `gets()` to `fgets()` (with `stdin`)
would work but it did not because `fgets()` stores the newline and `gets()` does
not. The code was relying on not having this newline. With `fgets()` the code
`if(A(Y)) puts(Y);` ended up printing an extra line which made the generation of
some files (like `adhead.c`) fail to compile. Why? There was a blank line after
a `\` at the end of the first line of a macro definition! Thus the code now
first trims off the last character of the buffer read to get the same correct
functionality but in a safe way and non obnoxious way.
But the improvement so that it uses `gets()` could not be changed to have the
macro do the removal of the extra line (as in with a comma operator or a `&&`)
as this caused compilation errors with another generated file (`adwc.c`). Thus
after the `gets()` call in the line that looks like:
as this, as might be expected from the above, caused compilation errors with
another generated file (`adwc.c`)! Thus after the `gets()` call in the line that
looks like:
```c
while( gets(Y) ){ Y[strlen(Y)-1]='\0'; if(A(Y)) puts(Y); }
@@ -1327,8 +1374,8 @@ while( gets(Y) ){ Y[strlen(Y)-1]='\0'; if(A(Y)) puts(Y); }
one must keep the `Y[strlen(Y)-1]='\0';` part and keep it there.
This is a complex change due to the way the program and Makefile generate
additional tools.
These fixes are complex changes due to the way the program and Makefile generate
the additional tools.
## [1992/albert](1992/albert/albert.c) ([README.md](1992/albert/README.md]))