From c0e237528bef56fbff5bede5a1744c11fe267874 Mon Sep 17 00:00:00 2001 From: Cody Boone Ferguson <53008573+xexyl@users.noreply.github.com> Date: Wed, 19 Apr 2023 10:40:29 -0700 Subject: [PATCH] Fixes and updates in 1998/dorssel Format fixes in README.md. Rename dorssel.hint -> dorssel.md and update formatting. Updated Makefile DATA variable for rename of dorssel.hint. --- 1998/dorssel/Makefile | 2 +- 1998/dorssel/README.md | 71 +++++++++-------- 1998/dorssel/{dorssel.hint => dorssel.md} | 94 +++++++++++++++-------- 3 files changed, 102 insertions(+), 65 deletions(-) rename 1998/dorssel/{dorssel.hint => dorssel.md} (70%) diff --git a/1998/dorssel/Makefile b/1998/dorssel/Makefile index 8d941e59a..3bfaeef52 100644 --- a/1998/dorssel/Makefile +++ b/1998/dorssel/Makefile @@ -111,7 +111,7 @@ ENTRY= dorssel PROG= ${ENTRY} # OBJ= ${PROG}.o -DATA= dorssel.hint +DATA= dorssel.md TARGET= ${PROG} # ALT_OBJ= diff --git a/1998/dorssel/README.md b/1998/dorssel/README.md index df4113100..10180692d 100644 --- a/1998/dorssel/README.md +++ b/1998/dorssel/README.md @@ -1,29 +1,35 @@ # Obsolescent Feature - Frans van Dorsselaer - Bakker Industrial Automation - Kaiserstraat 35 - NL-2311 GP Leiden - The Netherlands +Frans van Dorsselaer +Bakker Industrial Automation +Kaiserstraat 35 +NL-2311 GP Leiden +The Netherlands ## To build: - make all +```sh +make all +``` ## To run: - ./dorssel - # input some text +```sh +./dorssel +# input some text - echo text | ./dorssel +echo text | ./dorssel +``` ## Try: - echo Simple as 123 | ./dorssel - echo 'Binary as simple as 1, 10 11.' | ./dorssel - echo '.... --- .-- -.. --- . ... .. - -.. --- - .... .- -' | ./dorssel - echo 'HOW DOES IT DO THAT' | ./dorssel - echo 'HOW DOES IT DO THAT' | ./dorssel | ./dorssel +```sh +echo Simple as 123 | ./dorssel +echo 'Binary as simple as 1, 10 11.' | ./dorssel +echo '.... --- .-- -.. --- . ... .. - -.. --- - .... .- -' | ./dorssel +echo 'HOW DOES IT DO THAT' | ./dorssel +echo 'HOW DOES IT DO THAT' | ./dorssel | ./dorssel +``` ## Judges' remarks: @@ -52,15 +58,14 @@ You might also want to read the [dorssel.hint](dorssel.hint) file. ## Author's remarks: -Usage -===== +### Usage This program converts ASCII to Morse code, and vice versa. The program reads in lines from standard input and writes its result to standard output. The direction of conversion is determined on a per line basis. Each line must consist of a maximum of 952 characters -followed by a new line character, they must not contain any NUL +followed by a new line character and they must not contain any NUL characters. Otherwise, some characters may be lost in the conversion. - Encoding to Morse code: The program knows the Morse codes for all Latin @@ -69,40 +74,46 @@ are replaced by an ASCII space ' '. Morse code per converted character is written as a sequence of '.' and '-' characters, each sequence followed by a space. -- Decoding from Morse code: Morse code should be typed as sequences of '.' and -'-' characters, separated by a space. Extra spaces will be echoed back (to -separate words). Unknown sequences will be converted to a '?' character. +- Decoding from Morse code: Morse code should be typed as sequences of `.` and +`-` characters, separated by a space. Extra spaces will be echoed back (to +separate words). Unknown sequences will be converted to a `?` character. -Notes -===== +### Notes The program assumes both source and execution character sets to be ASCII. As far as I can tell, the program is conforming and does not invoke undefined behavior. -Hints -===== +### Hints -Try +Try: - ./dorssel < dorssel.c | ./dorssel +```sh +./dorssel < dorssel.c | ./dorssel +``` to see what variables, numbers, and functions are used. (Does a better job than most C beautifiers). - Filter your text twice (as in the previous sample) to obtain a version of your -text that can be reversibly morsified, i.e. +text that can be reversibly Morsified, i.e. - ./dorssel | ./dorssel | ./dorssel | ./dorssel +```sh +./dorssel < dorssel.c | ./dorssel | ./dorssel | ./dorssel +``` is the same as - ./dorssel | ./dorssel +```sh +./dorssel < dorssel.c | ./dorssel +``` - The file "dorssel.hint" contains all the information needed to understand this program. For an obfuscated spoiler, try - ./dorssel < dorssel.hint +```sh +./dorssel < dorssel.hint +``` ## Copyright and CC BY-SA 4.0 License: diff --git a/1998/dorssel/dorssel.hint b/1998/dorssel/dorssel.md similarity index 70% rename from 1998/dorssel/dorssel.hint rename to 1998/dorssel/dorssel.md index 0a92fa954..957bc1cfc 100644 --- a/1998/dorssel/dorssel.hint +++ b/1998/dorssel/dorssel.md @@ -1,14 +1,18 @@ +# Information about this entry + Morse codes for letters and digits use sequences of 1 to 5 (inclusive) Morse symbols (denoted as '.' or '-'), e.g. +``` E . Z --.. 3 ...-- +``` -(BTW: digits always have 5 symbols, letters 1 to 4. I didn't use this fact) +(BTW: digits always have 5 symbols, letters 1 to 4. I didn't use this fact.) -Of course we can encode these symbols with bits in a byte. I chose '-' -to be a bit set to 0, '.' to be a bit set to 1. Characters in C are at +Of course we can encode these symbols with bits in a byte. I chose `-` +to be a bit set to 0, `.` to be a bit set to 1. Characters in C are at least 8 bits, but they are either signed or unsigned. To be portable, I've used only values from 0 to 127 (inclusive) (both in the encoding of Morse codes, as well as in subexpressions in the program). Hence we @@ -18,42 +22,52 @@ the source code excessively large ;-) The least significant bit contains the first symbol of a Morse code, so we have +``` E > 0......1 Z > 0...1100 3 > 0..00111 +``` The dots have still to be determined. But we also need to encode the length of the sequence. Since there are only 4 different kind of lengths, two bits would suffice, e.g. +``` E > 000....1 Z > 010.1100 3 > 01100111 +``` That means bits 6 and 5 set to 00 is length 1, set to 01 is length 2, etc. But this would mean that we would have to extract the length into a counter before shifting out each symbol, such as +```c length = (encoding & 96) >> 5; do { putchar(encoding & 1 ? '-' : '.'); encoding >>= 1; } while (length--); +``` We can also use the remaining bits to *include* the counter +``` length 1: 00000.1x E > 00000.11 length 2: 0000.1xx length 3: 000.1xxx length 4: 00.1xxxx Z > 00.11100 length 5: 0.1xxxxx 3 > 0.100111 +``` With 1 bit still to be chosen freely (let's assume it is 0). Now we can do +```c do { putchar(encoding & 1 ? '-' : '.'); encoding >>= 1; while (encoding > 1); +``` As you see, the encoded Morse sequence can be the counter itself. But this is not very pleasant, as some Morse sequences will end up having @@ -61,66 +75,75 @@ an ASCII value < 32, i.e. they will be control codes. This would make the initializing (encoding) string awkward. Hence I came up with the following +``` length 1: 0010010x E > 00100101 (= '%') length 2: 001010xx length 3: 00110xxx length 4: 0100xxxx Z > 01001100 (= 'L') length 5: 011xxxxx 3 > 01100111 (= 'g') +``` Notice these encodings in the initialization string. Only the coding for -'5' needs an escape (5 = ..... > 01111111 = '\177'). The way the length +`5` needs an escape (`5 = ..... > 01111111 = '\177'`). The way the length is encoded still allows the encoding to be its own counter: +```c do { putchar(encoding & 1 ? '-' : '.'); encoding = (encoding + 32) >> 1; } while (encoding > 35); +``` -The last shift will always result in 00100010, hence the test for >35. -Note that the calculation for the next value of `encoding' is done at int -precision, so that the actual value of the char remains <= 127. +The last shift will always result in `00100010`, hence the test for `>35`. +Note that the calculation for the next value of `encoding` is done at int +precision, so that the actual value of the char remains `<= 127`. -The character '-' (ASCII 45) and '.' (ASCII 46) are conveniently in +The character `'-'` (ASCII 45) and `'.'` (ASCII 46) are conveniently in sequence, so the putchar can simply be +```c putchar(45 + encoding % 2); +``` -I used the modulo operator this time, because '+' takes precedence over '&'. +I used the modulo operator this time, because `+` takes precedence over `&`. The assignment can be incorporated into the while control test, and a do-while loop with a body of a single expression statement can be written as a while loop with a comma operator in the test: +```c while(putchar(45 + encoding % 2), encoding = encoding + 32 >> 1); +``` Voila. -Now, the decoding of Morse codes is just the reverse. We shift in 0's -and 1's as we scan the line. But: we have to scan from back to front +Now, the decoding of Morse codes is just the reverse. We shift in `0`s +and `1`s as we scan the line. But: we have to scan from back to front (the left most Morse symbol is in the least significant bit, and thus has to shifted in last). The test for the scanning while loop must -stop when there are no more '.' or '-' on the left. Hence we need a +stop when there are no more `.` or `-` on the left. Hence we need a terminating character in front of the line. The input line is read one past the encoding string (into the same character array) the terminating -character for the reverse scan is therefore provided by the implicit '\0' +character for the reverse scan is therefore provided by the implicit `'\0'` of the array initializer ! The test also stops when at most 5 symbols have been shifted in. The encoded character is then looked up in the encoding string with -memchr (why memchr is explained later). Either a '?' (when there are +`memchr()` (why `memchr()` is explained later). Either a `?` (when there are more than 5 consecutive Morse symbols, or if the sequence is unknown) -or the decoded character is printed. Or a ' ' if there was an extra +or the decoded character is printed. Or a `' '` if there was an extra space in the source line. Note the use of "or" in these last sentences, as they explain the use of the conditional operator. -The strspn calls need either " .-" or ".-" as argument. Of course they +The `strspn` calls need either `" .-"` or `".-"` as argument. Of course they are overlapped. I put these strings between the encodings for digits -and letters. This is OK, since neither of these characters, or the '\0', -is a valid encoding ! We must, however, use memchr instead of strchr in -the decoding lookup because of this '\0'. Since it is far more obscure +and letters. This is OK, since neither of these characters, or the `'\0'`, +is a valid encoding ! We must, however, use `memchr()` instead of `strchr()` in +the decoding lookup because of this `'\0'`. Since it is far more obscure this way, I hope you'll forgive me. The outline of the program is roughly as follows +``` while (more lines) { if (line is in morse) { while (more characters) { @@ -137,23 +160,25 @@ while (more lines) { putchar(' '); } } +``` -Because of the *two* while(more characters) loops are the same, I've -exchanged the if(line is in morse) and the two while loops (so there -is *single* while loop). However, we must now store the if(line is in -morse) test for efficiency (it would be stupid to test the entire line -for every character). This is done in l[0] ! (the first encoding is -thus in l[1]). The memchr decoding lookup is not disturbed, because the -test is always 0 for Morse lines, and '\0' is not a possible encoding. -l[0] is also used as the temporary shift character while encoding a -line of text. This does not disturb the if (line is in morse) test, -as the temporary l[0] will remain != 0. +Because of the *two* `while(more characters)` loops are the same, I've +exchanged the `if(line is in morse)` and the two while loops (so there +is *single* while loop). However, we must now store the `if(line is in +morse)` test for efficiency (it would be stupid to test the entire line +for every character). This is done in `l[0]` ! (the first encoding is +thus in `l[1]`). The `memchr()` decoding lookup is not disturbed, because the +test is always `0` for Morse lines, and `'\0'` is not a possible encoding. +`l[0]` is also used as the temporary shift character while encoding a +line of text. This does not disturb the `if (line is in morse)` test, +as the temporary `l[0]` will remain `!= 0`. -I've put the if's in the above layout into a single switch. Hope you don't +I've put the `if`s in the above layout into a single switch. Hope you don't mind ;-) Summary of the use of l[999]: +``` l[0] result of "is this line in morse" test l[0] temporary shift during encoding l[0] - l[33] Morse codings (letters and digits, plus garbage) @@ -161,13 +186,14 @@ l[11] - l[14] strspn argument l[12] - l[14] strspn argument l[34] terminator for backward decoding l[35] - l[998] input line +``` -***** +--- Selected notes: -- We don't need to declare isalnum() (or include ctype.h) as its implicit +- We don't need to declare `isalnum()` (or include ctype.h) as its implicit declaration is correct (int argument, int result). This is not so for - strlen(), strspn(), and memchr() as they use size_t. + `strlen()`, `strspn()`, and `memchr()` as they use `size_t`. - During development, my gcc 2.7.2.3 had an internal compiler error (signal 6) on code that was correct ! - lclint 2.4b thinks that "the observer is modified" a couple of times, @@ -176,7 +202,7 @@ Selected notes: - lclint 2.4b is positive the second while loop is infinite, whereas it is not. -***** +--- .. - .... .. -. -.- - .... .. ... ... .--. --- .. .-.. . .-. .. ... -- --- .-. . --- -... ..-. ..- ... -.-. .- - . -.. - .... .- -. - .... . .--. .-. --- --. .-. .- --