About Encoding And Decoding Base-64 In FORTH

In This Previous Posting, I wrote that I had written some source-code in the language FORTH, that decodes standard Base-64 into a binary array of data, in output sizes that are multiples of 36 Bytes. For my own purposes, there might be no need to output Base-64, because I can use command-line utilities to prepare Base-64 strings, and then only use those as a means to enter the data, and embed it into future, hypothetical source code.

But the purposes of other, hypothetical software-developers have not been met with this exercise, because those people may need to be able to output Base-64, which means they’d need a matching encoder.

Unfortunately, the language does not lend itself to that easily, if a standard Base-64 radix is being implied, because 6-bit output-numerals would need to be bit-aligned, and trying to align fields of bits in FORTH is difficult.

(Edit 07/25/2017 : )

One subject which I have investigated more completely now, is the fact that the numeral-to-text conversion utilities built-in to FORTH, seem to continue to produce output, even if a Base of 64 has been set. In theory, the FORTH developers could have adopted a custom radix, in order to be able to state, that their binary-to-FB64 conversion is computed faster, than standard Base-64 could be. But OTOH, the characters output, could just become garbage, by the time 24-bit numerals are to be streamed:


dirk@Klystron:~$gforth Gforth 0.7.2, Copyright (C) 1995-2008 Free Software Foundation, Inc. Gforth comes with ABSOLUTELY NO WARRANTY; for details type license' Type bye' to exit : list-forth-b64 [ base @ decimal ] 64 base ! &255 &0 do i . space loop [ base ! ] ; ok list-forth-b64 0 1 2 3 4 5 6 7 8 9 A B C D E F G H I J K L M N O P Q R S T U V W X Y Z [ \ ] ^ _  a b c d e f g h i j k l m n o p q r s t u v 10 11 12 13 14 15 16 17 18 19 1A 1B 1C 1D 1E 1F 1G 1H 1I 1J 1K 1L 1M 1N 1O 1P 1Q 1R 1S 1T 1U 1V 1W 1X 1Y 1Z 1[ 1\ 1] 1^ 1_ 1 1a 1b 1c 1d 1e 1f 1g 1h 1i 1j 1k 1l 1m 1n 1o 1p 1q 1r 1s 1t 1u 1v 20 21 22 23 24 25 26 27 28 29 2A 2B 2C 2D 2E 2F 2G 2H 2I 2J 2K 2L 2M 2N 2O 2P 2Q 2R 2S 2T 2U 2V 2W 2X 2Y 2Z 2[ 2\ 2] 2^ 2_ 2 2a 2b 2c 2d 2e 2f 2g 2h 2i 2j 2k 2l 2m 2n 2o 2p 2q 2r 2s 2t 2u 2v 30 31 32 33 34 35 36 37 38 39 3A 3B 3C 3D 3E 3F 3G 3H 3I 3J 3K 3L 3M 3N 3O 3P 3Q 3R 3S 3T 3U 3V 3W 3X 3Y 3Z 3[ 3\ 3] 3^ 3_ 3 3a 3b 3c 3d 3e 3f 3g 3h 3i 3j 3k 3l 3m 3n 3o 3p 3q 3r 3s 3t 3u 3v ok bye dirk@Klystron:~$




My conclusion is, that This pseudo- Base-64 streaming remains usable, even when 24-bit numerals are given.

This conclusion reverses a negative, tentative conclusion, which I had only given yesterday.

I have by now coded both the encoder and decoder for standard Base-64, which I’ve named ‘b64-stream’ and ‘b64-parse’ respectively, but as well the encoder and decoder for the pseudo- Base-64, which I call ‘fb64-stream’ and ‘fb64-parse’. At this point, Base-64 has been implemented in a way software-experts would consider complete, with a full non-standard version of Base-64. This is what the code ultimately does:


dirk@Klystron:~$cd ~/Programs dirk@Klystron:~/Programs$ gforth fb64-parse-6.fs
Gforth 0.7.2, Copyright (C) 1995-2008 Free Software Foundation, Inc.
Gforth comes with ABSOLUTELY NO WARRANTY; for details type license'
Type bye' to exit
S" generate1234TERRIBLE.,?-" 4 / b64-stream b64-parse 0 4 * type generate1234TERRIBLE.,?- ok
S" generate1234TERRIBLE.,?-" 4 / fb64-stream fb64-parse 0 4 * type generate1234TERRIBLE.,?- ok
bye
dirk@Klystron:~/Programs\$




My custom-semantics assume that on the stack, a binary array exists, with a numeric value placed on top of it, which warns each encoder, how many 32-bit words each array holds. OTOH, the input to each decoder expects a standard, full string, which the corresponding encoder outputs, and which also exist as two items on the stack each time, where the top numeral states how many characters long the string is, as per standard FORTH.

And below is the source-code (Updated 08/02/2017 : )