End-to-End Workflow & Clean-Up | Bitwise Episode Guide

0:09Recap and set the stage for the day

🗩

0:09Recap and set the stage for the day

🗩

0:09Recap and set the stage for the day

🗩

1:10Review the C code generation work done on Day 010 Extra¹

📖

1:10Review the C code generation work done on Day 010 Extra¹

📖

1:10Review the C code generation work done on Day 010 Extra¹

📖

2:37Run the program to see our generated output

🏃

2:37Run the program to see our generated output

🏃

2:37Run the program to see our generated output

🏃

3:16Toggle on example_test()

3:43Run it to see the result of example_test()

🏃

3:43Run it to see the result of example_test()

🏃

3:43Run it to see the result of example_test()

🏃

3:54Continue to review the C code generation work done on Day 010 Extra²

📖

3:54Continue to review the C code generation work done on Day 010 Extra²

📖

3:54Continue to review the C code generation work done on Day 010 Extra²

📖

8:45Determine to work on the end-to-end workflow

🗩

8:45Determine to work on the end-to-end workflow

🗩

8:45Determine to work on the end-to-end workflow

🗩

10:46Create ion.c and introduce ion_compile_file() and ion_compile_str(), and the general utility read_file()

27:28badflydog pervognsen You should probably use "rb" in fread, I believe the CRT messes with newlines otherwise

🗪

27:28badflydog pervognsen You should probably use "rb" in fread, I believe the CRT messes with newlines otherwise

🗪

27:28badflydog pervognsen You should probably use "rb" in fread, I believe the CRT messes with newlines otherwise

🗪

27:31Fix read_file() to pass "rb" as the second argument to fopen()

27:42Run it to see that it works

🏃

27:42Run it to see that it works

🏃

27:42Run it to see that it works

🏃

27:50Explain how fopen() works on Windows

🗩

27:50Explain how fopen() works on Windows

🗩

27:50Explain how fopen() works on Windows

🗩

28:44nothings2 Unless it's actually a single-line file, I suspect it's actually the ^Z at the end of the file, not the CR-LF stuff

🗪

28:44nothings2 Unless it's actually a single-line file, I suspect it's actually the ^Z at the end of the file, not the CR-LF stuff

🗪

28:44nothings2 Unless it's actually a single-line file, I suspect it's actually the ^Z at the end of the file, not the CR-LF stuff

🗪

29:37Introduce ion_test()

30:18nothings2 pervognsen The loop should be totally unnecessary for reading an actual file, although reading from a tty or network stream redirected or something, sure

🗪

30:18nothings2 pervognsen The loop should be totally unnecessary for reading an actual file, although reading from a tty or network stream redirected or something, sure

🗪

30:18nothings2 pervognsen The loop should be totally unnecessary for reading an actual file, although reading from a tty or network stream redirected or something, sure

🗪

31:11Create test1.ion

31:38nothings2 For what it's worth, Posix says "Upon successful completion, fread() shall return the number of elements successfully read which is less than items only if a read error or end-of-file is encountered."

🗪

31:38nothings2 For what it's worth, Posix says "Upon successful completion, fread() shall return the number of elements successfully read which is less than items only if a read error or end-of-file is encountered."

🗪

31:38nothings2 For what it's worth, Posix says "Upon successful completion, fread() shall return the number of elements successfully read which is less than items only if a read error or end-of-file is encountered."

🗪

31:51Make read_file() read in the entire file in one call

32:29james4k Zeal looks nice, thanks whoever pointed that out

🗪

32:29james4k Zeal looks nice, thanks whoever pointed that out

🗪

32:29james4k Zeal looks nice, thanks whoever pointed that out

🗪

32:51Run it to see that it still works

🏃

32:51Run it to see that it still works

🏃

32:51Run it to see that it still works

🏃

33:10Enable ion_compile_file() to handle empty files

34:37Run it to see that it actually worked

🏃

34:37Run it to see that it actually worked

🏃

34:37Run it to see that it actually worked

🏃

35:08Enable read_file() to handle 0-length files

36:14Run it to see that it works

🏃

36:14Run it to see that it works

🏃

36:14Run it to see that it works

🏃

36:36Add a load of test code to test1.ion

38:30badflydog Probably Intel shortcuts, I believe you can disable them in the Intel control panel

🗪

38:30badflydog Probably Intel shortcuts, I believe you can disable them in the Intel control panel

🗪

38:30badflydog Probably Intel shortcuts, I believe you can disable them in the Intel control panel

🗪

39:06Finish up adding test code to test1.ion

39:38Run it to see our generated output

🏃

39:38Run it to see our generated output

🏃

39:38Run it to see our generated output

🏃

39:43twitchplaysp0ng Will there be a syntax definition for Ion?

🗪

39:43twitchplaysp0ng Will there be a syntax definition for Ion?

🗪

39:43twitchplaysp0ng Will there be a syntax definition for Ion?

🗪

40:03Introduce write_file()

43:34Run it to find that it works

🏃

43:34Run it to find that it works

🏃

43:34Run it to find that it works

🏃

44:02Enjoy MSCV's regex-matching search

💢

44:02Enjoy MSCV's regex-matching search

💢

44:02Enjoy MSCV's regex-matching search

💢

45:05Introduce get_ext() and replace_ext()

52:17Run it to see how these file naming functions work

🏃

52:17Run it to see how these file naming functions work

🏃

52:17Run it to see how these file naming functions work

🏃

53:04Enable ion_compile_file() to write out to file

54:12Run it to find that we successfully write out a .c file

🏃

54:12Run it to find that we successfully write out a .c file

🏃

54:12Run it to find that we successfully write out a .c file

🏃

55:00badflydog In before POSIX strdup warning

🗪

55:00badflydog In before POSIX strdup warning

🗪

55:00badflydog In before POSIX strdup warning

🗪

55:09badflydog pervognsen You should use "wb" in write file as well

🗪

55:09badflydog pervognsen You should use "wb" in write file as well

🗪

55:09badflydog pervognsen You should use "wb" in write file as well

🗪

55:54Temporarily change write_file() to pass "wb" to fopen(), and inspect the (undesired) generated output

56:59Create generate_test.py to facilitate the production of many lines of Ion code

1:02:01heasummn pervognsen Do we want to focus heavily on optimisation in the C version since we will rewrite it eventually?

🗪

1:02:01heasummn pervognsen Do we want to focus heavily on optimisation in the C version since we will rewrite it eventually?

🗪

1:02:01heasummn pervognsen Do we want to focus heavily on optimisation in the C version since we will rewrite it eventually?

🗪

1:03:00Change generate_test.py to use regex

1:05:20Run it to see that it works

🏃

1:05:20Run it to see that it works

🏃

1:05:20Run it to see that it works

🏃

1:05:25Enable generate_test.py to generate a fair amount of code

1:07:42Determine to compare the compilation times of gcc and ion

🏃

1:07:42Determine to compare the compilation times of gcc and ion

🏃

1:07:42Determine to compare the compilation times of gcc and ion

🏃

1:08:10Enable syntax_error() and fatal_syntax_error() to report the line number containing the error

1:10:02Run it and hit an error on line 1

🏃

1:10:02Run it and hit an error on line 1

🏃

1:10:02Run it and hit an error on line 1

🏃

1:10:11Enable main() to call init_keywords(), the latter to set arena_end after calling KEYWORD(typedef), and increase the ARENA_BLOCK_SIZE from 1024 to 1024 * 1024

1:12:55Run it to find that it does the right thing

🏃

1:12:55Run it to find that it does the right thing

🏃

1:12:55Run it to find that it does the right thing

🏃

1:13:25Introduce ion_main(), and print out compilation success / failure

1:16:00Run it (successfully) on test1.ion and (unsuccessfully) on test2.ion

🏃

1:16:00Run it (successfully) on test1.ion and (unsuccessfully) on test2.ion

🏃

1:16:00Run it (successfully) on test1.ion and (unsuccessfully) on test2.ion

🏃

1:16:52Enable init_stream() and the error handling functions to handle anonymous sources (i.e. ones lacking a file path)

1:20:14Run it and no longer hit the "FATAL: Undeclared name" error

🏃

1:20:14Run it and no longer hit the "FATAL: Undeclared name" error

🏃

1:20:14Run it and no longer hit the "FATAL: Undeclared name" error

🏃

1:20:58Make ion_compile_file() pass the path to init_stream()

1:21:04Run it to find that it all works as desired

🏃

1:21:04Run it to find that it all works as desired

🏃

1:21:04Run it to find that it all works as desired

🏃

1:21:18Introduce SrcLoc struct, and pull fatal(), syntax_error() and fatal_syntax_error() out into a new error.c file, and introduce loc_syntax_error()

1:28:00Run it to find that it all works

🏃

1:28:00Run it to find that it all works

🏃

1:28:00Run it to find that it all works

🏃

1:28:08Rename loc_syntax_error() to error(), make resolve_typespec() call it, and provoke a "must denote a type" error

1:31:09Run it, correctly hit the error, but segfault

🏃

1:31:09Run it, correctly hit the error, but segfault

🏃

1:31:09Run it, correctly hit the error, but segfault

🏃

1:31:16Introduce fatal_error() for resolve_typespec() to call

1:31:57Run it to find that it works

🏃

1:31:57Run it to find that it works

🏃

1:31:57Run it to find that it works

🏃

1:32:06Q&A

🗩

1:32:06Q&A

🗩

1:32:06Q&A

🗩

1:32:20heasummn pervognsen You do sizeof(p), that should have a (?)

🗪

1:32:20heasummn pervognsen You do sizeof(p), that should have a (?)

🗪

1:32:20heasummn pervognsen You do sizeof(p), that should have a (?)

🗪

1:32:29Fix generate_test.py as per heasummn's suggestion

1:33:11Run it find that it worked

🏃

1:33:11Run it find that it worked

🏃

1:33:11Run it find that it worked

🏃

1:33:13Make generate_test.py produce much more code

1:33:18Run it to find that it compiles slowly

🏃

1:33:18Run it to find that it compiles slowly

🏃

1:33:18Run it to find that it compiles slowly

🏃

1:34:45praet_a51 Also path is declared but not used

🗪

1:34:45praet_a51 Also path is declared but not used

🗪

1:34:45praet_a51 Also path is declared but not used

🗪

1:35:04Fix ion_main() to pass path to ion_compile_file()

1:35:13cubercaleb pervognsen Can you run a test with -O2 or -O3 to see how fast ion is with optimizations?

🗪

1:35:13cubercaleb pervognsen Can you run a test with -O2 or -O3 to see how fast ion is with optimizations?

🗪

1:35:13cubercaleb pervognsen Can you run a test with -O2 or -O3 to see how fast ion is with optimizations?

🗪

1:35:22Compile Ion with -O3 to see that it's still slower than it should be

1:35:51xanatos387 As far as speed goes, we still don't have a hashtable interning right?

🗪

1:35:51xanatos387 As far as speed goes, we still don't have a hashtable interning right?

🗪

1:35:51xanatos387 As far as speed goes, we still don't have a hashtable interning right?

🗪

1:36:09Make generate_test.py produce 16× more code

1:36:41Run it to find that it takes forever

🏃

1:36:41Run it to find that it takes forever

🏃

1:36:41Run it to find that it takes forever

🏃

1:36:51praet_a51 Could trading from disk itself be an issue as well

🗪

1:36:51praet_a51 Could trading from disk itself be an issue as well

🗪

1:36:51praet_a51 Could trading from disk itself be an issue as well

🗪

1:37:40twitchplaysp0ng Where in the code is n^2

🗪

1:37:40twitchplaysp0ng Where in the code is n^2

🗪

1:37:40twitchplaysp0ng Where in the code is n^2

🗪

1:38:58heasummn Could you use Visual Studio's profiler to see what's causing the most slowdown?

🗪

1:38:58heasummn Could you use Visual Studio's profiler to see what's causing the most slowdown?

🗪

1:38:58heasummn Could you use Visual Studio's profiler to see what's causing the most slowdown?

🗪

1:39:01See how long it took

🏃

1:39:01See how long it took

🏃

1:39:01See how long it took

🏃

1:39:35praet_a51 Are we optimising a bit today or will that start of tomorrow

🗪

1:39:35praet_a51 Are we optimising a bit today or will that start of tomorrow

🗪

1:39:35praet_a51 Are we optimising a bit today or will that start of tomorrow

🗪

1:39:41godling72 pervognsen Isn't a bit of premature optimization? You aren't close to having 650k of real ion code?

🗪

1:39:41godling72 pervognsen Isn't a bit of premature optimization? You aren't close to having 650k of real ion code?

🗪

1:39:41godling72 pervognsen Isn't a bit of premature optimization? You aren't close to having 650k of real ion code?

🗪

1:40:06Stop recording

🗩

1:40:06Stop recording

🗩

1:40:06Stop recording

🗩

Bitwise

Keyboard Navigation

Global Keys

Menu toggling

In-Menu Movement

Quotes and References Menus

Quotes, References and Credits Menus

Filter Menu

Filter and Link Menus

Credits Menu