Longest pristine(ish) codepoint run

Question

Your challenge is to create the longest run of characters with incrementing Unicode codepoints in a pristine program as you can. To keep things simple, I'll abridge the definition for this challenge to be "any program which produces no errors but will error if any one character is removed".

Let's define a codepoint run as any contiguous substring of N characters, where 1 <= N <= program length, where the Unicode/UTF-8 codepoints increment by 1 for each character.

For example: abc, <=>, {|, and ƐƑƒƓƔ are all codepoint runs.

Your program (must be a standalone program) doesn't have to be purely a codepoint run. Instead, its score will be the length in characters of the longest codepoint run in its source code. A program dabc's score would be 3.

Note that your program does not need to actually output anything, as long as it does not error.

• Compiler warnings do not count as errors
• Your program does not need to (but still can) error if more than 1 contiguous character is removed
• Your program may take input only if the contents of the input do not affect whether or not the program errors
• Any code which is not part of the longest codepoint run is ignored in the scoring
• Highest score, as in code golf, is per-language

Answer 1

Python 3, run of 1,114,111

#coding:utf
assert len(open(__file__).read(1061))==1061
"""
	

 !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~€‚ƒ„…†‡ˆ‰Š‹ŒŽ‘’“”•–—˜™š›œžŸ ¡¢£¤¥¦§¨©ª«¬­®¯°±²³´µ¶·¸¹º»¼½¾¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖ×ØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõö÷øùúûüýþÿĀāĂ㥹ĆćĈĉĊċČčĎďĐđĒēĔĕĖėĘęĚěĜĝĞğĠġĢģĤĥĦħĨĩĪīĬĭĮįİıIJijĴĵĶķĸĹĺĻļĽľĿŀŁłŃńŅņŇňʼnŊŋŌōŎŏŐőŒœŔŕŖŗŘřŚśŜŝŞşŠšŢţŤťŦŧŨũŪūŬŭŮůŰűŲųŴŵŶŷŸŹźŻżŽžſƀƁƂƃƄƅƆƇƈƉƊƋƌƍƎƏƐƑƒƓƔƕƖƗƘƙƚƛƜƝƞƟƠơƢƣƤƥƦƧƨƩƪƫƬƭƮƯưƱƲƳƴƵƶƷƸƹƺƻƼƽƾƿǀǁǂǃDŽDždžLJLjljNJNjnjǍǎǏǐǑǒǓǔǕǖǗǘǙǚǛǜǝǞǟǠǡǢǣǤǥǦǧǨǩǪǫǬǭǮǯǰDZDzdzǴǵǶǷǸǹǺǻǼǽǾǿȀȁȂȃȄȅȆȇȈȉȊȋȌȍȎȏȐȑȒȓȔȕȖȗȘșȚțȜȝȞȟȠȡȢȣȤȥȦȧȨȩȪȫȬȭȮȯȰȱȲȳȴȵȶȷȸȹȺȻȼȽȾȿɀɁɂɃɄɅɆɇɈɉɊɋɌɍɎɏɐɑɒɓɔɕɖɗɘəɚɛɜɝɞɟɠɡɢɣɤɥɦɧɨɩɪɫɬɭɮɯɰɱɲɳɴɵɶɷɸɹɺɻɼɽɾɿʀʁʂʃʄʅʆʇʈʉʊʋʌʍʎʏʐʑʒʓʔʕʖʗʘʙʚʛʜʝʞʟʠʡʢʣʤʥʦʧʨʩʪʫʬʭʮʯʰʱʲʳʴʵʶʷʸʹʺʻʼʽʾʿˀˁ˂˃˄˅ˆˇˈˉˊˋˌˍˎˏːˑ˒˓˔˕˖˗˘˙˚˛˜˝˞˟ˠˡˢˣˤ˥˦˧˨˩˪˫ˬ˭ˮ˯˰˱˲˳˴˵˶˷˸˹˺˻˼˽˾˿̴̵̶̷̸̡̢̧̨̛̖̗̘̙̜̝̞̟̠̣̤̥̦̩̪̫̬̭̮̯̰̱̲̳̹̺̻̼͇͈͉͍͎̀́̂̃̄̅̆̇̈̉̊̋̌̍̎̏̐̑̒̓̔̽̾̿̀́͂̓̈́͆͊͋͌̕̚ͅ͏͓͔͕͖͙͚͐͑͒͗͛ͣͤͥͦͧͨͩͪͫͬͭͮͯ͘͜͟͢͝͞͠͡ͰͱͲͳʹ͵Ͷͷ͸͹ͺͻͼͽ;Ϳ΀΁΂΃΄΅Ά·ΈΉΊ΋Ό΍ΎΏΐΑΒΓΔΕΖΗΘΙΚΛΜΝΞΟΠΡ΢ΣΤΥΦΧΨΩΪΫάέήίΰαβγδεζηθικλμνξοπρςστυφχψωϊϋόύώϏϐϑϒϓϔϕϖϗϘϙϚϛϜϝϞϟϠϡϢϣϤϥϦϧ"""

Try it online!

The triple-quoted contains the entire Unicode range except for the null byte \0 which terminates a program. (Thanks to Chas Brown for suggesting triple-quoting, which allowed 13 more characters.) The TIO link uses a shortened version; assume that the number 1061 is replaced with the code length.

The program opens itself and checks that it has the appropriate length, which guards against any deletion. No one-character deletion can disable this logic in the code. Lowering either the number of characters read or the number checked for equality will make the assertion fail.

I'm not totally sure if all Unicode characters except '\0' are valid in Python code, and I haven't been able to test it. Maybe this is system dependent. Please let me know if this fails at some point.

We can also do this without reading the file:

s="""
	

 !"#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_`abcdefghijklmnopqrstuvwxyz{|}~€‚ƒ„…†‡ˆ‰Š‹ŒŽ‘’“”•–—˜™š›œžŸ ¡¢£¤¥¦§¨©ª«¬­®¯°±²³´µ¶·¸¹º»¼½¾¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖ×ØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõö÷øùúûüýþÿĀāĂ㥹ĆćĈĉĊċČčĎďĐđĒēĔĕĖėĘęĚěĜĝĞğĠġĢģĤĥĦħĨĩĪīĬĭĮįİıIJijĴĵĶķĸĹĺĻļĽľĿŀŁłŃńŅņŇňʼnŊŋŌōŎŏŐőŒœŔŕŖŗŘřŚśŜŝŞşŠšŢţŤťŦŧŨũŪūŬŭŮůŰűŲųŴŵŶŷŸŹźŻżŽžſƀƁƂƃƄƅƆƇƈƉƊƋƌƍƎƏƐƑƒƓƔƕƖƗƘƙƚƛƜƝƞƟƠơƢƣƤƥƦƧƨƩƪƫƬƭƮƯưƱƲƳƴƵƶƷƸƹƺƻƼƽƾƿǀǁǂǃDŽDždžLJLjljNJNjnjǍǎǏǐǑǒǓǔǕǖǗǘǙǚǛǜǝǞǟǠǡǢǣǤǥǦǧǨǩǪǫǬǭǮǯǰDZDzdzǴǵǶǷǸǹǺǻǼǽǾǿȀȁȂȃȄȅȆȇȈȉȊȋȌȍȎȏȐȑȒȓȔȕȖȗȘșȚțȜȝȞȟȠȡȢȣȤȥȦȧȨȩȪȫȬȭȮȯȰȱȲȳȴȵȶȷȸȹȺȻȼȽȾȿɀɁɂɃɄɅɆɇɈɉɊɋɌɍɎɏɐɑɒɓɔɕɖɗɘəɚɛɜɝɞɟɠɡɢɣɤɥɦɧɨɩɪɫɬɭɮɯɰɱɲɳɴɵɶɷɸɹɺɻɼɽɾɿʀʁʂʃʄʅʆʇʈʉʊʋʌʍʎʏʐʑʒʓʔʕʖʗʘʙʚʛʜʝʞʟʠʡʢʣʤʥʦʧʨʩʪʫʬʭʮʯʰʱʲʳʴʵʶʷʸʹʺʻʼʽʾʿˀˁ˂˃˄˅ˆˇˈˉˊˋˌˍˎˏːˑ˒˓˔˕˖˗˘˙˚˛˜˝˞˟ˠˡˢˣˤ˥˦˧˨˩˪˫ˬ˭ˮ˯˰˱˲˳˴˵˶˷˸˹˺˻˼˽˾˿̴̵̶̷̸̡̢̧̨̛̖̗̘̙̜̝̞̟̠̣̤̥̦̩̪̫̬̭̮̯̰̱̲̳̹̺̻̼͇͈͉͍͎̀́̂̃̄̅̆̇̈̉̊̋̌̍̎̏̐̑̒̓̔̽̾̿̀́͂̓̈́͆͊͋͌̕̚ͅ͏͓͔͕͖͙͚͐͑͒͗͛ͣͤͥͦͧͨͩͪͫͬͭͮͯ͘͜͟͢͝͞͠͡ͰͱͲͳʹ͵Ͷͷ͸͹ͺͻͼͽ;Ϳ΀΁΂΃΄΅Ά·ΈΉΊ΋Ό΍ΎΏΐΑΒΓΔΕΖΗΘΙΚΛΜΝΞΟΠΡ΢ΣΤΥΦΧΨΩΪΫάέήίΰαβγδεζηθικλμνξοπρςστυφχψωϊϋόύώϏϐϑϒϓϔϕϖϗϘϙϚϛϜϝϞϟϠϡϢϣϤϥϦϧ"""
assert len(s)==999

Try it online!

There isn't a way to delete a character from the assertion line while keeping it syntactically valid and preserving the length number.

Answer 2

C# (Visual C# Interactive Compiler), Score: 1114076

var s=@"Insert all characters from character 36 then on";if(s.Length!=1114076)throw new Exception();

Try it online!

Answer 3

x86 machine code, score: 1,114,112

E9 81 F7 42 00 00 01 02 ... F4 8E BF BF C3

In assembly:

main:
    jmp $+4388742 # i.e. to the end
    .byte 0,1,...,0xF4,0x8F,0xBF,0xBF
end:
    ret

The code that is executed is just a relative jump followed by a return. All the unicode codepoints are in between.

If the jump itself or any of the first three bytes of its offset are deleted, then the unicode codepoints are executed as code until it crashes. If any byte after that is removed, then it jumps past the end of the program and eventually crashes.