|
Members in Shoutbox
None.
Shoutbox Search
Shoutbox Commands
/w [name] > Whisper
/r > Reply to last whisper /me > Marks as action Shoutbox Information
Moderators may delete any and all shouts at will.
|
Global Shoutbox
Please log in to shout.
[2019-5-30. : 12:49 pm]  NudeRaider -- NudeRaiderNudeRaider shouted: okay, so at best you can double the number of chars in the string section, but only if you find a way to fill the ONE string with useful substrings? my reasoning is this: Can only address strings until 64k, but when the last string starts at 64k you can adress substrings again 64k deep. that's twice the normal limit. Now if that last string, is 64k and contains useful substrings you essentially double the amount of useful chars. But...[2019-5-30. : 12:47 pm] jjf28 -- NudeRaiderNudeRaider shouted: " the one string is only 130053 bytes long," why? Aren't you saying it's almost 4GiB long? let's say you didn't care about hitting maximum number of strings, then you use no offset that contains a zero byte, ergo no '\0' byte, then if you avoid using any nul terminators your only limit is the section size (4gbs?, maybe 2gbs?)[2019-5-30. : 12:45 pm] NudeRaider -- " the one string is only 130053 bytes long," why? Aren't you saying it's almost 4GiB long?[2019-5-30. : 12:45 pm] jjf28 -- and when I focus on maxing number of unique strings I need to use offsets that contain a zero (contain a null) and forcibly break up the last string after 130053 bytes[2019-5-30. : 12:44 pm] jjf28 -- but the 4gbs model is focused on maxing characters, I'm focused on maxing number of strings[2019-5-30. : 12:44 pm] jjf28 -- NudeRaiderNudeRaider shouted: jjf28 but only under 1 condition: the last substring of that ONE string, starts somewhere "before the byte" AND is "6,375,767,045" (-65536) bytes long, right? no, the one string is only 130053 bytes long, back here you're asking a question that's ambiguous to me NudeRaiderNudeRaider shouted: okay, so at best you can double the number of chars in the string section, but only if you find a way to fill the ONE string with useful substrings? your one string could be 4gbs long if you wanted to[2019-5-30. : 12:41 pm] NudeRaider -- jjf28jjf28 shouted: NudeRaider at best you could pack 6,375,767,045 characters worth of strings there but only under 1 condition: the last substring of that ONE string, starts somewhere "before the byte" AND is "6,375,767,045" (-65536) bytes long, right? [2019-5-30. : 12:40 pm] jjf28 -- NudeRaiderNudeRaider shouted: okay, so at best you can double the number of chars in the string section, but only if you find a way to fill the ONE string with useful substrings? at best you could pack 6,375,767,045 characters worth of strings there [2019-5-30. : 12:38 pm] NudeRaider -- okay, so at best you can double the number of chars in the string section, but only if you find a way to fill the ONE string with useful substrings?[2019-5-30. : 12:38 pm] jjf28 -- yes, so long as your substrings start somewhere after the start of that one string and before the 65536th byte[2019-5-30. : 12:35 pm] jjf28 -- NudeRaiderNudeRaider shouted: " green being all bytes after 65536" didn't you just say you can't address them? jjf28 same here, even if you could, you couldn't address it, right? yes you can't address them, but you can point to a byte before 65536 and have the string continue past that until the next NUL character[2019-5-30. : 12:34 pm] NudeRaider -- " green being all bytes after 65536" didn't you just say you can't address them? jjf28jjf28 shouted: and you'll never actually have a STR section where every string fits perfectly as a substring of one giant string same here, even if you could, you couldn't address it, right?[2019-5-30. : 12:33 pm] jjf28 -- NudeRaiderNudeRaider shouted: jjf28 just a note: when someone asks you to explain in simpler terms, they are likely not interested in even more technical explanations, they want it dumbed down. Like here jjf28 he also wanted to understand the ellipses, unless he was talking about the "- " comments at the end Oh_ManOh_Man shouted: i dont even know what half those dot points mean , but w/e[2019-5-30. : 12:31 pm] NudeRaider -- jjf28jjf28 shouted: cumulatively the comma separated values are the values of the STR structure http://www.staredit.net/wiki/index.php/Scenario.chk#.22STR_.22_-_String_Data https://puu.sh/DziSD/a564687330.png, red being the number of strings, blue being all bytes before 65536, green being all bytes after 65536 just a note: when someone asks you to explain in simpler terms, they are likely not interested in even more technical explanations, they want it dumbed down. Like here jjf28jjf28 shouted: I did this to better understand the limits of the string section, and the result was a bit horrific [2019-5-30. : 12:31 pm] jjf28 -- and you'll never actually have a STR section where every string fits perfectly as a substring of one giant string[2019-5-30. : 12:30 pm] jjf28 -- it has no useful application - the fact that the last string can be 4gbs can't be abused to gain more string data, except to have one really long string there, because you can't point past the 65536th byte[2019-5-30. : 12:28 pm] NudeRaider -- question is, is that just theorycrafting - testing the limits, or could it also have a useful application as I assumed here NudeRaiderNudeRaider shouted: I'm not sure either, but I *think* they'Re talking about a quirk in the String section where the last string has no size limit. Futher I'm assuming that this could help circumventing both string limits by putting all used substrings there. ?[2019-5-30. : 12:27 pm] jjf28 -- the blue-green values are simultaneously the string offsets and the string data[2019-5-30. : 12:26 pm] jjf28 -- cumulatively the comma separated values are the values of the STR structure http://www.staredit.net/wiki/index.php/Scenario.chk#.22STR_.22_-_String_Data https://puu.sh/DziSD/a564687330.png, red being the number of strings, blue being all bytes before 65536, green being all bytes after 65536[2019-5-30. : 12:24 pm] jjf28 -- NudeRaiderNudeRaider shouted: jjf he was talking about layman's terms that's too bad, cause it's a technical layout [2019-5-30. : 12:23 pm] jjf28 -- I'm tracking imaginary hexidecimal row numbers on the left, the first row, and the 0x100th row are two bytes and 512 bytes, while every other row is 510 bytes[2019-5-30. : 12:21 pm] jjf28 -- Oh_ManOh_Man shouted: i dont even know what half those dot points mean the comma separated values are each 16-bit numbers, the values are shown in hexadecimal with little endian byte ordering, so it's a bit hard to read even if you know what you're doing, but in the dots one byte is simply counting upwards, either the left (value *1, least significant byte - LSB) or the right (value *256, most significant byte - MSB)[2019-5-30. : 12:20 pm] jjf28 -- I did this to better understand the limits of the string section, and the result was a bit horrific[2019-5-30. : 12:19 pm] jjf28 -- Oh_ManOh_Man shouted: u guys really suck at breaking things down for the layman im just gonna say that XD it's a lot to break down tbh, the summary is I've created a string section with the maximum possible number of strings, and the maximum possible amount of character data while having the maximum possible number of strings[2019-5-30. : 11:02 am] NudeRaider -- but yeah I'm confused as well and this is only speculation on my part at the moment.[2019-5-30. : 11:00 am] NudeRaider -- I'm not sure either, but I *think* they'Re talking about a quirk in the String section where the last string has no size limit. Futher I'm assuming that this could help circumventing both string limits by putting all used substrings there.[2019-5-30. : 10:18 am]  Oh_Man -- maybe this is a case of not caring how the sausage is made and only wanting to know how it tastes?[2019-5-30. : 10:17 am] Oh_Man -- u guys really suck at breaking things down for the layman im just gonna say that XD[2019-5-29. : 10:38 pm] jjf28 -- but... that would require me actually testing whether a section size over 2gb (signed shenanigans) or a section size of 4gbs even works in SC, and that would cross into the un-fun side of theorycrafting [2019-5-29. : 10:31 pm] jjf28 -- Suicidal InsanitySuicidal Insanity shouted: Ya but since the length of the last string is just limited by the 4GB process ram limit.... I suppose that's a different max, this one maxes number of strings over number of characters, could make a second version that maxes number of characters, it just leaves off all the offsets that contain a 0 byte so it's 510 fewer strings [2019-5-29. : 9:25 pm] Suicidal Insanity -- Ya but since the length of the last string is just limited by the 4GB process ram limit.... [2019-5-29. : 10:01 am] Lanthanide -- that was around the time I was investigating the EUDs for my map and I figured that the mac population playing SC was going to decline pretty drastically so I didn't bother (amongst other reasons)[2019-5-29. : 10:01 am] Lanthanide -- once Apple stopped supporting Rosetta, you couldn't play SC on new versions of MacOS without some sort of emulator[2019-5-29. : 10:00 am] Lanthanide -- Suicidal InsanitySuicidal Insanity shouted: But isn't endianness moot ever since apple used x86 and not powerPC? yes, we're talking about the original Mac port of SC to PowerPC[2019-5-29. : 8:02 am] Suicidal Insanity -- But isn't endianness moot ever since apple used x86 and not powerPC?[2019-5-29. : 6:37 am] Lanthanide -- jjf28jjf28 shouted: but yea, one and two byte values are the sticky ones that I was hoping someone was able to confirm jjf28jjf28 shouted: the addresses were always at the same offsets from the death table in the old Windows/Mac hybrids, it was just the endianness we needed to account for correct[2019-5-29. : 6:36 am] Lanthanide -- oh, I seriously investigated MAC-compatible EUDs for my desert strike map[2019-5-29. : 4:33 am] jjf28 -- but yea, one and two byte values are the sticky ones that I was hoping someone was able to confirm[2019-5-29. : 4:32 am] jjf28 -- we were running against a non-wine mac SC version on SnowLeopard and similar Mac versions (with my on-bnet play testers) when we observed that endianess needed to be flipped[2019-5-29. : 3:37 am]  O)FaRTy1billion[MM] -- afaik, back when people were trying to test there weren't a lot of mac users around so idk how much of our understanding is just conjecture[2019-5-29. : 3:37 am] O)FaRTy1billion[MM] -- if you were reading dword values it'd be fine, but endianness would matter for smaller values[2019-5-29. : 3:19 am] Arta(M) -- Trgk uses Mac and there was no problem playing EUD map together with Windows player. idk he uses Wine or something. |