IRC logs for #openttd on OFTC at 2018-11-16
⏴ go to previous day
00:00:19 <frosch123> squirrel.cpp:453 looks very much like the BOM is not accounted in "size"
00:00:54 <LordAro> that's what i was thinking
00:01:54 <LordAro> the file is 14848, which is 29*512
00:02:05 <LordAro> hits a lot of the power of 2 boundaries
00:02:20 <LordAro> maybe a file like that has never happened before
00:02:51 <frosch123> BOMs are quite uncommon, aren't they?
00:03:31 <LordAro> i think i can patch this reasonably
00:06:00 <frosch123> c90 code is so ugly
00:06:29 <frosch123> also fix the "if (ret != 2) us = 0;" :)
00:06:43 <LordAro> i caught that while putting debugging in :p
00:06:44 <frosch123> it should check "size" instead of just reading 2 bytes
00:09:11 <LordAro> hmm, i don't think this would be able to read a UTF32 file :p
00:10:34 <frosch123> only oracle and other "enterprises" use utf32
00:11:04 <LordAro> still better than utf16 though
00:11:32 <LordAro> utf16 has the all the disadvantages of byte ordering in utf32, and it's also variable width
00:12:54 <frosch123> utf16 is said to be faster, but i think it is more complicated than utf8, so for me it's a case of optimsation negatively affecting the code
00:13:41 <LordAro> utf16 is usually only used because the system predates utf32
00:14:02 <LordAro> they were expecting fixed width, but then ucs4 happened
00:18:00 <LordAro> "Each file object includes any file data, and is preceded by a 512-byte header record. The file data is written unaltered except that its length is rounded up to a multiple of 512 bytes. The original tar implementation did not care about the contents of the padding bytes, and left the buffer data unaltered, but most modern tar implementations fill the extra space with zeros."
00:18:16 <LordAro> yeah, had to be a multiple of 512 AND with a BOM to cause this
00:18:22 <LordAro> that's pretty edgecasey :p
00:19:54 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro opened pull request #6971: Fix #6969: Account for BOM when reading script files https://git.io/fpGbM
00:24:38 <DorpsGek_II> [OpenTTD/OpenTTD] frosch123 requested changes for pull request #6971: Fix #6969: Account for BOM when reading script files https://git.io/fpGNU
00:30:12 <LordAro> frosch123: i'm not sure i quite understand your comment about sizes
00:30:49 <frosch123> the code uses the result of fread (number of bytes read) to check for eof
00:31:00 <frosch123> but that is pointless in tars, you have to check "size"
00:31:37 <frosch123> if you store a file of size "1" in the tar, the "size -= 2" will wrap around to 0xFFFFFFFF :)
00:31:57 <frosch123> the checks before do not catch the size "1" since fread is happy in a tar
00:36:18 *** itsallvoodoo has joined #openttd
00:38:30 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro updated pull request #6971: Fix #6969: Account for BOM when reading script files https://git.io/fpGbM
00:38:44 <LordAro> frosch123: i'm afraid the diff got a bit bigger :> telling gh to hide whitespace should make it easier
00:39:02 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro dismissed a review for pull request #6971: Fix #6969: Account for BOM when reading script files https://git.io/fpGNU
00:39:14 <frosch123> they added that to the gui
00:39:24 <frosch123> in the past you had to enter a magic url
00:40:13 *** PlanckWalk has joined #openttd
00:40:42 <frosch123> LordAro: now it skips the fread, but still does a seek -2 ?
00:43:28 <frosch123> i don't think there is a valid squirrel file with size < 3
00:43:41 <frosch123> so, maybe just check it first and return invalid file
00:43:50 <LordAro> presumably the empty file is valid?
00:43:56 <LordAro> or a couple of spaces
00:44:07 <LordAro> does nothing, but technically valid
00:48:29 <LordAro> it occurs to me that technically speaking the tar file could be lying about the length of file
00:49:02 <LordAro> i'm not sure how the tar parser handles that
00:49:35 <frosch123> if that is not checked earlier, then many things are likely broken :p
00:50:01 <LordAro> ok, i'll assume size will be happy :p
00:50:43 <frosch123> yes, i think if "size" is wrong, that should be handled earlier/elsewhere
00:54:51 <frosch123> shall we mirate from gh to sr.ht?
00:56:23 <DorpsGek_II> [OpenTTD/OpenTTD] LordAro updated pull request #6971: Fix #6969: Account for BOM when reading script files https://git.io/fpGbM
00:56:27 <frosch123> i wonder what "sr" stands for
00:59:07 <LordAro> peter1138: it's suspiciously every single thing TB wanted
01:00:13 <DorpsGek_II> [OpenTTD/OpenTTD] frosch123 approved pull request #6971: Fix #6969: Account for BOM when reading script files https://git.io/fpGx4
01:01:42 <frosch123> except it does mailing lists instead of web-based prs
01:02:49 <LordAro> he does a lot of stuff
01:03:25 <frosch123> i do not know any celebs :)
01:03:36 <LordAro> i keep coming across him
01:03:48 <LordAro> i'm not sure i've actually used anything of his
01:04:30 <LordAro> that's the one listed on his website :p
01:04:44 <LordAro> ah, his actual is in the description
01:04:50 <frosch123> well, the profile text links to another one
01:06:26 <frosch123> i wonder whether the "sir" is just to make "drew" unambiguous
01:08:09 *** Maarten has joined #openttd
01:08:58 <frosch123> he, some of his git repos say "moved to sr.ht" :)
01:11:09 <frosch123> maybe we should make a featurerequest for mattermost.sr.ht
01:31:07 <peter1138> Hmm, Muse have gone synthwave on their latest album, heh.
01:31:32 <peter1138> Complete with magenta/blue artwork.
01:31:50 <LordAro> Something Human was my favourite
01:32:06 <peter1138> I've not heard it yet, first listen.
01:32:16 <peter1138> Didn't know they even had a new album until I saw it on the shelf.
01:48:47 *** wodencafe has joined #openttd
01:56:09 *** wodencafe has joined #openttd
01:59:23 *** pitastrudl29 has joined #openttd
02:12:20 *** WWacko1976-work has joined #openttd
02:16:21 *** wodencafe has joined #openttd
03:08:15 *** TheCharlatan has joined #openttd
03:52:55 *** sim-al2 has joined #openttd
05:18:24 *** root_____ has joined #openttd
06:28:02 *** lourfire has joined #openttd
07:27:08 <sponix> /!\ AΤTN︓ Ꭲhis chaᥒᥒel һɑѕ ⅿоved to irc.frееᥒοde.net #/joiᥒ /︕\
08:23:54 <Eddi|zuHause> maybe i should join #openttd on freenode just for giggles?
09:13:42 *** chomwitt has joined #openttd
09:20:15 *** rocky1138 has joined #openttd
09:21:11 *** rocky11384497 has joined #openttd
09:40:46 <Xaroth> Eddi|zuHause: not that much going on there.
09:47:36 *** modrobert has joined #openttd
10:06:42 *** matt21347 has joined #openttd
10:20:37 *** esainane has joined #openttd
10:56:34 *** Progman has joined #openttd
11:43:37 *** Own[cpp] has joined #openttd
11:43:40 <Own[cpp]> /!＼ ATTN: This ⅽhaᥒnel һɑs moved to іrⅽ.frеenⲟⅾe․ᥒet #/join ᜵!⧹
11:47:21 <Xaroth> they do keep trying, don't they
12:01:14 <LordAro> it's been going on for months
12:01:25 <LordAro> you'd think they'd have got bored by now
12:01:45 <LordAro> you'd've thought they'd've gotten bored now
12:08:25 <Eddi|zuHause> i think the right way to handle these is to hide the messages from the user, but don't tell the spambots they've been muted, to avoid putting evolution pressure on them to circumvent the filters
12:09:28 <LordAro> the hard bit is detecting them before they message anywhere
12:09:49 <LordAro> as you can see, the existing load are immediately getting killed after their first message
12:15:25 <Eddi|zuHause> yes, it needs serverside filtering before the message is passed on to the users, not just a bot listening to messages as they're already being passed on
12:17:25 <LordAro> the trouble is that irc networks generally aren't structured to be able to do that
12:21:49 <Eddi|zuHause> maybe my understanding is a bit naive, but IRC networks are a tree, right? so every message passes at least one server before it hits any client
12:23:13 <LordAro> it's certainly possible, but it requires significant rewriting of whatever ircd they're using
12:23:22 <LordAro> which they understandably would like to avoid
12:34:04 <Eddi|zuHause> anybody use Falkon? i'm missing a function i had in Konqueror where i could right click an image and block it...
13:57:31 <DorpsGek_II> [OpenTTD/OpenTTD] matthijskooijman commented on pull request #6912: Fix #6854: Compilation with ICU 62 https://git.io/fpZwL
14:03:51 <DorpsGek_II> [OpenTTD/OpenTTD] matthijskooijman commented on pull request #6912: Fix #6854: Compilation with ICU 62 https://git.io/fpZwr
14:18:52 *** skywavesurfer has joined #openttd
14:18:55 <skywavesurfer> /!\ ATΤΝ: Thіs chaᥒnel has mഠveԁ to irc.freеnodᥱ.ᥒet #/join /!\
15:47:29 *** frosch123 has joined #openttd
16:08:57 <DorpsGek_II> [OpenTTD/OpenTTD] matthijskooijman commented on pull request #6912: Fix #6854: Compilation with ICU 62 https://git.io/fpZS6
16:15:41 *** sla_ro|master has joined #openttd
16:57:45 *** Wormnest has joined #openttd
17:27:03 *** ChanServ sets mode: +v tokai
18:20:18 *** andythenorth has joined #openttd
18:20:25 *** HerzogDeXtEr has joined #openttd
18:20:55 *** triolus has joined #openttd
19:02:46 *** andythenorth has joined #openttd
19:02:52 *** theglass10 has joined #openttd
19:02:53 <theglass10> /！⧹ AΤTN: Τhiѕ channel һas mo⋁eԁ tഠ іrс．frеeᥒοdе．net #/ϳഠin /ǃ\
19:09:59 <erikd> /!\ AТΤN： This chɑᥒᥒеⅼ һaѕ ⅿoᴠeԁ to irⅽ.frᥱеᥒoԁe．ᥒet #/ϳⲟіn ／!\
19:12:56 <andythenorth> oh we've moved? o_O
19:13:06 <andythenorth> which unicode characters do I move to?
19:21:22 *** WWacko1976-work has quit IRC
20:11:02 *** gelignite has joined #openttd
20:23:12 <andythenorth> shall we make openttd?
20:25:20 <Eddi|zuHause> openttd is overrated
20:34:03 <andythenorth> over-rating is overrated
20:34:10 *** Wacko1976 has joined #openttd
20:34:10 <andythenorth> so that levels everything
22:08:56 *** Dunedan has joined #openttd
22:08:56 <Dunedan> /﹗\ ATTΝ: Thiѕ chaᥒᥒel hɑs mഠ⋁ed tο irc.frᥱeᥒoԁе.nᥱt #/jഠⅰn /!﹨
22:53:40 *** Samu is now known as Xarick
22:55:04 <Xarick> isn't there a way to automatically identify myself
22:55:19 *** Xarick is now known as Samu
22:59:00 <Samu> NoNoCAB isn't doing that great of a job imo
22:59:10 <Samu> was supposed to be the king of profit
23:00:38 <Samu> was hoping to find him here
continue to next day ⏵