1
Everything-Else / Re: Infocom Source Code Released (in ZIL language)
« on: April 16, 2019, 10:11:13 PM »
I for one had a blast trying to parse the ZIL files. In the "read and make sense" sense, not the "make a compiler" sense.
Show Posts
This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.
2
The Games and other Sierra Adventure stuff / Re: What are we working on?
« on: April 13, 2019, 04:10:13 AM »
Perfectly well, thank you. It's not exactly drunk coding.
As you know, text resources are quite under-used in SCI11, in favor of messages. The only text resources you can expect to find are some debug format strings. Now, my toolset includes a message converter that can go from the raw message resource to a text file in the same format as SV and back again, with the added ability to have defined names instead of numbers, comment lines, references, and of course UTF8 if the file has a magic marker as the first line.
If I were to make a text resource counterpart and made it use SV-style output too, you'd get this visually strange case where each entry has a number key, but they're not actually part of the file and don't add anything when editing the data for later conversion back to a resource. The only reason to keep them in is to allow multi-line entries: if a line starts with a tab, it's part of the current entry (the message table parser uses a bunch of tabs, and still allows more to be part of the line proper). But if I remove the superfluous numbers, I can't easily tell if a newline ends the current entry, or is actually part of the entry.
As you know, text resources are quite under-used in SCI11, in favor of messages. The only text resources you can expect to find are some debug format strings. Now, my toolset includes a message converter that can go from the raw message resource to a text file in the same format as SV and back again, with the added ability to have defined names instead of numbers, comment lines, references, and of course UTF8 if the file has a magic marker as the first line.
If I were to make a text resource counterpart and made it use SV-style output too, you'd get this visually strange case where each entry has a number key, but they're not actually part of the file and don't add anything when editing the data for later conversion back to a resource. The only reason to keep them in is to allow multi-line entries: if a line starts with a tab, it's part of the current entry (the message table parser uses a bunch of tabs, and still allows more to be part of the line proper). But if I remove the superfluous numbers, I can't easily tell if a newline ends the current entry, or is actually part of the entry.
Code: [Select]
2 "Leisure Suit Larry in the Land of the Lounge Lizards" contains some elements of plot which may not be considered appropriate for some children.
3 Use the TAB key to select,
then ENTER to continue.
4
5 Sorry, but this game can only be played by adults, or with an adult. Please find an adult, come back, and try again.I could of course just shrug and make it use escaped newlines to mean the latter?Code: [Select]
"Leisure Suit Larry in the Land of the Lounge Lizards" contains some elements of plot which may not be considered appropriate for some children.
Use the TAB key to select,\nthen ENTER to continue.
Sorry, but this game can only be played by adults, or with an adult. Please find an adult, come back, and try again."But Kawa, what if you want to use the text files to look up which tuple to use?" Well, besides comment lines, obviously you could enable line numbers in your text editor and subtract one to make it zero-based... or you just frikken import the resource and look it up from inside SCI Companion?
3
The Games and other Sierra Adventure stuff / Re: What are we working on?
« on: April 12, 2019, 10:20:13 AM »4
The Games and other Sierra Adventure stuff / Re: What are we working on?
« on: April 11, 2019, 05:01:43 PM »
What makes considerably less sense than a polygonal pathfinding patent is the one feature in SCI11+ that's the least likely to appear in ScummVM if only because The Dating Pool doesn't use it¹, and why it keeps shooting me in my goddamn face.
I mean, of course, the part where double or triple-byte UTF8 sequences cause text to measure wider than it really is. I've tried rewriting the mentioned function in C but that made it eat the first character of every line after the first (when compiled with UTF8 support, that is), with garbage taking its place after. I backported SCI32's C++ version of the same, but found I was actually pretty close in my attempt, in the sense that it gave exactly the same results.
The shit's on my github if you want it.
(¹: There are plenty more features like that. I have a branch ready to PR.)
I mean, of course, the part where double or triple-byte UTF8 sequences cause text to measure wider than it really is. I've tried rewriting the mentioned function in C but that made it eat the first character of every line after the first (when compiled with UTF8 support, that is), with garbage taking its place after. I backported SCI32's C++ version of the same, but found I was actually pretty close in my attempt, in the sense that it gave exactly the same results.
The shit's on my github if you want it.
(¹: There are plenty more features like that. I have a branch ready to PR.)
5
The Games and other Sierra Adventure stuff / Re: What are we working on?
« on: April 11, 2019, 10:43:05 AM »
Someone does. But if it was based on incomplete information anyway, that regression would just be two different implementations based on incomplete information where one of them just happens to give better results. If the plan is to rewrite it based on complete information, that'd be even better.
At least that's how I understood it.
At least that's how I understood it.
6
The Games and other Sierra Adventure stuff / Re: What are we working on?
« on: April 10, 2019, 08:11:19 PM »
I don't feel like that counts :\
7
The Games and other Sierra Adventure stuff / Re: What are we working on?
« on: April 08, 2019, 06:10:07 AM »
Did I?
8
SCI Syntax Help / Re: Parsing Noun/Qualifying Adjective Words
« on: April 06, 2019, 10:27:50 AM »
I'll just write a general-use minisembler for starters... oops, I already did!
9
SCI Syntax Help / Re: Parsing Noun/Qualifying Adjective Words
« on: April 06, 2019, 08:21:29 AM »
If I were to make a tool to compile those grammar files, should it be a minisembler for the 80s version or a compiler for the 90s version?
10
The Games and other Sierra Adventure stuff / Re: What are we working on?
« on: April 05, 2019, 04:56:31 PM »
So it seems they're working on remaking ScummVM's pathfinder in SSCI's image. I'm gonna be watching the commits even closer now!
And hey. Thanks for the ❤️, OmerMor
And hey. Thanks for the ❤️, OmerMor
11
SCI Syntax Help / Re: Parsing Noun/Qualifying Adjective Words
« on: April 05, 2019, 03:12:18 PM »
MASM is Microsoft's Macro Assembler, which they also used for the SCI16 (going by archive name) interpreter proper. Not all assemblers are made equal, and the specific syntax won't work on various other assemblers. The alternative from 1990 is Lisp S-Expressions, much like SC script code, not BNF. It'd be trivial to make a grammar compiler that can handle the 1990 version, and I've written a minimal data-only "assembler" for a completely unrelated project years back so I can confirm it'd be trivial to make one for the older grammars too, without the link and exe2bin steps.
Also, going by language.h you missed Italian and Portuguese, I guess.
Edit: I just assembled a vocab.900 myself with MASM (though it's ML here), with only the ASM file as a parameter. I got some linkage warnings, but otherwise I have an OBJ and EXE with the raw data inside. For the OBJ, the raw data started at offset 0x30, offset 0x200 in the EXE. The latter has no trailing bits. From a cursory comparison with a randomly-chosen SCI0 game it seems about right.
Also, going by language.h you missed Italian and Portuguese, I guess.
Edit: I just assembled a vocab.900 myself with MASM (though it's ML here), with only the ASM file as a parameter. I got some linkage warnings, but otherwise I have an OBJ and EXE with the raw data inside. For the OBJ, the raw data started at offset 0x30, offset 0x200 in the EXE. The latter has no trailing bits. From a cursory comparison with a randomly-chosen SCI0 game it seems about right.
12
SCI Syntax Help / Re: Parsing Noun/Qualifying Adjective Words
« on: April 05, 2019, 06:47:37 AM »
So they went from assembly to lisp, huh?
13
The Games and other Sierra Adventure stuff / Re: What are we working on?
« on: April 03, 2019, 04:22:54 PM »File a bug report to them?What I'm wondering is
how did nobody discover this several years ago?
14
The Games and other Sierra Adventure stuff / Re: What are we working on?
« on: April 03, 2019, 11:29:41 AM »
I figured out it's something to do with the determination of the distance between the actor and the considered maybe-nearest point. Since the game checks like 30000 pixels away and the distance function cuts off at 4096, every point was considered equally near. So it picked the first, which was off to the left.
And only for contained access.
And only for contained access.
15
The Games and other Sierra Adventure stuff / Re: What are we working on?
« on: April 03, 2019, 08:08:06 AM »
I've been trying to figure out why in a bunch of different SCI11 games, both fan and original, keyboard control is shot when playing in ScummVM. And I have officially reached my limit.
https://twitter.com/NoxicoDev/status/1112671891246977024 - the discovery
https://twitter.com/NoxicoDev/status/1112796555310972929 - the curveball
https://twitter.com/NoxicoDev/status/1112868008949682176 - the endless logging
https://twitter.com/NoxicoDev/status/1113407856236216320 - the wall... and a revelation!
Turns out ScummVM doesn't handle Contained polygons very well.
https://twitter.com/NoxicoDev/status/1112671891246977024 - the discovery
https://twitter.com/NoxicoDev/status/1112796555310972929 - the curveball
https://twitter.com/NoxicoDev/status/1112868008949682176 - the endless logging
https://twitter.com/NoxicoDev/status/1113407856236216320 - the wall... and a revelation!
Turns out ScummVM doesn't handle Contained polygons very well.
SMF 2.0.14 |
SMF © 2017, Simple Machines
Simple Audio Video Embedder
Page created in 0.134 seconds with 21 queries.