Does anyone have a clue what .8xo (as opposed to, e.g., .8xp or .8xg) files are about? We all know about .8xp/.8xg; .tig are ZIP files containing other TI files; .TIProgram files are (as far as I know) Mac-only at the moment and represent programs in some sort of funky XML syntax (and are currently rejected on ticalc.org); but what in the world are .8xo files? I'm seeing them submitted to ticalc.org more and more lately, but haven't come across any software that can actually deal with them.

They look very much like normal .8xp/.8xg in a hex viewer to me, but I think they must have a different header or some kind of structure differences because software that can handle the usual .8x? stuff balks on these.

If they're Mac-only or whatever, then I can simply reject them like I currently do for .TIProgram (unless someone releases some sort of .TIProgram-to-.8?? converter), but I have no clue exactly where .8xo's come from or whether there is anything out there that actually handles these, so I have no idea how to handle them. Anyone know?

*.8xo seems to be a Mac-only TI Group file extension, according to this: http://tibasicdev.wikidot.com/forum/t-43978/lunar-ide

The link guide doesn't list them so it must be a more recent creation or at least a rare one. I wonder if they share the same file structure with the rest of the 8x* family. If so then it would be fairly simple to convert them to w/e format was needed. If someone with a Mac can make on and send it to me I can see what libtifiles says about it.

Travis, have you tried running one through SourceCoder? If so, what has it done with it? I wonder how difficult it would be to make a tool that would convert these to normal .8xg files.

souvik1997: Thanks for the pointer. That's some interesting information.

TheStorm: Not sure how much this might help, but this file appears to have both normal .8xg's and .8xo's of the same groups. Substantial portions of the corresponding files appear identical bytewise, though there are inserted/deleted bytes in a few places.

KermMartian: Yes. I get a message like this:



It seems to partially recognize the .8xo files, and the reported size looks plausible, but I guess the file format code is something unexpected.

(Edit: On a side note, I get the impression that .TIProgram would also be fairly easy to convert as well—the main program content appears to be TI-BASIC source code as ordinary text (with UTF-8 characters). The rest of the XML just appears to be Mac- or application-specific data that could probably be mostly ignored.)

8xo file support is a fairly recent feature request for TILP
http://sourceforge.net/tracker/?func=detail&aid=3095473&group_id=18378&atid=368378

When the feature request was posted, I downloaded several 8xo files from ticalc.org (one of them is linked in the feature request), with file names "baseballti84plus.8xo.zip", "boundti84plus.8xo.zip" and "drawti84plus.8xo.zip". I haven't seriously looked at them though, feel free to beat me

[EDIT: oh, and for the time being, I have a MacBook Pro for work, so I can try transferring the same files through TI-Connect Mac and TILP.]

Lionel, did you add the feature yourself? Travis, funnily enough, 23="GroupObj EQU 17h ;group.", which means it's sort of a group inside a group? I'm confused-ish.

Nope: as I wrote, I haven't seriously looked at the files yet

I just spent a bit more time looking at the file format for .8xo based on the sample file that Travis shared above, and it is indeed a TI-formatted file (which are all phrased as if they were groups) with a single variable inside. That variable is of type 0x17 (group), and contains the on-calc representation of a 0x17 group file. There is then a series of fields containing variables, each starting with this header:

offset 0: variable type
offset 1: ?? I'd imagine some of this is the archived flag and the version byte
offset 6: the name seems to start here but zero-terminated instead of zero-padded? Will bear further examination
offset 9: variable contents

I'll continue to investigate.

Edit: I'm an idiot. It's just the VAT entry for each variable, reversed, just as archived variables are stored in ROM. This will be somewhat nontrivial to add to SourceCoder, but I'll give it a try.

I've just created a group on the 84+SE, from the identical trivial BASIC programs "ABCDEFGH" and "HGFEDCBA" (you're all free to laugh at my TI-Z80 BASIC skills - I didn't even bother using a loop to reduce the size ), and transferred it through TI Connect X, yielding a 8xo file. You can download it from http://lpg.ticalc.org/prj_tilp/beta/GRP_TI-84_Plus_Silver_Edition.8xo

Yup, it's as I expected, it contains the literal on-calc group file rather than the normal .8xg format. Why they would decide to suddenly break the standard file format is beyond me, and utterly ridiculous. Nevertheless, once I finish the Obliterate bits that I'm working on right now, I'll update SourceCoder to understand this new format,.

Yeah, TI is strange. At one point, BrandonW and I realized that my version of Mac TI connect likes to send apps back as if they were 68k binary formated. And it doesn't like to do it all the time. Sometimes, it sends back just fine, others, it fails. Same data, just not intel hex. I had to figure out how to convert it at one point.

Yes indeed. Also, I've spent two or three hours now while watching Die Hard 2 and adding .8xo support to SourceCoder 2.5, and it seems to be working very well. Pictures, programs, and all that fun stuff all seem good. It took a lot of tweaking to get everything working, unfortunately, and hopefully I succeeded. Please feel free to test it out and tell me how different .8xo files work.

Awesome, and thanks. I think I might add a page in the ticalc.org FAQ about these unusual file types and what to do about them.

Just out of curiosity, has anyone looked at .TIProgram files yet? I tried copying and pasting program text from a few into SourceCoder's editor, and it almost works except for minor syntax differences (e.g., it uses actual subscript characters for Y1, etc., and I think SourceCoder expects something like “{Y1}”. It seems that some basic XML parsing and fairly simple substring substitutions could make these openable in the editor, at least, for conversion into a proper 8x? file.

Here are a few random ones that previously slipped into the archives:
http://www.ticalc.org/archives/files/fileinfo/411/41159.html
http://www.ticalc.org/archives/files/fileinfo/406/40681.html
http://www.ticalc.org/archives/files/fileinfo/412/41266.html
http://www.ticalc.org/archives/files/fileinfo/332/33210.html

And yeah, I don't get TI's thing with coming up with incompatible file formats for every PC platform for no apparent reason. It's as if they're just making a lame attempt to try to fragment the TI community or something.

Most welcome! I hope that it works for you, and that I didn't miss anything in particular, given the limited subset of samples that I'm working with. It seemed to work for that physics program and for Lionel's two-program .8xo, for instance. Sounds like TI is somehow trying to compete with SourceCoder and Tokens with their .TIProgram nonsense, which I would find downright entertaining. I'm depressed at the massive horror of spaghetti code that SourceCoder is becoming with all these different formats and features.

Kerm, actually they've had the TI-Program Editor for a while, it was actually removed from the latest windows version of Ti-Connect but the OS X version was actually expanded in the last release iirc. I know tifreak8x has a copy of it that can be made to run with the new version of TI-Connect if you prod it hard enough.

Though .TIProgram predates both by quite a while (I was using it as early as Jan/Feb 2005) It seems more like lackadaisical planning on their part.

@both Jon and Thomas: Ah, I didn't realize that was just the editor that's been in TI-Connect for quite some time. My mistake; thanks for pointing that out.

No worries, it's easy to forget about with the fragmentation between TI-Connect X and TI-Connect.

Yeah they are removing features from one while adding them to the other.