If I wrote an autopatch module, would anyone bother to use it?

I'm thinking something like this:
1. Developer writes a cool game, but knows he might have to modify some of the scripts or other resources in the future, so he sticks my autopatch module in his code.
2. Each time the game is run, it generates an MD5 digest of the files the developer wants to allow automatic updating with, and writes the filenames and digests to an update file.
3. When the game is released, the developer switches the autopatcher from 'update file generation' to 'update file checking' mode and puts the message digest file online somewhere.
4. When the game is run, it (optionally - I'd be very annoyed if my game checked for updates every time I played -_-) checks the known location of the update file using the netcode and gets the list of filenames and digests.
5. The module compares the digests for the online files and the current files. If they are different, it downloads the online files into a temporary folder.
6. When the downloads are complete, the message digests are checked again. If any of the files were incompletely downloaded, or not the version the game expected to download, then the whole update is abandoned.
7. Otherwise, a backup is made of the affected files and the new files are copied over the old.

So, basically what the patchers for recent games do. I could write this (at least, once I go look up the MD5 digest algorithm) because I think it's nifty, but if there's no real demand for it, or someone else is already doing it, then I'll concentrate on something more useful.

Security considerations are also a problem. I could code a prohibition against downloading new copies of the Verge3 executable to keep people from spreading virii like that, but it wouldn't take much to work around it, since they'll have the source code. Perhaps rather than a central update file, have a single config file on the user's computer like:
art.pak http://www.awesomegame.com/update/art.md5
code.pak http://www.awesomegame.com/update/code.md5
// ...etc

and allow the user to delete any that aren't acceptable to him from the config file? I'm thinking that all of this could be handled in the code with a single function call at startup and a configuration file for the development side of things...I could write a standard interface for the updating stuff (and the user interface), and all the developer would have to do is write a configuration file showing which files need to have digests generated and find some webspace to put the digests and stuff on.

I'd better get back to work on my term project, but thoughts, concerns and suggestions are welcome.

Rather than distribute the full source, maybe you could dstribute it, or at least the sensitive components as a non-overrideable VPK?

Secondly, rather than providing the full system, perhaps just make it a series of functions that the coder can call? So they can make it a menu option, or make the auto-updating off by default.

Finally, it really ought to download the digests to the folder, then apply it from there - then you could make it check to see if there's a digest in the folder first, and apply it - that way, those who are offline can autoupdate by getting a hold of the digest.

that would be cool :)
all I would want is structs within structs and that not hard with the lists but strings are a stumbling block :P

I would make a game in verge if I would have fun doing it, but with the limitations verge programming doesn't seem fun.
Not that I don't think verge has potential and I do know of at least one full rpg that is being made now (I am helping on it) .So everyone doesn't have to make games if they don't want too.

in verge 3 strings have unlimited size so you can't assume 255 size max unlike verge 2

I've been working on something similar to this that will bomb out if you can't connect or if you get an unexpected response from the server (ie, it isn't a patch server). I haven't been working on it a whole lot though. I am planning on using it for online-only stuff.

Quote:Originally posted by Zip

Personally, I'd far prefer you made a game, that might in future need patching. Seeing as verge sports a grand total of one finished rpg, and that doesn't support netcode, I see more of a demand for kewl stuff like the linked list. Hell, grue can't even keep a version number right, so use for this seems limited.
Mak gamz not utils!

SOmeONeeLsE


there are people who do not want to make games. they are allowed to make fun things that might reduce the time it takes someone else to make a game. so i'd prefer feyr does what he wants to, and does it well like he tends to.