Aeomin

See my reply in page 6

Interesting. so something to do with permission then. Packing works here, maybe you need move patcher out ro client first or run packer in administrator. Ah, yeah it was a permission issue. I haven't disabled UAC on this new computer. 2.6.4.13b does not have any errors when executed outside OR inside the RO Client directory on a Windows 10 machine. Nice job. Oh that's the cause then.. I will need check updates on that component then

Interesting. so something to do with permission then. Packing works here, maybe you need move patcher out ro client first or run packer in administrator.

The main challenge I would say is that input box. To style windows's input box to a good looking is hard if not impossible. Custom implementation takes a lot of time. If people are okay with whatever windows's built in one, then I can just export to script API and implement the logic themselves.

What loader? Yes I do receive similar bug report. However I couldn't figure out the reason as I wasn't able to reproduce it on both windows 10 and 8Hmm, that's strange. I'll provide more information if I ever figure out anything else. Is it possible to release a version that's in debug mode so that I can get you more information on where the error is occuring at runtime? If not I suppose I'll patiently wait for version 3. The latest build included the exception handler. It's a third party component I have used long ago. Maybe the version is not compatible with windows 10. I'll reupload see if remove it would solve the problem.

Yes I do receive similar bug report. However I couldn't figure out the reason as I wasn't able to reproduce it on both windows 10 and 8

That's odd.. there is no related change between 2.6.4.9 and 2.6.4.8. I will take a look though,

Oh cool, you figured out thor format. I took a glance your implementation, it seems fine. One recommendation though. The "Mode" you are referring (0x21, 0x30) is actually version number. While there are earlier versions, they are already dropped. The plan is drop out 0x21 in the future. So even for patching patcher and client, use 0x30 is good enough. The reason I use 0x21 in thormaker only because when 0x30 came out, people needs upgrade old patcher but those of course won't able to read it. All the version after that will recognize it, there is no special treatment during patching. And what happens after all these years? "If it's not broken, don't fix it" (terrible btw). So always use 0x30 when creating thor file if that makes your life easier. Will there be any more format change? Maybe.. I do have some ideas floating around to have some kind extensible format, but I don't think it justify the needs right now. Will likely open specification if it ever happens though.

Just give me the files, I will include it. There is no special approval process.

That feature is been there for years. 3.0 will use json instead ini files though. You will need add a json file in Configuration/Languages. Take Default.json as base template.. Do note that json specification doesn't allow comment, but patcher's parser is generous enough to allow that at this moment. You will also need to modify Configuration/LanguageMap.json to map the language to locale ID. ID can be found in LCID Dec column at http://msdn.microsoft.com/goglobal/bb964664.aspx And very important: the language file needs to save in UTF-16 LE with BOM, otherwise patcher might display gibberish.

A quick update. After days of rushing through, I hereby releasing alpha version of 3.0. This is mainly for those who want get familiar with new setup earlier than completely puzzled later as document is scarce as usual. There is another reminder to make sure it sticks: ONLY USE FOR TEST PURPOSE, DO NOT RUN IN PRODUCTION ENVIRONMENT.

Thanks. I'm still scratching my head over the design for 2.7. I'm already not very happy on the new configuration. Bad sign.

Hey guys. Sorry for not being active for a long time (like 2 years?). I'm still struggling with time and patience to continue on. But will get there! Working on get back to track. Anyhow, I have decided to abandon version 2.7 and move to v3 due to major changes that are not even marginally close to backward compatible. For some, this might be good, for others maybe is "WTF R U DOIN?" For those who already knew what would be in 2.7, yes v3 is going to use json and config (maybe something else, but definitely not ini) and use lua scripts to finer control behaviors.

in 2.6.1.66 You mean that's the cause of corruption? I doubt so... https://www.mediafire.com/?rymv44z6l2122c9 The "Corrupted" folder is the original patch that failed, I believe it was the textureÀ¯ÀúÀÎÅÍÆäÀ̽ºmap[email protected]<script cf-hash='f9e31' type="text/javascript"> /* */</script> and [email protected]files that were corrupt. The "Clean Repatch" folder was my attempt to fix it by repatching the original files, but it failed to repatch them. I left the original data files in there as well so you can see that the originals weren't corrupt. As a note, I found out that if I add a patch with a different file using the same name, then patch the original file again, it usually fixes it. So for example, if HeadgearA.spr gets corrupted, I could just rename RandomHG.spr to HeadgearA.spr, patch it, then patch in the original HeadgearA.spr again and it wouldn't be corrupted anymore. It definitely seems like there's something preventing it from repatching identical files, so maybe it's integrated into the GRF library you're using or something. I haven't tested that file yet, but my speculation that somehow renaming fixed corruption would be the new file is larger than the old one? When new file file is larger, it goes to a different mechanism.. There isn't anything that prevents repatching identical file, and the GRF library is custom made.. I will test it out.

Sorry for this long delay, I have not been monitoring the forum. Was expecting getting email notification but doesn't seem to be the case. It used to be run as admin by default, but mechanism changed to run as admin when needed. Is it not working properly? Yea, that is how previous versions enforces admin privilege. But mechanism changed.. the new method might be crap though..There is another way that is more user friendly but haven't try it yet. There isn't any known issue, but I have changed a lot in the code. So it could be regression happening. If you have reproducible problematic patch file, send them to me. Well, there isn't any duplicate file detection because it would decrease performance. Yea , older versions might work.. but I do remember they have memory problem, so it might work for you but crash for others. Which old version?

Hashing related function is marked deprecated.

Currently it is at feature freeze. New feature would be possible starting next major release, which is bit overdue now. Hopefully I have some time next 2 months. I hope everything is going fine to you =) this is still the best patcher ever xD I find myself working on this project every now and then. Time and laziness kills.

Currently it is at feature freeze. New feature would be possible starting next major release, which is bit overdue now. Hopefully I have some time next 2 months.

loader part? 2.6.1.66? There are known bugs in this version... [EDIT] Oh you are updating to current version. Have you tried the one I provided from last reply?

@PunkBuster What was the version before? Also, can you try this? https://www.dropbox.com/s/ep3hg5pnrd5bgxc/Thor.exe There is a change back then with memory manager during patching. I switched back see if that's the problem. @Fluffle Puff I can't reproduce it. Does it occur in default start button or custom ones?

That's not good... The earliest time I can test this would be tomorrow.

I figure that's my version. If you can prove you are packed the config data, I might able to extract it.

Which version?

Unfortunately no, I didn't make one.

parameter? from where to where? from external login to patcher to exe..like Ai4rie rsu patcher it can pass parameter.. Oh, I will take note of that.