Saturday, 29 March 2014

RPGMaker Trans v2.04 Released

Another new release. Two things:

  1. Support loading patches with a UTF-8 BOM
  2. Support saving patches with a UTF-8 BOM (via an option)
For those who don't know what a UTF-8 BOM is, it's something Microsoft invented to identify UTF-8 documents. It's almost universally not recommended for technical reasons, but certain text editors on Windows require it. Of course, using a UTF-8 BOM on editors that don't require it could break those editors, hence why it's an option and not enabled-by-default.

Microsoft: Breaking standards since (insert some arbitrary date here)

Sunday, 23 March 2014

RPGMaker Trans v2.03 Released

So apparently, there's a bug on the specific version of cx_freeze that I used. I'm not sure how to trigger this bug, because it didn't show up for me, but if it does show up you can't use the patcher at all.

Anyhow, the workaround seems to be using an older version of cx_freeze, so I've done that, and hence v2.03 is released. If you have v2.02 and it works for you, there's no need to upgrade. Otherwise, grab v2.03 and hopefully it will work.

Saturday, 22 March 2014

RPGMaker Trans v2.02 Released

OK, so quick changelog:
  • Re-enabled the version checker - I have a feeling it could be useful going forward.
  • Fixed error windows popping up. This wasn't caused by what I initially thought, but it's actually a bug in Python. Fortunately, whilst the fix is not in Python yet, there is a patch available.
  • Nonfatal errors: Now there are errors which pop up and don't cause RPGMaker Trans to completely blow up. Specifically, bad patches/games and not being able to copy files.
  • Readme now included. It's a bit brief, so maybe there will be updates later.
Now, onto VX and multilingual support... (crosses fingers that RPGMaker Trans does not explode...)

Monday, 17 March 2014

Current work

For those wondering what I've been doing, I've mainly been working on the known bugs that were listed in the previous post. So, the itemised list:

  • Most of the way through fixing the error messages that briefly pop up
  • Infrastructure for displaying error messages without the entire patcher bombing out is now in place. Practically, this should mean that error messages that can be anticipated leave helpful error messages in a new error log that appears when necessary rather than cryptic debug tracebacks. Debug tracebacks will still be there, but should only be displayed when there are genuine bugs rather than something on the users system being slightly off (e.g. bad patch, problem writing files, etc)
  • Started a readme
Hoping for a release with these features in quite soon, and then onto the more interesting stuff i.e. VX patching.

Monday, 3 March 2014

RPGMaker Trans v2.01 Released

As the title says. Bugs fixed:

  • The odd error that was affecting high core count PCs
  • Memory usage should be lower
  • Read only files crashing the file copier - hopefully 
  • Progress bar no longer gets stuck at 99%
Bugs that remain
  • On closing, you may see some error boxes pop up. I'm working on fixing this, but it's not a critical thing.
  • Some malformed RPGMaker games may cause the translator component to crash (specifically ones which embed NUL characters in strings - not even sure how one does this with RPGMaker)
  • A readme: A number of recent questions seem to suggest a readme is in order.#
  • EDIT: There is an issue with copying thumbs.db files, it seems. Mitigation: don't have the translated folder open whilst patching and you'll probably be OK; if not try again.
So, theoretically, this one should fix the remaining issues with RPGMaker Trans v2. I've said that before, obviously, but hopefully this time it sticks. When someone finds an error (inner cyniscm taking over here), bug reports in the comments please.  Otherwise, apologies for the very bumpy ride from 1.x to 2.01. Hopefully it's over now.

Friday, 28 February 2014

Bugs and stuff

First off, as people apparently cannot google/read previous comments anymore, this:

IF YOU ARE MISSING MSVCR100.DLL INSTALL THE MS VC++ REDISTRIBUTABLE HERE: http://www.microsoft.com/en-us/download/details.aspx?id=26999

Now that's out of the way - for the people with error in Pythons multiprocessing libraries (the one that ends with "TypeError: WriteFile() argument 1 must be integer, not None"): thanks for the system specs. One thing I notices is that you're running Core i7s which have 8 threads, which leads me to think that RPGMaker Trans is going too parallel for Windows to handle. Hence the next version of RPGMaker Trans will have a limit on the number of processes it uses, and hopefully that will be a cure.

As to me testing this out: I don't have a Windows machine with an 8 thread processor in, so that could be tricky. I'll see if I can set something up, but there's a very high chance that'll I'll be working blind on this one.

ETA on RPGMaker Trans 2.01: Sunday/Monday morning 

EDIT: I've managed to do some testing and managed to produce a similar (although not entirely the same) bug, so I'm pretty certain that this is related to Python spawning too many processes on systems with a high thread count. The reason for this is due to some default behavior in Python that I wasn't aware of (specifically multiprocessing.pool does not spin up processes as necessary, but by default starts one process for each hardware thread). The fix is a bit more involved than I thought, but should be done shortly. It wasn't caught before because testing is on a 2 core VM, which doesn't exhibit this problem. 

EDIT2: And it wasn't that... it seems to be a more fundamental bug in Python, but it's fixed anyway.

Wednesday, 26 February 2014

On version 1.99

Seeing as a few people have commented "it worked on 1.99" now...

v1.99 is not coming back. The codebase is unmaintainable, I do not have the time or inclination to maintain it, and I am unable to open source it. It doesn't have the architectural features that I need for new features. Further, it was announced very early on that v1.99 would be killed off and I did ask people to test v2. Unfortunately, very few people did.

If more people had tested v2, then I'm pretty certain that all the bugs would be ironed out by now. That isn't the case. I fully intend to get v2 bug free, but that will take time. Until then, people will have to put up with bugs, because if I put v1.99 back out I won't get the bug reports I need to fix v2. To anyone who doesn't like this policy: sorry, but that's tough on you, I guess. If I could get bug reports for v2 any other way, I would. But no-one wanted to engage with the beta testing report unless they were one of the people who couldn't use v1.99.

So in summary: v1.99 is gone, and it isn't coming back.