JBand progress log.

[PaulBlay]

Visual Studios half-hearted support of Unicode got me again today. Compiler errors are generated if you don't have the BOM (Byte Order Marker) in UTF-8 files and you have certain characters in the text.

Of course lots of programs (particularly in Linux) will complain if UTF-8 files have a BOM. So I'm stuffed either way.

[EDIT] Can't load savefiles, but that's no big surprise. It probably isn't saving them right in the first place.

[EDIT]x2 Yay! Savefiles load. (fingers crossed).

Other new bugs have arrived
Walking into walls doesn't always give the default message.
Crash when running in corridors (my teachers always said you shouldn't do that).
Auto identify just works too well.

[PaulBlay]

Well, I've just got a character to level 5 without the game crashing - I think that deserves a post.

[PaulBlay]

Apparently JBand only likes warriors. It doesn't always crash when a non-warrior is selected, but the arguments passed to character generation are odd (female mage = male warrior with all stats at 18/1024 )

[Nolendil]

Nice improvements!

[PaulBlay]

Nice improvements!

It's, how you say, "a learning experience".

Latest developments on the "Warriors only" bug. Apparently Rogues don't have sex. Who'd a thought it? :P

[EDIT] Fixed that one now.

[PaulBlay]

Hey, I now have Japanese displaying (almost) correctly!

(Actually each character is split down the middle with a small gap. Probably a font issue?)

[EDIT] I added code to select a font and now it isn't working again. :-(

[Nolendil]

Great!
Isn't the gap due to some problem with tile width being one pixel wider than the font width?

[PaulBlay]

Great!
Isn't the gap due to some problem with tile width being one pixel wider than the font width?

Possibly, but I've since fixed that problem and created a new problem.

I'm currently seeking help from a C programming forum.

Oh well, I think I'm generally moving forwards anyway.

[EDIT] Well it's now (approximately) working. The screen is just a bit flickery.

[Pete Mack]

Very, very impressive. How are you localizing the messages? Is it portable to another language than Japanese?

[PaulBlay]

Very, very impressive. How are you localizing the messages? Is it portable to another language than Japanese?

Haha, well sort of. I may well replace the 'system' (if you could call it that) that I've got at the moment later.

Basically it's set up for two languages and you just switch between them with <ctrl>+L. (It also saves the language used last in the .INI and starts with that as default next time). At the moment it only supports complete sentences/messages. Support for all the "You swing your %s at the %s" type messages has yet to be implemented (although I've got a good idea about how I intend to go about it).

The program is a Unicode compiled Windows program (portability will be a pain) so there is relatively little special handling that would need to be added for a different language.

I introduced a very hacky work-around to get Japanese to display correctly - each Japanese character takes up two spaces so I add a 'fake space' after each Japanese character (via a function called 'spaceme()' and make sure that the Term_queue_chars() function does not write anything for those spaces. This has the advantage that length calculations on the strings come to the 'right' answers. Unfortunately I had to do a very ugly hack to get the right half of each character to display correctly.

[Nick]

I introduced a very hacky work-around to get Japanese to display correctly - each Japanese character takes up two spaces so I add a 'fake space' after each Japanese character (via a function called 'spaceme()' and make sure that the Term_queue_chars() function does not write anything for those spaces. This has the advantage that length calculations on the strings come to the 'right' answers. Unfortunately I had to do a very ugly hack to get the right half of each character to display correctly.

This sounds like the ugly hack the produces bigtile mode - maybe there's a generalisation possible here.

[PaulBlay]

This sounds like the ugly hack the produces bigtile mode - maybe there's a generalisation possible here.

XAngband uses a modification of the bigtile mode code. Frankly I had trouble following it which is why I thought up my hack that (almost) worked.

The only problem (so far) is that I have to force a repaint after displaying Japanese text so the window flickers a bit.

[PaulBlay]

OK, progress reports again.

I've started translating in earnest now. Estimated completion ratio ... 0.5%

Put up a 'help wanted' for translators, but not overly hopeful of getting a response.

[artanis]

Looks like you're making good progress. I haven't had time to do anything lately, but I'll see what I can do to help again in a couple days.

[PaulBlay]

News page now displays (mostly) correctly. It actually looks better than this, but it seems that SourceForge.JP shrunk it a bit.

[EDIT] Woot! 100th SVN commit!