more /root/Features/Non-ASCII characters.thread
2010-07-04 17:53:57IsmAvatar
I'll look into it. It may be a problem with the font. LGM uses Monospace. Perhaps try another application that uses Monospace, like a text editor (you may need to tell it to use a specific font), and test the characters. If the problem is reproduced there, I would recommend obtaining a replacement Monospace font that does support your accented characters.

Also, it may interest you that many of LGM's strings are externalized, meaning that they can be replaced with the French equivalents, so you can have a (mostly) french LGM without needing to recompile LGM.
2010-07-04 19:12:36IsmAvatar
Unable to reproduce on Windows XP. Will try Ubuntu next.
2010-07-04 19:21:55IsmAvatar
Monospace is a class of font. The actual font implementation varies from platform to platform, but Windows will probably choose "Courier" or "Courier New" as your monospace. Try that font instead.

Also, could you please tell me what operating system you're using, your Java version, and the LGM revision (r###, as visible from the Help>About box)?
2010-07-07 16:59:54IsmAvatar
I've reproduced this on Linux, Fedora 12 and LGM r444 (latest from the repo). I suspect this is a problem with the save method losing the characters.

Thank you for your bug finding. We'll be sure to try and correct this for the next version. I'll keep this thread updated with progress with a bugfix.
2010-07-08 02:52:31IsmAvatar
	public void writeStr(String str) throws IOException
		write(str.getBytes("ascii")); //$NON-NLS-1$

Here, String.getBytes("ascii") returns the '?' character for our special characters. This piece of code has been around ever since the inception of our SVN repository, believe it or not, as can be seen in revision 4.

Obviously, the workaround would be to use something other than "ascii" as the character encoding. The question is what to use. I'm afraid to use the default encoding for fear that it may decide to produce 2-byte strings on some machines.
2010-07-08 03:36:54IsmAvatar
I've solved this by replacing "ascii" with "ISO-8859-1", also known as ISO-LATIN-1, an 8-byte character encoding scheme that recognizes special characters like "é", "è" or "à" with the same byte representations as GM does. I believe GM may also be using Iso-Latin-1.
Fortunately, "ISO-8859-1" is a java standard charset, so I don't need to worry about some machines not supporting it.

I've corrected this in r446 of the versioning branch.
2010-07-08 19:55:44IsmAvatar
No. You'll either have to wait, or build it yourself

Eventually, we'll probably build it and include it with the Enigma project, since they have been interested in keeping with the bleeding edge LGM. However, you won't find a download for it on our website until the versioning branch is merged with the trunk again.
2010-07-21 20:45:15IsmAvatar
A newer revision of LGM versioning, r447 (corrects some GM8 file problems) was built and included with the latest revision of Enigma, r290. It can be downloaded from their SVN. Here's a hotlink for you:


(Note, despite the "b4", it is not actually beta - it's an unofficial build for Enigma)