Hi Alberto,
oops, I found a mistake in last changesets. 01LanguageBackport.2.cs.gz and 02LanguageBackport.2.cs.gz include LF as new line mark, please download newer version from; http://www.languagegame.org:8080/ggame/17/ I think there would be another better place to upload these files though... > well, well done Takashi! thanks! > I'll wait for you to implement also an UTF-8 import:) Actually, UTF-8 import has already done. > I'm going to publish the italian translation along with a howto about > how to load it into the image and maybe contribute back. > I think to publish also these changesets, i we don't find a better place > for them. Great! > Another issue i'm facing right now is phrases duplication. Inside the > exported .po i find many duplicated phrases which differ only for the > number of escapes ( \ characters) that they contain. Here is an example: That's interesting. Indeed, this was because gettext importing problem. There is a button named 'r-unused' most right side on LanguageEditor. With this button, you can eliminate bad phrases. But it takes too long time (10-20 min in my computer). Backslash related problem will be fixed in new changeset fixGetTextEscape-tak.1.cs.gz. Anyway, I found another bug. Current gettext support doesn't handle phrases with new lines well. "new line\nphrases" should be denoted as like (at least with po-edit); "" "new line\n" "phrases" This is a time to use mantis... Cheers, - Takashi fixGetTextEscape-tak.1.cs.gz (1K) Download Attachment |
In reply to this post by Takashi Yamamiya
please create or use the one I created on squeaksource!!!!!!
http://www.squeaksource.com/LanguageEditor Stef On 22 mars 06, at 12:39, Takashi Yamamiya wrote: > Hi Jeroen, > > These are another change sets. (I should upload these fixes on some > website > as stéphane said...) > > - 01LanguageBackport.2.cs.gz > almost same as Alberto's patch, I added a preamble and merged > 0473LocaleRefactoring-tak.cs > > - 02LanguageBackport.2.cs.gz > almost same as Alberto's patch, I added a preamble. > > - 03fastLanguageEditor-tak.2.cs.gz > fix NaturalLanguageTranslator >> translated: > - 04GetTextUTF8.1.cs.gz > export gettext file in UTF-8 > > Maybe, your problem will be solved with these change sets. I guess the > problem was untranslated phrases are registered in a dictionary > whenever untranslated phrases are found in String>>translated. > >> ----- Original Message ----- From: "Hilaire Fernandes" >> <[hidden email]>: >>> . Then the big bad point, is about the lost of context when >>> dealing with >>> translation. When using .po file to translate you are totally >>> blind, you >>> don't know the context of the message, then the translation may >>> not be >>> accurate (and well it is not accurate) >> So i was wildly thinking: >> Would it be possible to do something like: >> Activating some kind of 'translation mode' (say by pressing a >> button in the >> language editor). While in this mode, i would be able to select >> any string i >> see and be able to call the translate function on that... (I dont >> know how - >> from the menu or by placing mouse over and press special key?) >> That would >> give me a translate fillInTheBlank. Like now in the editor, but >> called from >> the spot where the actual string is, ie visible in its own >> context.I have no >> idea if this is possible or how, but it seemed like a nice option >> to me. >> That way i could just translate what comes to me first. Because as is >> obvious translating all from the editor or .po is not done easilly / >> quickly. > > That is a nice idea! But current translation mechanism is not > sophisticated enough to do such trick (although, current simplicity is > an advantage in programmer's point of view). And I think we should > start to more easy goal. > > Hilaire Fernandes wrote: >> Now I think that the best way to deal for translation is to split >> the translation effort in department: one person to translate E- >> toys, another one for the menu, another one for a specific package >> (connectors, kedema, ...) > > I agree. It is useful if a translation file can be separated in each > package (PackageInfo). I think we need a keyword like gettext_noop to > mark untranslated phrase, because sometimes #translated is called > indirectly and it does not help to find a word. > > Cheers, > - Takashi > <01LanguageBackport.2.cs.gz> > <02LanguageBackport.2.cs.gz> > <03fastLanguageEditor-tak.2.cs.gz> > <04GetTextUTF8.1.cs.gz> > |
Stéphane,
OK, I published it. I have hesitated to upload as a Monticello package because LanguageEditor depends on a lot of other classes. But anyway, it is better than nothing. - Takashi stéphane ducasse wrote: > please create or use the one I created on squeaksource!!!!!! > > http://www.squeaksource.com/LanguageEditor > > Stef |
Free forum by Nabble | Edit this page |