Problem with MT 5.0.5pl5 and Sunbird installer files
Iacopo Benesperi
I've downloaded MT 5.0.5pl5 made by Cedric, and I've used it to import
Sunbird CVS files. I've made two products, one for the chrome folder and
one for the installer folder, since MT doesn't support more than 2 level
deep for directory.
I have no problems with chrome folder, MT read all files and I can
translate them painless. I have problems with installer directory: there
are three files in that dir, but if I open chrome view in MT there are
no files listed in the installer directory, so I can't use MT to
translate them.
Did anyone find the same problem before? Or maybe I'm doing something wrong?
bye,
Iacopo
Iacopo Benesperi
> Imade by Cedric
I meant Ricardo, sorry :-P
Ricardo Palomares Martinez
> Iacopo Benesperi ha scritto:
>
> I meant Ricardo, sorry :-P
Now I'll answer you. :-)
As long as I don't finish my testing review of what will be MT 5.1,
you are going to have problems to get MT capture those files. MT 5.0x
versions not only are handicapped with more than two levels depth; it
also are unable to grab files directly under the "root" directory.
So, you're probably using a path ending in ".../en-US/installer" for
that MT product. You could stop at ".../locales/en-US/", but then
you'll probably find yourself with duplicated files.
I think you're using Windows; if you were using a Unix-like filesystem
you could try to do a symlink to add one extra directory depth.
Anyway, I'm just in the middle of MozillaTranslator 5.1 testing, and I
hope to have it ready in about a week. It will be able to cope with
any arbitrary directory level, and (Axel is gonna love this) :-) will
preserve original file licenses, besides allowing to include the
localizer himself as a contributor.
Ricardo.
--
If it's true that we are here to help others,
then what exactly are the OTHERS here for?
Robert Kaiser
> Anyway, I'm just in the middle of MozillaTranslator 5.1 testing, and I
> hope to have it ready in about a week. It will be able to cope with
> any arbitrary directory level, and (Axel is gonna love this) :-) will
> preserve original file licenses, besides allowing to include the
> localizer himself as a contributor.
Wow, cool!
So will you officially take over the MT project now? ;-)
Robert Kaiser
Ricardo Palomares Martinez
Well, my plan is to bring up MT to life again and reach a functional
state, but not add too many features unless they are needed or very
easy to implement.
That's because, as soon as I finish this review, I intend to start
seriously with my brand-new replacement for MT, which would also be
based on Java but will use a real SQL DB (I think of HSQLDB or Derby
for small installations and MySQL for larger ones) and, if I find the
way to do it in a simple way from a end-user perspective, it will
implement web services so other people can write web-based front-ends
(a little science fiction at this moment, though). :-)
I need a brand-new program to present it myself as my final year
project, but as I didn't know almost anything about Java or UML, I
thought that starting by modifying existing code would let me learn
more quickly (and it turned out that I was right).
In no way, however, I plan to discontinue support on MT until I can
trust myself in my new program, :-) so I'm not going to let you down
in the foreseeable future. What brings me to a related topic: move the
sources again to a public repository (hopefully, the MT existing
repository). You, Robert, are listed as member of the MT project at
sourceforge.net, although not as a developer. What that means? Is any
chance that, once we're confident in the new MT version, you can help
me to get understood by Serhiy or other MT developers (there are five
listed including you)?
Ricardo.
PD: mt51.jar weighs 400K, 28K less than mt505pl5.jar. You were right,
we manage to make each version lighter than previous. :-)
Robert Kaiser
> In no way, however, I plan to discontinue support on MT until I can
> trust myself in my new program, :-) so I'm not going to let you down
> in the foreseeable future. What brings me to a related topic: move the
> sources again to a public repository (hopefully, the MT existing
> repository). You, Robert, are listed as member of the MT project at
> sourceforge.net, although not as a developer. What that means?
I didn't even know I was listed there, and I'm not sure what it actually
means...
Robert kaiser
David Fraser
> Robert Kaiser escribió:
>
>> Ricardo Palomares Martinez schrieb:
>>
>>> Anyway, I'm just in the middle of MozillaTranslator 5.1 testing, and I
>>> hope to have it ready in about a week. It will be able to cope with
>>> any arbitrary directory level, and (Axel is gonna love this) :-) will
>>> preserve original file licenses, besides allowing to include the
>>> localizer himself as a contributor.
>>>
>> Wow, cool!
>>
>> So will you officially take over the MT project now? ;-)
>>
>>
>
>
> Well, my plan is to bring up MT to life again and reach a functional
> state, but not add too many features unless they are needed or very
> easy to implement.
>
> That's because, as soon as I finish this review, I intend to start
> seriously with my brand-new replacement for MT, which would also be
> based on Java but will use a real SQL DB (I think of HSQLDB or Derby
> for small installations and MySQL for larger ones) and, if I find the
> way to do it in a simple way from a end-user perspective, it will
> implement web services so other people can write web-based front-ends
> (a little science fiction at this moment, though). :-)
>
> I need a brand-new program to present it myself as my final year
> project, but as I didn't know almost anything about Java or UML, I
> thought that starting by modifying existing code would let me learn
> more quickly (and it turned out that I was right).
Please at least consider looking at the Pootle project. It is currently
adding the ability to deal with multiple different source formats for
translations, so that you could use it for dtd, properties files, etc.
(The support for handling these formats is mature, its just getting it
into the web interface)
Of course, it may not be you thing, but if you find it is, it's always
better to collaborate...
Cheers
David
Iacopo Benesperi
> As long as I don't finish my testing review of what will be MT 5.1,
> you are going to have problems to get MT capture those files. MT 5.0x
> versions not only are handicapped with more than two levels depth; it
> also are unable to grab files directly under the "root" directory.
>
> So, you're probably using a path ending in ".../en-US/installer" for
> that MT product. You could stop at ".../locales/en-US/", but then
> you'll probably find yourself with duplicated files.
Thank you Ricardo!
I was not sure that this was a known-bug, so I reported it. Now I'll try
your tip, maybe I'll get it work.
bye,
Iacopo
Ricardo Palomares Martinez
>>
>> I need a brand-new program to present it myself as my final year
> Wow. Sounds like a lot of work...
> Of course, it may not be you thing, but if you find it is, it's always
> better to collaborate...
As I said, the problem with contributing to an existing project, or
even receiving help with a new one, is that I need to draw a clear
line about what I've done. For a developer, extracting a CVS log with
his contributions would be fine, but tell that to a college
professor... :-) That's one of the reasons why I never intended to
build upon MT (the other one being that its datamodel doesn't
completely fit my view).
I don't know how far I will go with that program before wrapping it up
as the final year project, because it involves a huge documentation
work, but after I get the degree, I intend to advance on it further (a
LOT further, hopefully). :-)
Ricardo.
Robert Kaiser
> I don't know how far I will go with that program before wrapping it up
> as the final year project, because it involves a huge documentation
> work, but after I get the degree, I intend to advance on it further (a
> LOT further, hopefully). :-)
BTW, does it need to be in Java?
I'd think that a XULRunner app would fit the Mozilla model even better :)
Robert Kaiser
Ricardo Palomares Martinez
I've spent a year trying to get myself used to Java, don't even
suggest I should change the programming language. ;-)
Seriously, I always thought of using a database, so writing in XUL was
not possible at that time. Besides, I'd really like to learn something
that I can use when I'm not contributing Mozilla. :-)
David Fraser
> Robert Kaiser escribió:
>
>> Ricardo Palomares Martinez schrieb:
>>
>>> I don't know how far I will go with that program before wrapping it up
>>> as the final year project, because it involves a huge documentation
>>> work, but after I get the degree, I intend to advance on it further (a
>>> LOT further, hopefully). :-)
>>>
>> BTW, does it need to be in Java?
>> I'd think that a XULRunner app would fit the Mozilla model even better :)
>>
>
>
> I've spent a year trying to get myself used to Java, don't even
> suggest I should change the programming language. ;-)
*best* reason to change it ;-)
David
Robert Kaiser
>> I'd think that a XULRunner app would fit the Mozilla model even better :)
>
> I've spent a year trying to get myself used to Java, don't even
> suggest I should change the programming language. ;-)
Programming language?
/me ducks ;-)
> Seriously, I always thought of using a database, so writing in XUL was
> not possible at that time. Besides, I'd really like to learn something
> that I can use when I'm not contributing Mozilla. :-)
Well, using databases is easy with XULRunner (or should be, we have
SQLite support) - and programming knowledge is actually quite universal,
the structure is quite similar between languages, basically only the
syntaxes differ.
That said, good C++ and XML (think XUL) knowledge seems to be even more
important nowadays than good Java knowledge...
Robert Kaiser
Ricardo Palomares Martinez
> Well, using databases is easy with XULRunner (or should be, we have
> SQLite support) - and programming knowledge is actually quite universal,
> the structure is quite similar between languages, basically only the
> syntaxes differ.
I'd rather say the exact opposite. the syntaxs are similar with minor
differences, and anyway, you only need two weeks or so to get used to
the fact that VB uses select for what Java or C is switch, and
everything else...
However, the real difference I see and where I've found more problems
to make myself not feeling a complete idiot is the underlying OOP
philosophy and the API. In my case, the OOP philosophy problem was
simply that I hadn't used before a nearly fully OOP language; for
C/C++ programmers, the problems would come from multiple inheritance,
pointer arithmetic, etc.
Regarding the API components, there are some that are more or less
standardized (DOM and SAX come to my mind), but others vary from
language to language, and with so huge APIs nowadays, I don't think it
can be easy changing from one to another without losing productivity.
Anyway, let's face it, the language was not the one to blame in this
case, but the individual. :-)
> That said, good C++ and XML (think XUL) knowledge seems to be even more
> important nowadays than good Java knowledge...
What I like about Java is that it keeps being more portable than any
other option, and is advancing in the open source direction.
Restricting to XML/XUL/JS is yet too limiting as long as Gecko doesn't
reach a leading position in browsers quota (right now, you would be
forcing to 3 out of 4 people to install a Mozilla product to use a
XULRunner-based program).
.NET is something I don't really trust on; on one hand, it is a MS
product, and we have seen enough of MS tactics. Besides, people
working in Mono lacks API experience when you place them in a pure
.NET environment, and .NET people uses that APIs, thus being
handicapped to write programs that run outside a Windows environment.