29.01.2011 3:29, Marcelo B de Paula ?????:

Please check if LazDE translation works when you rename its .pb.po file to .pt_BR.po.

If it works, we can solve all these problems by renaming all .pb.po files to .pt_BR.po.

Hi list,

I am having some trouble using my native language with Lazarus applications (Brazilian Portuguese).
IDE itself works fine with its translated language files, but the tools included with it (lazde for example) and programs using DefaulTranslator, do not recognize the correct language files.

The portuguese language files have an extension ".pb.po". Unfortunatelly, function LCLGetLanguageIDS returns "pt_BR" or "pt" as a locale identification. Its correct in OS point of view, but make translations routines unusable since they can?t find the correct files.

I really don?t know why these files was standardized with "pb" instead of "pt".

I intend to submit a patch to modify LCLProc.pas to change the "pt" to "pb", when detected, to solve this problem once and for all.

Did anyone else are having similar problem ?

Any suggestions on this matter are welcomed.

Marcelo.

Marcelo B de Paula wrote

>> I really don?t know why these files was standardized with "pb" instead >> of "pt".

the answer seems to be here
http://wiki.lazarus.freepascal.org/Translations_/_i18n_/_localizations_for_programs#Translating

Out of curiosity, if you change it in lcl, what happens to people in Portugal, who might actually have pt? or pt_PT ?
And it wouldn't just be brazil, which other of the list below may cause trouble? (I do NOT know the code behind it, did not look at the LCL part in question)
http://www.termwiki.com/TWSpecial:ISO_Language_Code_Comparison -------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20110129/aa9f1817/attachment.html>

Yes, it works. But this not solve our problem. If i do the same with lazaruside.pb.po it disrupts the IDE translations. IDE itself maps the ".pb.po" files.
See attached image.
I still think it would be better to put a line of code in lclproc.pas.

Marcelo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20110128/607ce1e7/attachment-0001.html> -------------- next part --------------
A non-text attachment was scrubbed...
Name: options.PNG
Type: image/png
Size: 62240 bytes
Desc: not available
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20110128/607ce1e7/attachment-0001.png>

I found this on lazarus wiki:

http://wiki.lazarus.freepascal.org/Translations_/_i18n_/_localizations_for_programs#Translating

Thats why these files were standardized this way.

Oh man, this will be a big headache. Other portuguese language countries will have variations on locale identification. So, IMHO, we will need to really change the lclproc.pas

Regards,

Marcelo.

Why a distinction between Brasilian and other Portuguese here?

How are en_US, en_GB etc. handled?

DoDi

?

Hans-Peter Diettrich hat am 29. Januar 2011 um 07:05 geschrieben:

The algorithm first searches for the "en_US", then "en", finally the default translation.
Because the main language id 'pt' differ from 'pb' a different po file is used. Maybe pb could be used as fallback for pt and vice versus in translations.pas. ?
Mattias
?
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lazarus.freepascal.org/pipermail/lazarus/attachments/20110129/dc1cf5d3/attachment-0001.html>

As far as i know, reading Lazarus wiki, these files were standardized as two letters for country ("xx.po"). It seems that the idea was not to have language variations. Today, we already have some variations implemented, like "af_ZA" and "zn_CN"

So, what to follow. What?s stated on wiki or implementing language variations ?

I see the following possibilities:

- Review the 2 letters standard,
- Release implementation of language variations (IDETranslations will be updated for every new language variation),
- Normalize locale returns in LCLProc to map the correct 2 letters. Currently the code for Linux returns only the first 2 letters of Locale (ex. "pt_BR" returns "pt")
- Any portuguese language variations here could be solved only mapping "pt" to "pb". Or we can rename the files from "pb.po" to "pt.po"
- Other languages have to follow the same rule. So, "en_US" and "en_GB" will be simple "en", thus mapping to "en.po" files.

Regards,

Marcelo

29.01.2011 5:27, Marcelo B de Paula ?????:

Yes, it works. But this not solve our problem. If i do the same with lazaruside.pb.po it disrupts the IDE translations. IDE itself maps the ".pb.po" files.

Changed. I have renamed .pb.po files to .pt_BR.po. Please test.

See attached image. I still think it would be better to put a line of code in lclproc.pas.

No.

Brazilian portuguese is very different from "portuguese" portuguese. It's not US-GB difference, mostly about slight spelling variations like colour and color. In Brazil FILE and SCREEN are "arquivo" and "tela"; in Portugal they are "ficheiro" and "ecr?".

2011/1/29 Hans-Peter Diettrich

Maxim Ganetsky wrote

Changed. I have renamed .pb.po files to .pt_BR.po. Please test.

Will this change cause that pt_PT users end up with the english translations by default? If there is no pt_PT translation, then it should default to pt_BR. Maybe you should have renamed to pt.po instead?

Brazilian portuguese is very different from "portuguese" portuguese. It's not US-GB difference, mostly about slight spelling variations like colour and color. In Brazil FILE and SCREEN are "arquivo" and "tela"; in Portugal they are "ficheiro" and "ecr?".

That's from old times when people didn't care about keeping the language united. Now we have the spelling reform and in wikipedia for example there is no pt_PT and pt_BR. The spelling reform also says that in the future we should have a common dictionary of preferred technical terms.

That's why I asked for the reason for the special handling of pt_BR, with the *unsystematic* shortcut "pb".

Maybe pb could be used as fallback for pt and vice versus in translations.pas.

IMO "pb" simply should be removed, and the Brasilian users can vote for their own pt_BR translation, if this ever makes sense.

DoDi

Hans-Peter Diettrich wrote

That's why I asked for the reason for the special handling of pt_BR, with the *unsystematic* shortcut "pb".

This "pb" shortcut is ridiculous.

IMO "pb" simply should be removed, and the Brasilian users can vote for their own pt_BR translation, if this ever makes sense.

Why removed? Then there will be no portuguese my eyes are tricking me, I don't a pt.po translation file.

If I am correct and there is no pt_PT translation, then it should be moved to pt.po instead.

I vote for Option 1

Please, revert this submit. This will not work.
By renaming the files you disrupted the IDE translations.

Before taking any actions, we must decide the best thing to do.

Renaming the files, as you did, will frustrate our collegues from Portugal, since their locale will not find the correct
files and they will end up with the english version.

Did you read my previous post ?

Option 01: Rename the files to "*.pt.po" instead of "pt_BR.po" Change IDETranslations.pas ( Portuguese ) "pb -> pt"

Option 02: Make a statement on Lazarus wiki to permit language variations. Brazilizan Portuguese files will be renamed to ".pt_BR.po" Portugal Portuguese should provide their own translations files (".pt_PT" or whatever.)
Change IDETranslations to identify Portuguese Brazilian, Portuguese Portugal, .....
Modify LCLProc (Linux) to return the entire Locale variable, instead of Copy(1,2)..

Option 03: Do not touch any files or any code, except LCLProc to return locale which starts with "pt" to "pb".
Do not change what is stated on wiki.
Do not permit language variations. Mantain the already implemented rule. So Portugal Portuguese will have to use Brazilian Portuguese language files.

Please, think on this with care. This is not a matter to rename files only. This is about to change a rule.
Will be great if other Lazarus Team take a look on this, before any actions.

Regards,

MarceloB. Paula

Maxim Ganetsky wrote

Yes, it works. But this not solve our problem. If i do the same with lazaruside.pb.po it disrupts the IDE translations. IDE itself maps the ".pb.po" files.

Changed. I have renamed .pb.po files to .pt_BR.po. Please test.

See attached image. I still think it would be better to put a line of code in lclproc.pas.

No.

After thinking and writing the options,

I will vote to Option 01 too.

Marcelo B. Paula

Not true! While the spelling is unified, the words that we use for each case are completely different.

Examples:

English -> Portuguese (European) -> Portuguese (Brazilian)

Application -> Aplica??o -> Aplicativo
Desktop publishing -> Edi??o Electr?nica -> Editora??o eletr?nica Save -> Gravar -> Salvar
Spreadsheet -> Folha de c?lculo -> Planilha eletr?nica Database > Base de dados -> Banco de dados
Mouse -> Rato -> Mouse
etc...

And wikipedia is very lame in having only one "Portuguese" version, many times it becomes very Brazilian centered...

Paulo Costa

Paulo Costa wrote

Not true! While the spelling is unified, the words that we use for each case are completely different.

http://pt.wikipedia.org/wiki/Acordo_Ortogr?fico_de_1990#Vocabul.C3.A1rio_comum

29.01.2011 20:33, Marcelo B de Paula ?????:

Please, revert this submit. This will not work. By renaming the files you disrupted the IDE translations.

Your translation is Brazilian Portuguese, so my commit will work OK. When there will be proper Portuguese translation it would be added as appropriate.

Hello all,

The Language is the same but the each side has many prefered local words much specialy in computer related translations:

brazilians tend to adapt original expressions for their use, like:

Delete -> Deletar
Video Monitor -> Monitor de video
Format -> Formatar
Mouse -> Mouse
etc...

while portuguese prefer create their own portuguese words

IMO two different translations are the best option, unfortunately my knoledge do not permits me to go trough technical issues yet

30.01.2011 2:41, Marcelo B de Paula ?????:

Please, take a look at /Lazarus/ide/IDETranslations.pas, function GetLazarusLanguageLocalizedNames, line 110 and tell me how it will work.

Please update your Lazarus SVN copy. Line 110 of this file hase exactly pt_BR value since r29245.

I see no pt_BR or pt_PT or other portuguese language reference there. Simple 'pb'. How this function and associate functions will "guess" that the files to loaded are the ".pt_BR.po" ones.

30.01.2011 3:07, Marcelo B de Paula ?????:

So, our folks from Portugal and other portuguese language countries are on their own now !!

Yes, but keep in mind that our Portuguese fellows can base their efforts on your work.

Please, take a look at /Lazarus/ide/IDETranslations.pas, function GetLazarusLanguageLocalizedNames, line 110 and
tell me how it will work.
I see no pt_BR or pt_PT or other portuguese language reference there. Simple 'pb'.
How this function and associate functions will "guess" that the files to loaded are the ".pt_BR.po" ones.

Regards,

Marcelo B Paula.

Maxim Ganetsky wrote

Please, revert this submit. This will not work. By renaming the files you disrupted the IDE translations.

Your translation is Brazilian Portuguese, so my commit will work OK. When there will be proper Portuguese translation it would be added as appropriate.

So, our folks from Portugal and other portuguese language countries are on their own now !!

Sorry guys, i did what i could. You will have some work to do now....

Thanks you all,

Marcelo B. Paula.

Maxim Ganetsky wrote

Please, take a look at /Lazarus/ide/IDETranslations.pas, function GetLazarusLanguageLocalizedNames, line 110 and tell me how it will work.

Please update your Lazarus SVN copy. Line 110 of this file hase exactly pt_BR value since r29245.

I see no pt_BR or pt_PT or other portuguese language reference there. Simple 'pb'. How this function and associate functions will "guess" that the files to loaded are the ".pt_BR.po" ones.

No coments on solution 1?

On Jan 29, 2011 11:21 PM, "Maxim Ganetsky" wrote:

29.01.2011 20:33, Marcelo B de Paula ?????:

Please, revert this submit. This will not work. By renaming the files you disrupted the IDE t...

Your translation is Brazilian Portuguese, so my commit will work OK. When there will be proper Portuguese translation it would be added as appropriate.

Your argument is that regional differences are so important that people in Portugal, Angola, green cape and east timor are better of with English? Thats completely against the spirit of the reform, which was approved in the parlament of all involved countries.

On Jan 30, 2011 12:15 AM, "Maxim Ganetsky" wrote:

30.01.2011 3:07, Marcelo B de Paula ?????:

So, our folks from Portugal and other portuguese language countries are on their own now !!

...
Yes, but keep in mind that our Portuguese fellows can base their efforts on your work.

Alexsander Rosa wrote

Brazilian portuguese is very different from "portuguese" portuguese. It's not US-GB difference, mostly about slight spelling variations like colour and color. In Brazil FILE and SCREEN are "arquivo" and "tela"; in Portugal they are "ficheiro" and "ecr?".

I fully agree, because there is a big difference between en_US, en_GB, en_ZA, en_AU etc.. too. You can not simply ignore those variations and only have one en.po file.

Translation should go from most specific to least specific (as a fallback).

Regards,
Graeme

30.01.2011 9:34, Felipe Monteiro de Carvalho ?????:

No coments on solution 1?

1. From Wikipedia I understood that there are substantial semantical differences between Brazilian and Portugal Portuguese. At least half of the people participating in this thread seem to share this opinion.

2. If people in Portugal want to select Brazilian Portuguese translation, they can always do it. Yes, locale autodetect will not work for them, but it didn't work before too, so the current situation is not worse than before.

Maxim Ganetsky wrote

1. From Wikipedia I understood that there are substantial semantical differences between Brazilian and Portugal Portuguese. At least half of the people participating in this thread seem to share this opinion.

That's ridiculous, a list of different prefered words (note that both versions are valid portuguese) doesn't make brazilian portuguese ununderstandable for people from other portuguese-speaking countries. Plus we aren't even talking about making a unified translation, although this would be possible. If there is enough space one can even add both variants, like: Mouse / Rato

Did you notice, by the way, that there is no brazilian or portugal portuguese wikipedia? Just portuguese wikipedia? Also Mac OS X does not have 2 separate translations.

Now, back to our topic, what is really at stake here is if the default language for all portuguese speaking countries should be the current translation or not. What is the default language made for? Obviously that for people that don't speak english or have a hard time reading it. For those you really intend to argue that english will be just as hard to read like brazilian portuguese? This afronts logic.

Anyway, we can always agree to disagree. And if there is no agreement, I propose a voting to solve this democratically. It can be added to the forum like we have already done previously. I propose the wording:

For users from Portugal, Angola, Mozambique, Green Cape and East Timor and other portuguese speaking countries, the default translation should be:
a> English
b> Brazilian portuguese

Then the vote can decide the issue instead of having a unilateral decision.

2. If people in Portugal want to select Brazilian Portuguese translation, they can always do it. Yes, locale autodetect will not work for them, but it didn't work before too, so the current situation is not worse than before.

<sarcasm>Yes, sure, a user from Portugal which doesn't speak english would have a much harder time trying to read brazilian portuguese then it will have trying to read english</sarcasm>

maybe add a .pt copy of .pt_BR so that other portuguese will have at least that translation

Marc

You are right. The semantical differences are huge specially in areas where there are many "new" words. Electronics/computers/software are some of those areas.

"ununderstandable" is a strong word, but it makes us uncomfortable and confused, that is a fact.
Most software is localized differently for Portugal and for Brazil.

Did you notice, by the way, that there is no brazilian or portugal portuguese wikipedia? Just portuguese wikipedia? Also Mac OS X does not have 2 separate translations.

And just like I said before that is a big wikipedia mistake. From your point of view it is good: it makes the Portuguese wikipedia Brazilian centric, but for us it is a loss.

I'm not very familiar with Mac OS X but I think it has separate translations:

http://macosxportugues.blogspot.com/2010/03/portugues-ou-portugues-pt.html

It depends, here in Portugal it is commonly thought that if you want to be a developer you must be able to use English tools and documentation.

For users from Portugal, Angola, Mozambique, Green Cape and East Timor and other portuguese speaking countries, the default translation should be: a> English

English

b> Brazilian portuguese Then the vote can decide the issue instead of having a unilateral decision.

Of course the best solution is to have a Portuguese (European) translation. I'm volunteering to do that.
I already searched (very briefly) for instructions on what is necessary but only found:
http://wiki.lazarus.freepascal.org/Localization
where the topic "Localization of Lazarus" isn't yet a link :(

Paulo Costa

30.01.2011 19:20, Paulo Costa ?????:

I can make a copy of Brazilian Portuguese files (.pt_BR.po) to .pt.po, then you will be able to adapt them just by editing .po files.

30.01.2011 19:07, Marc Weustink ?????:

I made the copy. Paulo Costa volunteered to maintain it.

Ok, good, the issue seams solved :)

On Jan 30, 2011 7:19 PM, "Maxim Ganetsky" wrote:

30.01.2011 19:07, Marc Weustink ?????:

On 30-1-2011 7:53, Felipe Monteiro de Carvalho wrote: > > Your argument is that regional diff...

I made the copy. Paulo Costa volunteered to maintain it.

I dunno about the implementation, but shouldn't it be possible to use a common .pt file and leave only the different translations in .pt_PT and .pt_BR?

DoDi

Did you notice, by the way, that there is no brazilian or portugal portuguese wikipedia? Just portuguese wikipedia? Also Mac OS X does not have 2 separate translations.

Documentation has one immanent problem: how to share common texts, and only exchange specific parts?

As long as automatic translation doesn't work reliably, all translated wiki entries will differ over time, with different updates in every language. In the PT/BR case it could help to add macros for the different terms, that are expanded when a page is rendered. It's up to you (the Portuguese speakers) to ask the wiki developers for an according feature. Or the Lazarus developers...

Why fix such general decisions, instead of leaving the individual preferences settings to the users?

DoDi