|
|
|
Home » U++ Library support » U++ MT-multithreading and servers » Witz internationalization
| Witz internationalization [message #39559] |
Fri, 29 March 2013 19:13  |
Zbych
Messages: 327
Registered: July 2009
|
Senior Member |
|
|
Hi,
I was playing little bit with translating texts in witz scripts and it appears that it is possible to use t_("xxx") syntax like in *.cpp files (e.x. $t_("some text")).
TheIDE even scans *.witz scripts for t_(""). So far so good. But I have some questions:
- GetLngString uses cache to speed up translation. Is it possible that different texts from *.witz can have the same location in memory? Or maybe the same text can have different address each time and this will make cache grow after each translation. Maybe it is better to use non-cached version (GetLngString(GetCurrentLanguage(), (String)arg[0]))
- is it possible to detect user agent language and automatically show him proper translation?
Modified Skylark11 example:
main.cpp:
#include <Skylark/Skylark.h>
#include <plugin/sqlite3/Sqlite3.h>
#define TFILE <Skylark11/lang.t>
#include <Core/t.h>
using namespace Upp;
SKYLARK(HomePage, "")
{
http("VAR", t_("Aborted by user."))
.RenderResult("Skylark11/index");
}
SKYLARK(SetLanguage, "setlanguage/*")
{
int lang = LNGFromText(http[0]);
if(lang)
http.SetLanguage(lang);
http.Redirect(HomePage);
}
struct MyApp : SkylarkApp {
MyApp() {
root = "myapp";
threads = 1; // Sqlite3 does not like threads...
#ifdef _DEBUG
prefork = 0;
use_caching = false;
#endif
}
};
Value translate(const Vector<Value>& arg, const Renderer *)
{
if(arg.GetCount() == 1 && IsString(arg[0]))
return GetLngString((String)arg[0]);
// return GetLngString(GetCurrentLanguage(), (String)arg[0]);
return arg[0];
}
INITBLOCK {
Compiler::Register("t_", translate);
};
CONSOLE_APP_MAIN
{
#ifdef _DEBUG
StdLogSetup(LOG_FILE|LOG_COUT);
Ini::skylark_log = true;
#endif
MyApp().Run();
}
index.witz:
<html>
<body>
<a href=$SetLanguage("en-us")>EN</a>
<a href=$SetLanguage("cs-cz")>CZ</a>
<a href=$SetLanguage("pl-pl")>PL</a>
<br>
$t_("Current language"): $.language<br>
$t_("English version")<br>
$t_("Variable value"): $VAR
</body>
</html>
lang.t:
#ifdef _MSC_VER
#pragma setlocale("C")
#endif
// lang.witz
T_("Current language")
plPL("Bieżący język")
// index.witz
T_("English version")
plPL("Wersja polska")
T_("Variable value")
plPL("Wartość zmiennej")
// main.cpp
T_("Aborted by user.")
plPL("Przerwano na życzenie użytkownika")
[Updated on: Fri, 29 March 2013 19:15] Report message to a moderator |
|
|
|
| Re: Witz internationalization [message #39562 is a reply to message #39559] |
Sat, 30 March 2013 10:16  |
|
|
Hi Zbych,
I wrote almost exactly the same function a while back 
| Quote: |
Value Translate(const Vector<Value>& arg, const Renderer *) {
if(arg.GetCount() != 1 || !IsString(arg[0]))
return String();
return GetLngString(String(arg[0]));
}
INITBLOCK {
Compiler::Register("t_", Translate);
}
|
I had no problem with this implementation, but I haven't use it too extensively, so I probably wouldn't notice cache problems. Also, it is only useful for shorter strings, anything bigger is better to be placed in separate files (*.xy-xy.witz).
User language can be detected by parsing the Accept-Language HTTP header. However, it is always good idea to let the user override the setting manually (e.g. for cases when they use someone else's computer with different language settings).
Best regards,
Honza
|
|
|
|
|
|
| Re: Witz internationalization [message #39569 is a reply to message #39562] |
Sat, 30 March 2013 12:22 |
Zbych
Messages: 327
Registered: July 2009
|
Senior Member |
|
|
| mirek wrote on Sat, 30 March 2013 11:24 |
- there definitely IS a problem with GetLngString and cache. t_ macro and GetLngString are only defined when being used with string literal. You are using them with String...
|
As you can see insane people like me are always trying push your code to the limits (and even beyond). Maybe there should be overloaded version of GetLngString that takes String and uses hash instead of address to cache translations? I can make it.
| mirek wrote on Sat, 30 March 2013 11:24 |
- to get HTTP header fields, use Http::GetHeader
|
| dolik.rce wrote on Sat, 30 March 2013 10:16 |
User language can be detected by parsing the Accept-Language HTTP header.
|
Thank you both. I will try it.
|
|
|
|
Goto Forum:
Current Time: Thu Dec 26 13:26:33 CET 2024
Total time taken to generate the page: 0.04713 seconds
|
|
|
Members
Pages
Help
Register
Login
Home
