|
|||
|
.NET, Delphi, ... - Sisulizer Localization Tool Support > Technical Support (You need to be registered at the forum to write) > Bugs and Quirks in Sisulizer > Strings are trancated or mysteriously filtered
|
| Moderated by: Renate.Reinartz, Markus.Kreisel, Jaakko.Salmenius, Ilkka.Salmenius | Page: 1 2 |
|
|||||||||||||||
| Strings are trancated or mysteriously filtered - Bugs and Quirks in Sisulizer - Technical Support (You need to be registered at the forum to write) - .NET, Delphi, ... - Sisulizer Localization Tool Support | |||||||||||||||||
| Author | Post | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|||||||||||||||||
|
Janusz Grzybek Super Moderator
|
Could you send on support (at) sisulizer.fi or attach to forum post your project file? BTW I have reproduced on other PC and Windows system (Win 2000) issue with “4” string, and it could be other bug in SL or character set. Janusz Attachment: screen.png (Downloaded 128 times)
____________________ http://www.sisulizer.com - Three simple steps to localize |
||||||||||||||||
| |||||||||||||||||
| |||||||||||||||||
|
starr Member
|
It's strange as a friend of mine is also able to reproduce the bug on an English version of Windows though according to your result it is related to the system. I think it better for other person of your staff to try to reproduce it on another machine to confirm the bug. Thanks.
|
||||||||||||||||
| |||||||||||||||||
|
|||||||||||||||||
|
Janusz Grzybek Super Moderator
|
Thank you for information. We have now your detailed description and screenshots. Maybe Jaakko will find some solution for testing it. It's really strange because I tested it on two different PC and OS. Janusz
____________________ http://www.sisulizer.com - Three simple steps to localize |
||||||||||||||||
| |||||||||||||||||
| |||||||||||||||||
|
Jaakko.Salmenius Administrator
|
The reson for some character showing in rectangles was an font issue. Current Sisulizer build (270) uses Arial as default font in the sheet. This causes this rectagle feature. Starting from next build the default font of sheet is Tahoma. It has better support for all characters and it can use font cascading that Arial could not. Font cascading was used with Arial only of charset was specified (e.g. in the Chinese column) but did not work when original language was English but strings actually contained Chinese data. You can get current build showing the strings correctly if you change the font of the sheet to Tahoma. 1) Right click the sheet 2) Choose Sheet properties 3) Set font from Arial to Tahoma. Jaakko
____________________ http://www.sisulizer.com - Three simple steps to localize |
||||||||||||||||
| |||||||||||||||||
|
|||||||||||||||||
|
starr Member
|
My friend with English system made a test and it seemed that it had something to do with the system's default non-Unicode program codepage fallback. If it is set to Simplified Chinese (like the localized Simplified Chinese version of Windows), it is filtered while if set to English it is ok. And here is the project's content:
|
||||||||||||||||
| |||||||||||||||||
| |||||||||||||||||
|
banxian Member
|
Western European as CP_ACP will keep all chars between string/widestring hidden cast. once your code working with literal or other PChar/LStr/ShortString/SmallString, hidden cast between string/widestring must be triggered. I don't know if some of you using TNTUnicode control pack or some other componets in your sisiterlizer, I means I used an API based KOL library for my applications, some years ago I meet more problem like this when Unicode control support been added, and I have to keep my unofficial patch for it. I made some test with sisu in Virtual machines with Windows Server 2003 Std ENU/ Windows 2000 Professional, and be munitioned the "Language for non-Unicode programs" setting get the original column in sisulizer filled by different chars. in multibyte world, if we set this option to "English (US)", okay now the hidden cast can't fAke my Unicode strings but, most of ANSI programs will working like an modern arts which full of USASCIIExt. PS: Multilizer working fine on my actual text source (in Key File mode). this slp xml files still lost decectfileformat/fileformat="0" field in text source settings. I can't confirm there some typo in sisu codes. I have read some reply and found Ini mode works fine on my testcase, but my actualize program will output additional columns for debug proper, so I think the Ini workaround may work not so good, they can't ignore these columns like text source processor. sorry for my trouble post, Delphi 2009 will released tomorrow, all chars and string relative API now fixed to Wide Version like in past they FIXED to narrow ANSI calls. my API based applications can't compile with it at all, I lost my best comelier if I can't modify all windows.pas into an {$IFDEF ANSI} switch, it's more trouble than this post in my mind Oh no.
|
||||||||||||||||
| |||||||||||||||||
|
|||||||||||||||||
|
Jaakko.Salmenius Administrator
|
This boxing in Sisulizer sheet and when viewing SLP on text editor is caused when the current font does not have support for a characters. In that case Windows draws a box. In Sisuluizer sheet Sisulizer used Arial font that does not have a good support for all characters (but you can change it from Sheet's properties). Starting from next build the default font is Tahoma that has support for most characters. Internally Sisulizer handles eveything in Unicode and uses only Unicode version of Windows API. This boxing was caused by wrong font not by Unicode sting converted to Ansi string by mistake. it means that you data is safely in Unicode but Sisulizer only failed to show it correctly on its editor. If you open your text file on Notepad and set font to Arial exactly same things will happen. Notepad as Sisulizer can not shot it correctly but data itself is safely in Unicode. Jaakko
____________________ http://www.sisulizer.com - Three simple steps to localize |
||||||||||||||||
| |||||||||||||||||
| |||||||||||||||||
|
starr Member
|
So sir, the issue remains unresolved. If you follow my steps on an English version of Windows with non-Unicode fallback set to Chinese (Simplified) or simply on a localized Simplified Chinese version of Windows, you are sure to get the same result I described, so it is obvious there's something wrong with the parsing of strings. If everything is handled with Unicode API, some staff other than your part, for example, TNT control or other third-party component, will be guilty. Or it is also possible that there is something wrong during the interoperation between the main program and these components. Treating the txt file as ini works, but it is just a workaround, not good solution.
|
||||||||||||||||
| |||||||||||||||||
|
|||||||||||||||||
|
Jaakko.Salmenius Administrator
|
The point is that I really don't understand what you are trying to tell. The only thing I got was the boxing issue and as I told the problem was wrong font. I have tried to reproduce it several times without any luck. Please send me a Sisulized project file (.slp) that shows the problem. support (at) sisulizer.fi Jaakko
____________________ http://www.sisulizer.com - Three simple steps to localize |
||||||||||||||||
| |||||||||||||||||
| |||||||||||||||||
|
Markus.Kreisel Administrator
|
Hi Starr, as you see, we are all really willing to help with your problem. But none of us can get the point from your description. It might be a language problem. Can you please describe the problem step by step (a-b-c...). What do we need to do to reproduce the problem (step by step)? What does happen? What do you think should happen? Be as detailed as possible. If you have many screenshots please put them in a .zip and send them to support (at) sisulizer.fi. Markus
____________________ http://www.sisulizer.com - Three simple steps to localize |
||||||||||||||||
| |||||||||||||||||
|
|||||||||||||||||
|
banxian Member
|
I think fonts won't display a grayed (empty) to tell coder he has trouble to display all of chars in spec strings. and It seems no bytes between $034C and $034D. yes I can make sure consolas can display 0~9,A-F perfectly here is my text source
Attachment: Copy of consolas cannot display number.png (Downloaded 102 times)
|
||||||||||||||||
| |||||||||||||||||
| |||||||||||||||||
|
banxian Member
|
Hi Markus.Kreisel It's time to reproduce this bug. one weeks ago I make an (HD) video for describe how string been faked. I use an real file produced from itself by StringRipperLite. hxxp://my1993.com/PpsDES/Report/Sisuplayer.rar you can drag mp4 file to mplayer.exe to play it. this action can smothly replay on gbklist.txt upload by starr, too. Condition A (Environment A) Change "Language for non-Unicode programs" to English (United Stats) Condition B Change "Language for non-Unicode programs" to Chinese (PRC) Operation Step1: In Condition A, Create an new project for "gbklist.txt", all things seems fine. save the project. Step2: In Condition B, Load the saved project, press "scan for changes". save project. Step3: In Condition B, Create an new project from "gbklist.txt" with same operation as Step1, save as a new project. Result except All three steps have same fine line in Original Column. Result Occurred In Step 2 and 3, Orginal Column have many (empty) cell. Addition Operation Step 4: Switch to Condition A, load two broken project and press "scan for changes", all (empty) cell change back to content in text lines. It seems codepage affect over sisulizer text source settings. see your http://www.sisulizer.com/support/codepages.shtml (working for codepages) for more information. becare of WStrFromLStr / LStrFromWChar/LStrFromPWChar, they killed many anicode programs.
|
||||||||||||||||
| |||||||||||||||||
|
|||||||||||||||||
|
Janusz Grzybek Super Moderator
|
Hello Banxian, Thank you for detailed information about your issue. Your step by step reproduction information allows me at long last reproduced your problem. Below is screenshot from my PC with empty cells. We will inform you about results of our investigation. Best regards, Janusz Attachment: screen.png (Downloaded 65 times)
____________________ http://www.sisulizer.com - Three simple steps to localize |
||||||||||||||||
| |||||||||||||||||
| |||||||||||||||||
|
Jaakko.Salmenius Administrator
|
I fixed this in 271. The problem was as you explained an automatic conversion from ansi to wide during text file parsing. Delphi before 2009 did not have any warning about string conversion but fortunately the new Delphi has that and than warning would have pinpointed the bug automatically. Thank you for the bug report. Jaakko
____________________ http://www.sisulizer.com - Three simple steps to localize |
||||||||||||||||
| |||||||||||||||||
|
|||||||||||||||||
|
starr Member
|
Hope to see build 279 soon. Thanks for the fix.
|
||||||||||||||||
| |||||||||||||||||
| |||||||||||||||||
|
Janusz Grzybek Super Moderator
|
The new build will come out in a week. Janusz
____________________ http://www.sisulizer.com - Three simple steps to localize |
||||||||||||||||
| |||||||||||||||||
| Current time is 06:27 pm | Page: 1 2 |
| .NET, Delphi, ... - Sisulizer Localization Tool Support > Technical Support (You need to be registered at the forum to write) > Bugs and Quirks in Sisulizer > Strings are trancated or mysteriously filtered | |
Sisulizer software localization tool - Three simple steps to localize