This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
Platform: Solaris10u4(sparc/x86/x64) Locale: zh 1. launch nb60beta1 2. create cnd project result: Folder name's Chinese characters display "??", they are "Header Files" and "Resource Files" "Source Files" and "Important Files" are fine This issue doesn't exist on zh_CN.UTF8 locale and other platfrom This issue doesn't exist on ja locale. screenshot is attached
Created attachment 51260 [details] screenshot
Are they localization issues? Header Files and Resource Files are not handled any different than Source Files.
As we know, this issue only exists on Solaris zh locale. It doesn't exist on Solaris zh_CN.UTF8 or windows/linux... If it's l10n issue, all platform should have this issue. So i did the investigation for "Header Files", #1. the source is: makeproject/src/org/netbeans/modules/cnd/makeproject/api/configurations/Bundle.properties en: HeaderFilesTxt=Header Files zh: HeaderFilesTxt=头文件 ascii: HeaderFilesTxt=\u5934\u6587\u4ef6 头 => \u5934 文件 => \u6587\u4ef6 UI display：？文件 result: "头"(ascii code is \u5934) displayed "?" #2. launch the IDE on Solaris zh locale, create a C++ header file, please see screen shot in the attachment. result: "头" display well. I checked the source, it's also \u5934 Maybe the display method in the code are different between displaying folder name and create file menu, so it seemed coding issue.
Ken Frank: more experiment i set zh locale solaris, then dumpcs and found that char at cdb7 (I dont know if it relates to the \u5934 Will mentions so i created java project with name having 1 character to left of that char and one to the right, and it shows ok but using that char gives same ? in explorer and exception - java.lang.IllegalArgumentException: URI has a query component at java.io.File.<init>(File.java:372) at org.netbeans.modules.java.source.usages.RepositoryUpdater$CompileWorker$1.run(RepositoryUpdater.java:1372) at org.netbeans.modules.java.source.usages.RepositoryUpdater$CompileWorker$1.run(RepositoryUpdater.java:1122) at org.netbeans.modules.java.source.usages.ClassIndexManager.writeLock(ClassIndexManager.java:100) at org.netbeans.modules.java.source.usages.RepositoryUpdater$CompileWorker.run(RepositoryUpdater.java:1119) at org.netbeans.modules.java.source.usages.RepositoryUpdater$CompileWorker.run(RepositoryUpdater.java:1092) at org.netbeans.api.java.source.JavaSource$CompilationJob.run(JavaSource.java:1439) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:441) at java.util.concurrent.FutureTask$Sync.innerRun(FutureTask.java:303) at java.util.concurrent.FutureTask.run(FutureTask.java:138) at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:885) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:907) [catch] at java.lang.Thread.run(Thread.java:619) Ken Frank wrote: > not sure its a cnd thing; i create java project and use that string Will uses as part of name of > java class > > (first i am in ja locale; euc-jp proj encoding) > > at first the mbyte shows ok, then it changes and the first letter is shown as ? > > i change project encoding to gb18030 or gbk, then create a new class again > and same situation, and same using utf-8 project encoding > > thus some interpretattion is happening between when it first shows in explorer ok > and one second later when it shows the ? > > its true i have started nb from ja locale but i think project encoding value can emulate > what is being seen by Will. > > cnd does not have project encoding property so it uses (or should use) the encoding > of the locale the user is in, just as was done for pre nb6 projects. > > Will, is that character with ? perhaps not valid in some zh encodings ? > or in some range not part of some zh encodings ? > > no, that can;'t be it; i see the char ok in editor, in members view, in refactor rename > so seems like something about explorer, and why it looks ok at first and then the ? > > > Thanks - Ken > > > Will Zhang wrote: >> Hi Thomas, >> >> I also felt strange. >> As we know, this issue only exists on Solaris zh locale. It doesn't exist on Solaris zh_CN.UTF8 or windows/linux... >> If it's l10n issue, all platform should have this issue. >> So i did the investigation for "Header Files", >> #1. the source is: >> makeproject/src/org/netbeans/modules/cnd/makeproject/api/configurations/Bundle.properties >> en: HeaderFilesTxt=Header Files >> zh: HeaderFilesTxt=头文件 >> ascii: HeaderFilesTxt=\u5934\u6587\u4ef6 >> 头 => \u5934 >> 文件 => \u6587\u4ef6 >> UI display：？文件 >> result: "头"(ascii code is \u5934) displayed "?" >> #2. launch the IDE on Solaris zh locale, create a C++ header file, please see screen shot in the attachment. >> result: "头" display well. I checked the source, it's also \u5934 >> >> Maybe the display method in the code are different between displaying folder name and create file menu, so it seemed coding issue. >> Since i didn't meet this situation before, maybe i'm wrong. >> Hope it will be helpful.
Ken: I'm glad it was ok for you on solaris sparc; perhaps I saw the problem since i was running nb in ja locale, but then if so, I wonder why did I see the problem just with that character, and its the same char you saw problem with. (I think it is anyway; is the one I used from dumpcs \u5934 the same value as the char you used that caused the ? to appear ? or is it that on solaris sparc you see problem just in cnd area and not with java project ? and is there any exception in log or window ? also, when you see the problem in explorer, does the character look ok for one brief moment, and then become the ? Thanks - Ken Will Zhang wrote: > Hi Ken, > > Thank you for evaluation. > I tried to reproduce the issue you mentioned on Solaris sparc gb18030 and gbk locales with nb6beta1 + l10n jars, but i can't see ? > screenshot was attached. > I create two java project with name "头文件" and "头文件1" then create two java class in project "头文件" with name "头文件" and "头文件1" > Chinses characters display well in explorer, there is no ? > Are the steps above wrong? or please let me know the detail how to reproduce it. > > thank you, > Will
downgrading to original priority p3.
What is the status on this? Do you still see a problem?
This issue still can be reporduced in netbeans-trunk-nightly-200808241401-ml-cpp-solaris-sparc.sh
ok
See also CR 6863000.
I've checked generated project files, and have realized incorrect encoding is used in <project_directory>/nbproject/configuration.xml zhcn@gimli[195] $ head -3 configurations.xml <?xml version="1.0" encoding="EUC-JP"?> <configurationDescriptor version="62"> <logicalFolder name="root" displayName="root" projectFiles="true"> "EUC-JP" is a Japanese encoding. "UTF-8" (UNICODE) or "GB2312" (Chinese encoding) should be used. I would suggest to use "UTF-8" encoding for all the environment. I guess the following source fragment is causing this issue: cnd/src/org/netbeans/modules/cnd/api/xml/XMLDocWriter.java ----- 82 protected String encoding() { 83 String lang = System.getenv("LANG"); // NOI18N 84 String encoding = "UTF-8"; // NOI18N 85 if (lang != null) { 86 if (lang.equals("zh") || // NOI18N 87 lang.equals("zh.GBK") || // NOI18N 88 lang.equals("zh_CN.EUC") || // NOI18N 89 lang.equals("zh_CN.GB18030") || // NOI18N 90 lang.equals("zh_CN") || // NOI18N 91 lang.equals("zh_CN.GBK")) { // NOI18N 92 93 encoding = "EUC-JP"; // NOI18N 94 95 } else if (lang.equals("ja") || // NOI18N 96 lang.equals("ja_JP.eucJP")) { // NOI18N 97 98 encoding = "EUC-JP"; // NOI18N 99 } else { 100 encoding = "UTF-8"; // NOI18N 101 } 102 } 103 return encoding; 104 } ----- The "EUC-JP" in line #93 needs to be fixed. Strictly, the returned encoding name should be "GB2312" when 'lang' is one of "zh*" encodings. However, I would suggest encoding() method always return "UTF-8" because "UTF-8" is preferred as XML encoding for all the language environment.
Thanks for the evaluation. Keiichi, can you please confirm the following change: 82 protected String encoding() { 83 // String lang = System.getenv("LANG"); // NOI18N 84 String encoding = "UTF-8"; // NOI18N 85 // if (lang != null) { 86 // if (lang.equals("zh") || // NOI18N 87 // lang.equals("zh.GBK") || // NOI18N 88 // lang.equals("zh_CN.EUC") || // NOI18N 89 // lang.equals("zh_CN.GB18030") || // NOI18N 90 // lang.equals("zh_CN") || // NOI18N 91 // lang.equals("zh_CN.GBK")) { // NOI18N 92 // 93 // encoding = "EUC-JP"; // NOI18N 94 // 95 // } else if (lang.equals("ja") || // NOI18N 96 // lang.equals("ja_JP.eucJP")) { // NOI18N 97 // 98 // encoding = "EUC-JP"; // NOI18N 99 // } else { 100 // encoding = "UTF-8"; // NOI18N 101 // } 102 // } 103 return encoding; 104 }
Fixed.
Thank you Thomas for your prompt action. The fixing seems OK. I will verify in the development build.
Integrated into 'main-golden', will be available in build *201001271614* on http://bits.netbeans.org/dev/nightly/ (upload may still be in progress) Changeset: http://hg.netbeans.org/main/rev/41d1c384b499 User: Thomas Preisler <thp@netbeans.org> Log: #119431 - Chinese characters display "??" on project folder
Created attachment 94454 [details] screenshot in fixed build Verified in build 201002230200. Thank you for your fixing. I've also confirmed the saved XML encoding is now "UTF-8". Sorry for my late verification.
Change status to VERIFIED
Yeah.... Thanks.