Known Problems and Limitations


Cannot load uni-local collection and Asian-locale collection simultaneously

 

One K2 Server or indexing application cannot load a collection using the multilanguage (uni) locale and a collection using an Asian locale (japanb , koreab , simpcb , tradcb ) into the same process. (97590, 98187, 100603)

Workaround: Disable support for those Asian languages in the uni locale. See Incompatibility Between uni Locale and Asian Locales.

Installer fails on non-English Windows systems

 

The Locales installer does not function on Windows machines whose native operating-system language is not English, because registry entries specifying existing Verity products cannot be read. (97169)

Installer launch can be slow on UNIX and Windows

 

When the Locales installer is launched, it may take up to several minutes for the Installer Welcome screen to appear. During that interval, it may appear as if the installer is not running. (97138)

Workaround: None. Give the Installer screen several minutes to appear, rather than attempting to re-launch it.

Microsoft Word 4.0 (Mac) documents not indexable in uni locale

 

Documents created in Microsoft Word for Macintosh v. 4.0 have no character-set information. When such a document is indexed, its character set cannot be determined and the format filter (flt_kv ) assigns UTF-8 as the character set. Therefore, it cannot be indexed correctly. (83969)

-charmap error if collection name has different character set from collection content

 

In command-line tools such as vspider , mkvdk , and rcvdk , if you use the -charmap option to specify a collection name that uses a different character set from the locale’s internal character set, an error occurs. For example, to use rcvdk on a multilanguage-locale collection whose name is written in shift-jis, you might want to use this command:

rcvdk -locale uni -charmap sjis collection_name_in_sjis

In this situation, the search stops and an error is displayed. (84618)

Workaround: Do not specify the -charmap option. A collection’s name should be in either the same character set as the collection’s content, or (when searching) the same character set as the query string’s characters. Using only ASCII characters in collection names is strongly recommended.

Language detection is configured for only a subset of supported languages

 

Out of the box, language detection is configured to support fewer than all languages supported by Verity Locales. To go beyond those languages, change the file common/langid/langlist.cfg , based on the full set listed in the file langlist.all .

Installer does not accept multi-byte input

 

In general, the K2 Dashboard and K2 Services (K2 Administration Server, K2 Broker, etc.) support multi-byte characters. However, during installation, multi-byte characters cannot be used in input fields or paths; they should be composed of 7-bit ASCII characters only. (72738)

Language ID is required when using testqp with the uni locale

 

When you use the testqp command-line tool with the multilanguage locale, you must specify the language ID of the input query, as in

testqp -locale uni/ja

In this example the query is in Japanese. (90959)

Some character sets not compatible with some locales

 

By default, when working with a collection, you must use a character set that is one of the supported sets for the collection’s locale. For example, you cannot use the options -charmap 437 or -charmap 850 on a command-line tool that accesses a uni collection. This restriction exists because the standard character-set conversion files 437.cs , 850.cs , and 1252.cs do not contain the required conversion tables. (85462)

Workaround: Remove the appropriate file (437 .cs , 850.cs , or 1252.cs ) from the directory product_dir/common , forcing use of the default character-set driver kvcs.cs .

Incompatibilities between Internet-style query parser and the multilanguage locale

 

Because the Internet-style query parser template files (Internet, Internet_Basic, Internet_Advanced, Internet_AdvancedWeb) do not allow imbedding explicit VQL, you cannot by default use the LANG modifier with the Internet-style query parser to specify a language preference per query. 

Workaround: You can specify a default language in the K2 collection properties by using uni/ zz, where zz is a language code. For full integration of VQL, you can programmatically combine VQL with the Internet-style query parser as described in the query parsers appendix of the Verity Query Language and Topic Guide.

Names and paths for Verity structures should be ASCII only

 

It is recommended that pathnames specifying the locations of Verity topic sets, collections, parametric indexes, and so on contain only 7-bit ASCII characters (and no spaces). This recommendation is made because of the differences in handling of pathname characters between the Dashboard and the command-line tools. (85328)

Using non-1252 characters in the name of a uni collection causes errors

 

If you create a collection in the multilanguage locale and give it a name that contains non-1252 characters, indexing with vspider will cause truncation of the collection’s name, making it unsearchable using its original name. (86787)

Workaround: Do not use non-1252 characters in collection names.

Asian and multilanguage locale constraints

 

(For OEM customers) If your C code application uses Verity Asian locales or the multilanguage locale, the application must be compiled and linked in a C++ compiler. Otherwise, it can be built in C or C++ compiler.

Messages, keywords are in English only in multilanguage locale

 

Keywords, such as VQL operators, Date name, Week name, are in English only. All error messages are in English too. Date value must be in English too.

Workaround: If another language version is required, use the vdk30.rsd and vdk30.rst files from another locale.

Limitations on charset and document formats

 

UCS-2 or pure 2-byte unicode is not supported. (UTF-8 is supported). HZ for Chinese is not supported.

Documentation is not installed by Locales installer

 

The Verity Locale Configuration Guide is included on the Locales CD, but is not installed by the Locales installer. (87091)

Workaround: The Verity Locale Configuration Guide is installed as part of either K2 or VDK installation, so the copy on the Locales CD does not need to be installed. If you wish to use it, however, copy the directory docs/ on the Locales CD to an appropriate location in your installation (for example, rename the directory to locales_docs/ and place it in verity_dir/data ).