|
Документ взят из кэша поисковой машины. Адрес
оригинального документа
: http://www.astro.louisville.edu/software/sbig/archive/xmccd-4.1/xmccd-4.1e/xpa-2.1.14/doc/template.html
Дата изменения: Thu Dec 25 05:27:48 2003 Дата индексирования: Thu Feb 27 22:43:38 2014 Кодировка: Поисковые слова: п п п п п п п п п п п п п п |
XPA access points are composed of two parts: a general class and a specific name. Both parts accept template characters so that you can send/retrieve data to/from multiple servers at one time.
When XPA servers call XPANew(), or XPACmdNew() to define XPA access points, they specify a string identifier composed of a class and a name. When clients communicate with XPA access points, they specify which access points to communicate with using an identifier of the form:
class:nameAll registered XPA access points that match the specified identifier will be available for communication (subject to access control rules, etc.)
As of XPA 2.1.5, the length of both the class and name designations are limited to 1024 characters.
The XPA class:name identifier actually is a template: it accepts wild cards in its syntax, so a single specifier can match more than one XPA access point. (Note that the class is optional and defaults to "*".) The allowed syntax for clients to specify the class:name template is of the form shown below. (Note that "*" is used to denote a generic wild card, but other wild cards characters are supported, as described below).
template explanation -------- ----------- class:name exact match of class and name name match any class with this name *:name match any class with this name class:* match any name of this class *:* match any access point
In general, the following wild-cards can be applied to class and name:
wildcard explanation -------- ----------- ? match any character, but there must be one * match anything, or nothing [...] match an inclusive set
Although the class:name template normally is used to refer to XPA access points, these also can be specified using their individual socket identifiers. For inet sockets, the socket identifier is ip:port, where ip can be the DNS-registered name, the ASCII IP number (e.g. 123.45.67.890) or the hex IP number (e.g. 838f3a60). For unix sockets, the identifier is the socket file name. These socket identifiers are displayed as the fourth argument in the xpans display of registered access points. For example, consider the ds9 program started using inet sockets. The xpans name server will register something like this:
csh> xpaget xpans DS9 ds9 gs saord.harvard.edu:3236 ericYou can access ds9 using ip:3236 in any of the three forms:
csh> xpaget saord:3236 file /home/eric/data/snr.ev csh> xpaget 123.45.67.890:3236 file /home/eric/data/snr.ev csh> xpaget 838f3a60:3236 file /home/eric/data/snr.evIn the case of unix sockets, the socket identifier is a file:
csh> xpaget xpans DS9 ds9 gs /tmp/.xpa/DS9_ds9.2631 eric csh> xpaget /tmp/.xpa/DS9_ds9.2631 file /home/eric/data/snr.evThis feature can be useful in distinguishing between multiple instances of a program that all have the same class:name designation.