Documentation generated from fossil trunk
platform -
System identification support code and utilities
package require platform ?1.0.10? platform::generic platform::identify platform::patterns identifier
The platform package provides several utility commands useful for the identification of the architecture of a machine running Tcl.
Whilst Tcl provides the tcl_platform array for identifying the current architecture (in particular, the platform and machine elements) this is not always sufficient. This is because (on Unix machines) tcl_platform reflects the values returned by the uname command and these are not standardized across platforms and architectures. In addition, on at least one platform (AIX) the tcl_platform(machine) contains the CPU serial number.
Consequently, individual applications need to manipulate the values in tcl_platform (along with the output of system specific utilities) - which is both inconvenient for developers, and introduces the potential for inconsistencies in identifying architectures and in naming conventions.
The platform package prevents such fragmentation - i.e., it establishes a standard naming convention for architectures running Tcl and makes it more convenient for developers to identify the current architecture a Tcl program is running on.
This can be used to allow an application to be shipped with multiple builds of a shared library, so that the same package works on many versions of an operating system. For example:
package require platform # Assume that app script is .../theapp/bin/theapp.tcl set binDir [file dirname [file normalize [info script]]] set libDir [file join $binDir .. lib] set platLibDir [file join $libDir [platform::identify]] load [file join $platLibDir support[info sharedlibextension]]