Subject: Re: (dsssl) file I/O and system call in dsssl From: Brandon Ibach <bibach@xxxxxxxxxxxxxx> Date: Thu, 1 Nov 2001 10:09:31 -0600 |
Quoting Trent Shipley <tcshipley@xxxxxxxxxxxxx>: > In an ideal world DSSSL would be a SCHEME library so that the call would use > ANSI SCHEME file IO. > > As far as I know OpenJade can't be used that way . . . > No, it can't, and it shouldn't. Making DSSSL available as a library would have certain advantages, but doing so in order to allow access to these sorts of facilities from within DSSSL scripts would *not* be a good idea. I suppose I'm getting a bit theoretical, here, but I believe the designers of DSSSL had some very good reasons for basing it on the side-effect free subset of Scheme. For starters, this means that the document can be processed in a different order, if it suits the DSSSL engine for performance or other reasons, as long as the result order is maintained. This sort of reordering is made much simpler by the lack of side-effects in DSSSL processing rules. As an example, if your source document specifies a directory to be created, and several files to be copied into that directory, calling functions to do those operations during the processing of the document could cause a problem if the DSSSL engine, for some reason, chose to process the nodes that triggered the file copies before the node that triggered the directory creation. By using the approach of creating the shell script, instead, the DSSSL engine will guarantee that the operations in the script will be in the order that the document specifies them (assuming the stylesheet maintains that) when the processing is complete. -Brandon :) DSSSList info and archive: http://www.mulberrytech.com/dsssl/dssslist
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
Re: (dsssl) file I/O and system cal, Trent Shipley | Thread | Re: (dsssl) file I/O and system cal, roconnor |
Re: (dsssl) file I/O and system cal, Trent Shipley | Date | (dsssl) Abstract w/ tex backend, curt brune |
Month |