There are a large number of ways to implement what you want and the devil is in the details. I will assume that you will be dealing with a file system similar to that on Window's/Linux/Mac and not an embedded system, or an old style mainframe with an unusual file system.
Your implementation will need a small library of routines to automatically traverse directories, find source code files, and read them. If you have an IDE the code should also automatically time stamps of files to check whether they have been modified since the last time they were processed and reprocess them if necessary. In the long run in addition to source code files you may also want byte code files or other compressed forms of the source code files that are quicker to process. You will need to provide a means of defining the search path for the code. All of the following is possible 1. A tree based on the file location of your processor. This is typically where the language defined code is located (though some of the language defined "imports" are often implemented directly in the processor than as external code as this can allow additional optimizations and error reporting) 2. The directory or a tree based on the directory where you invoke your processor. If you have an IDE this might be changed dynamically through change directories calls. 3. directories or trees passed in through the command line 4. directories specified on the prompt of an IDE 5. paths specified in a preferences database file. A preferences file should be a high priority item. 6. paths specified through system variables. System variables are over used and awkward to change so I suggest using one only to specify the path to the database file. 7. Normally your users will want all of the above, but I suggest start with the tree based on your processor then implement the preferences file. You will need to define a default mapping of the language identifiers: functions, procedures, modules, classes, etc, to file and/or directory names. You will also in the long run want a language defined means of overriding that default in case of naming conflicts. You will need to define the order in which you search the file space of the files of interest. Some implementations stop when they find a file whose name is compatible with the desired file. I find that that can hide naming conflicts and unintentionally result in the wrong file being called. I suggest you search the entire file tree at the start and always report any conflicts. On Dec 16, 2009, at 4:52 AM, Bujji wrote: > Hi all > > how to implement import statement (like in python: import sys) using > antlr grammar > i have got syntax recognition in examples but how the action should be > performed. > > how the language is able to find that that file is there or not in the > library > > any clues ? > > help me > > Thanks in advance > > Siva > List: http://www.antlr.org/mailman/listinfo/antlr-interest > Unsubscribe: > http://www.antlr.org/mailman/options/antlr-interest/your-email-address -- You received this message because you are subscribed to the Google Groups "il-antlr-interest" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/il-antlr-interest?hl=en.
List: http://www.antlr.org/mailman/listinfo/antlr-interest Unsubscribe: http://www.antlr.org/mailman/options/antlr-interest/your-email-address
