This section contains some basic examples that illustrate what
SCA can do to help you with your programs. The examples have very
little explanation. If you want a more detailed explanation of the
underlying concepts, see the Basic_Query_Concepts help topic. The
remainder of this section is written in terms that are specific to
Pascal programs.
If you want to follow along and try the examples, you will need to
have an SCA library available. The SCA$EXAMPLE library provided
with SCA is based on Pascal, so you could use it. If you want
to use your own library, but do not know how to create an SCA
library, read the Building_An_SCA_Library help topic. The examples
in this section use variables from the SCA$EXAMPLE library. If you
use your own library, you will have to substitute variable names
that actually exist in your code when trying the examples.
The first example is the easiest query: It lets you find all the
items in your SCA library named i, and shows you all the places
where they appear (all occurrences of i).
FIND i
You can search for any name in this manner, including using
wildcard characters (for example, FIND i*).
Now let's say you are looking for an occurrence, and you know
that it occurs in a particular file. The following query finds all
occurrences of items that are named i but will then limit them to
those which happen to occur in the file named 'BUILDTABLE.PAS'.
FIND i AND FILE_SPEC="BUILDTABLE.PAS"
Another typical question one might ask is "Find all the places
where this item is assigned to (or read from, called, declared,
and so forth)." The next example finds all occurrences of items
that are named c, but then limits them to only those occurrences
where c is assigned a value.
FIND c AND OCCURRENCE=WRITE
(SCA understands many occurrence classes other then WRITE. See the
help subtopics under Getting_Started for tables containing all the
SCA attributes and their corresponding meanings in Pascal.)
Often, you only want to know where (in what file or module) a
particular procedure is, so that you can go to it and edit it. You
could use a query similar to the first (where i would be replaced
by the name of the procedure) and then look through the output.
The output would include all occurrences of the procedure, one
of which would be its declaration, which you could then select.
Or, you could ask SCA to limit the search for you by typing the
following query:
FIND build_table AND OCCURRENCE=PRIMARY
In SCA terms, a primary declaration is the most significant
declaration of an item. For a Pascal procedure, this means the
place where the procedure is actually implemented. This is in
contrast to FORWARD or EXTERNAL declarations, which are associated
declarations.
Another problem you might have is that there are many different
items in your system having a given name. Some may be variables;
others may be functions, constants, labels, and so forth. Suppose
you want to find only the procedures named 'build_table'. Again,
the query FIND build_table would give you what you wanted, but
it would also give you much more. It is preferable to issue the
following query:
FIND build_table AND SYMBOL_CLASS=PROCEDURE
The last four examples have all selected information based on
two attributes. The last example selected information based on a
name attribute (in this case, 'build_table') and a symbol class
attribute (in this case, PROCEDURE). Note how the attributes
are combined using the boolean operator AND. In general, you
can select items out of your library based on any combination
of attributes, using AND as well as the other logical operators
OR, XOR and NOT.
The next example shows another primary feature of SCA - the
ability to display relationships between items. The example
given here shows the most common use of this feature. It finds
the complete call tree (that is, all procedures called directly
and indirectly), of the procedure named 'build_table'.
FIND CALLED_BY (build_table, DEPTH=ALL)
If you want to limit the depth of the call tree, you can replace
the keyword ALL by any positive integer.
The final part of this section describes how to go directly to the
source code once you have issued a query. After issuing the query
FIND c, for example, you can have an LSE query buffer containing
something that looks like the following:
C variable
EXPAND_STRING\60 VAR (variable) declaration
EXPAND_STRING\75 write reference
EXPAND_STRING\79 read reference
EXPAND_STRING\95 read reference
EXPAND_STRING\122 read reference
EXPAND_STRING\144 write reference
EXPAND_STRING\146 read reference
EXPAND_STRING\149 write reference
EXPAND_STRING\149 read reference
EXPAND_STRING\150 read reference
EXPAND_STRING\166 read reference
The first two lines of this display will be highlighted. The first
line represents the item you looked for (c), and the rest of the
lines represent the different places in the code where this item
occurred (that is, the occurrences of c). By using the up and down
arrows on your keyboard, or by clicking on an occurrence with your
mouse, you can choose the occurrence you want to see. Then all
you have to do is type CTRL/G (the keyboard equivalent of the GOTO
SOURCE command), and LSE will bring the source file into a buffer
and position you at the occurrence you chose.
To obtain help on the following topics, request help as indicated.
o For help on query language, see the Basic_Query_Concepts help
topic.
o For help on libraries, see the Building_An_SCA_Library help
topic.