How can I unserialize a s-exp then seralize it and avoid upper-casing?

Question

I am supposed to read a complex s-expression tree, modify some nodes and save it somewhere.

It appears that in process the 'read function is "modifying" the input. For instance in a simple example:

CL-USER> (read-from-string "(seq_num 123)")
(SEQ_NUM 123)
13

You can see it capitalize nodes (as well as values).

As well it appears it can add pipes to the outmost left and right of a value. As in:

CL-USER> (read-from-string "(password 000006013H)")
(PASSWORD |000006013H|)
21

It adds pipes!

Is there a way to tell 'read not to do that? I guess such modification are done for good reason when the s-expression is actually a valid LISP program. But here it is not the case. Let see that file as an XML file. A simple configuration file which appears to be a s-expr. I don't need him to "intern" the symbols it reads. I just need him to unserialise it as a tree, since for me it is the easiest way to search the tree then ( 'car and 'cdr are no nice).

Well if the tree is formed then every symbols must be interned.. Told in another words, how can I tell him to intern no symbols, but instead keep it as strings. (thus it could form the cons-tree but instead of pointing to symbols, it would point to characters strings. You see what I mean?)

Answer 1

The reader will by default internalize symbols. Note that both the reader and printer affect how symbols are used and appear. If you want to see the real case of a symbol call (symbol-name some-symbol). The printer will try to escape a symbol if necessary, such that it can be read back and the same case gets used.

CL-USER 26 > 'foo
FOO

CL-USER 27 > 'f\oo
|FoO|

CL-USER 28 > (symbol-name 'f\oo)
"FoO"

The reader allows the control of how a symbol gets read. See below.

A few things to know:

• all symbols are by default uppercase internally. By default the reader uppercases lowercase characters.

• a symbol can contain arbitrary characters, including lowercase characters. The symbol then needs to use escapes:

Example:

|This is a valid symbol.|

Not that the vertical bars are not part of the symbol name. They are used to escape the symbol. Another escape character is the backslash:

1\a2

Above is also a symbol.

• note that things that contain numbers or characters can also be symbols or numbers, depending on the read base:

Example:

00a

Above is a symbol in the reader base 10.

Same Example, other read base:

00a

Above is a number in the reader base 16.

• a non-interned symbol (not in a package) is written like this:

Example:

#:non-interned-symbol

• a keyword symbol:

Example:

:keyword-symbol

How can you influence in which case a symbol is created/looked-up during read?

• you can escape the symbol, see above

• Use a readtable with a different case.

Example in the Common Lisp Hyperspec: 23.1.2.1 Examples of Effect of Readtable Case on the Lisp Reader

Turn off escaping in the printer

CL-USER 36 > (let ((*print-escape* nil))
               (write (read-from-string "(|passWord| 000006013H)")))
(passWord 000006013H)
...