The parser

Next: The visitor Up: Gasta architecture Previous: Gasta architecture
Gasta Homepage: Home page

The parser

The parser is written using the powerful tools bison and flex. The grammar is very simple. The compiler gcc produces a file like :

**Figure 2:** Output produced by gcc
$\fbox{\begin{minipage}{7cm} \begin{small} \begin{tabbing} tttt \= tttttttttttttt... ...\>\>extern\\ \>\>body:~@7\\ @4\>...\\ \end{tabbing}\end{small}\end{minipage}}$

Our grammar identifies the different parts of the file like :

@X is a NUMBER (the number of the node).
function_decl is an IDENT (the identifier of a node).
name: @2 is a tag with two parts, the TAGLABEL and the TAGVALUE. The tag label is "name:" and the tag value is "@2".

According to the previous definitions, a node of the graph is the association of three items: Its number, its identifier and a list of tags that describe it. So, during the parse, we construct a list of tags. When all tags are added to the list, we prepend the identifier and the node is stored in a hash table. The key to access a particular node in the hash table is its number.

Guillaume Thouvenin 2001-10-01