stolo93/php_subset_compiler

Project for Formal languages and compilers class

IFJ 2022

This is a project for Formal Languages and Compilers class at FIT BUT.

Implementation of compiler for a given subset of the php language into a wannabe assembly language, which is given in the assingnment.

More about the compiler:

The four main parts of the compiler:

• Lexical analysis
• Syntactic analysis
• Semantic analysis
• Code generator

php_subset_compiler's Issues

vec_ _resize end of a non-void function

DEFINE_VEC_FUNCTIONS_WITH_DESTRUCTOR in vec_resize control reaches end of a non-void function
src/vector.c: In function ‘vec_token_ptr_resize’:
src/headers/vector.h:320:1: error: control reaches end of non-void function [-Werror=return-type]
320 | }
| ^
src/vector.c:22:1: note: in expansion of macro ‘DEFINE_VEC_FUNCTIONS_WITH_DESTRUCTOR’
22 | DEFINE_VEC_FUNCTIONS_WITH_DESTRUCTOR(token_ptr, token_ptr, free);

Make print_err_info togleable with a #define

vec_push_back neuklada od nuli

Funkcia vec_push_back pri pushnuti prveho prvku ulozi prvok na index 1 nie nula co neviem ci je bug alebo to tak ma bit

Change capitalization of error macro

vec_push_front stale segfaultuje

Pri volani funkcie new_vec_with_capacity(n) sa do premenej capacity_ ulozi sizeof(type)*n ( sizeof(type) moze bit vecsi nez 1 ).
https://github.com/stolo93/IFJ_2022/blob/314fa6acaf2f8f286ca7dd04911d80b46b76fa2e/src/headers/vector.h#L54
https://github.com/stolo93/IFJ_2022/blob/314fa6acaf2f8f286ca7dd04911d80b46b76fa2e/src/headers/vector.h#L168

Cize sa do kapacity ulozi velkost alokovaneho priestoru v pamati miesto max poctu prvkov na ulozenie co by sa tam asi malo ukladat.
Tim padom podmienka vo funkcii vec_push_front nema zmysel:
https://github.com/stolo93/IFJ_2022/blob/314fa6acaf2f8f286ca7dd04911d80b46b76fa2e/src/headers/vector.h#L97

lebo ak len_ bude max pocet prvkov kolko mozem ulozit do pola tak to bude vzdy mensie nez alokovany priestor (jedine ak robime s datovim typom char tak by to fungovalo, neskusal som). Dalej ked je tato funkcia volana prvi krat tak len_ bude nula pri volani:
https://github.com/stolo93/IFJ_2022/blob/314fa6acaf2f8f286ca7dd04911d80b46b76fa2e/src/headers/vector.h#L43

cize nastane (size_t)0-1 co podtecie a for loop segfaultne.

Add recomendation to use --force-with-lease

In CONTRIBUTING.md we mention the --force flag we should instead mention --force-with-lease.

Interner may produce undefined behaviour if a new string is interned.

Details:
If the interner already has at least one string in it then interning a string which should be placed at the end of the backing array when sorting with strcmp, the intern function will produce a pointer one past the end of the backing array and then attempt to use it in a strcmp function. strcmp will then attempt to de-reference this pointer and if we're lucky segfault.

Investigate how to make gh actions run on private repository

print_err_info prints to stdout

print_err_info uses printf instead of fprintf.

Vec push operations are broken

• vec_push_front will segfault on empty vectotrs

htab.c not included in Makefile src path

Hashing table can not be built

htab.c should either be placed in the src file or path to this file has to be given to Makefile explicitly in VPATH variable (I may not be correct about this).