Segmentation fault (core dumped)

[Zunea]

When the .pyc file is large, pycdc outputs Segmentation fault (core dumped)

[rocky]

How large is your .pyc file?

[Zunea]

yes. >=27kb .pyc

[Zunea]

Is there any way to solve it?

[alxnegrila]

It seems to be linked with the *.pyc file size or the depth of the code. Maybe allocating memory on loading the file.

(gdb) run sk.pyc
Starting program: /usr/local/bin/pycdc sk.pyc
# Source Generated with Decompyle++
# File: sk.pyc (Python 3.10)


Program received signal SIGSEGV, Segmentation fault.
0x00005555555b59c0 in std::vector<PycRef<ASTNode>, std::allocator<PycRef<ASTNode> > >::size() const ()

[0x3C50]

Just now got the same crash,

Program received signal SIGSEGV, Segmentation fault.
0x00005555555b5d78 in std::vector<PycRef<ASTNode>, std::allocator<PycRef<ASTNode> > >::size() const ()

[MoodMonitor]

Hello, I felt into same issue, but somehow I found the clue. The problem was, that the stack was empty but still program want to get the value from.
My solution for this (hot fix, without further investigation):
File ASTree.cpp:

Line 1187:
PycRef<ASTBlock> tmp = curblock;
blocks.pop();

if (blocks.size() == 0) {
break; } else {
curblock = blocks.top();} // changes

if (should_add_for_block) {
    curblock->append(tmp.cast<ASTNode>());
}

Line 1268:
/* Special case */
if (blocks.size() != 0) {
prev = blocks.top(); // changes
} 

Line 1353:
/* Special case */
if (blocks.size() != 0) {
prev = blocks.top(); // changes
}

Line 1363:
if (prev->blktype() == ASTBlock::BLK_MAIN) {
    /* Something went out of control! */
    prev = nil;
}
if (blocks.size() == 0) {
prev = nil; // changes
}

Line 1398:
if (blocks.size() == 0) {
   curblock->setEnd(pos+offs); // changes
} else {
    curblock = blocks.top(); // changes
}

if (curblock->blktype() == ASTBlock::BLK_EXCEPT) {
  curblock->setEnd(pos+offs);
}

Adding ASTree.cpp:
ASTree.txt
Hope will work with your case!

[nataliawilsona10]

Hello, I felt into same issue, but somehow I found the clue. The problem was, that the stack was empty but still program want to get the value from. My solution for this (hot fix, without further investigation): File ASTree.cpp:

Line 1187:
PycRef<ASTBlock> tmp = curblock;
blocks.pop();

if (blocks.size() == 0) {
break; } else {
curblock = blocks.top();} // changes

if (should_add_for_block) {
    curblock->append(tmp.cast<ASTNode>());
}

Line 1268:
/* Special case */
if (blocks.size() != 0) {
prev = blocks.top(); // changes
} 

Line 1353:
/* Special case */
if (blocks.size() != 0) {
prev = blocks.top(); // changes
}

Line 1363:
if (prev->blktype() == ASTBlock::BLK_MAIN) {
    /* Something went out of control! */
    prev = nil;
}
if (blocks.size() == 0) {
prev = nil; // changes
}

Line 1398:
if (blocks.size() == 0) {
   curblock->setEnd(pos+offs); // changes
} else {
    curblock = blocks.top(); // changes
}

if (curblock->blktype() == ASTBlock::BLK_EXCEPT) {
  curblock->setEnd(pos+offs);
}

Adding ASTree.cpp: ASTree.txt Hope will work with your case!

Thank you so much brother you saved my works!

[SuperKenVery]

It seems that the code has been updated and I can't find the places to apply the patches...

Any updates? This problem makes pycdc pretty much unusable for some input...

[kmontag]

I've packaged the fix into a PR: #532. Thanks @MoodMonitor !