terminated by signal SIGSEGV (Address boundary error) #72
Comments
|
Few last lines of |
|
Thanks for the report. I think it’s a bug. What language are you parsing? Is the grammar open source? It’d be great to get a reproducible script that causes this. |
|
I parse php using this grammar [EDIT] For the record, tree-sitter-node version in use: |
|
This snippet currently reproduces the error. Delete last element from "keywords" array and error goes away. const Parser = require('tree-sitter')
const PHP = require('tree-sitter-php')
const parser = new Parser()
parser.setLanguage(PHP)
const tree = parser.parse('<?php //')
const keywords = [
'empty_statement',
'named_label_statement',
'expression_statement',
'if_statement',
'switch_statement',
'while_statement',
'do_statement',
'for_statement',
'foreach_statement',
'goto_statement',
'continue_statement',
'break_statement',
];
const query = keywords.reduce((prev, curr) => {
return prev + `(${curr}) @statement`
}, '');
(new Parser.Query(PHP, query)).matches(tree.rootNode) |
|
I can't reproduce the problem using this script. What platform are you on? |
Maybe try a bit larger query on your machine, like: |
|
Hmm, I still can't reproduce it. I also tried repeating the entire If you get a chance, could you rebuild the To rebuild the module: npm install -g node-gyp
cd node_modules/tree-sitter
node-gyp rebuild --debugThen, to run: lldb node -- test.jsSince the trace ends at |
|
Thank you for helping me out! Also, during issue (once you said you cannot reproduce), I updated |
|
So we mark this as bug ? |
|
I have been able to reliably reproduce a seg fault by passing in a single string to a query instead of an S expression, as in I added this to the top of my test: var SegfaultHandler = require('segfault-handler');
SegfaultHandler.registerHandler("crash.log"); |
|
Also note that I arrived here because of jestjs/jest#8769 (comment) I have a more pernicious seg fault that is ONLY triggered in a Docker environment, which has more limited memory. It's possible that this bug is separate from the one above, and is triggered by an out-of-memory error in Query. |
|
Reproduce case:
as in new Query(Ruby, ` (assignment
left: (_) @var
right: (method_call
method: (
call receiver: (constant) @class
(#eq? @class "I18n")
method: (identifier) @method
(#eq? @method "namespace")
)
arguments: (
argument_list (
(string) @namespace
)
)
)
)
`)but ONLY in a Docker context, and only inside a jest test. Note that the seg fault occurs in lib_pthread, so this is a threading issue. I am almost certain the NAPI PR will fix this. const Parser = require("tree-sitter");
const Ruby = require("tree-sitter-ruby");
const { Query } = Parser;
describe("createQuery", () => {
it("?", () => {
new Query(Ruby, `
(assignment
left: (_) @var
right: (method_call
method: (
call receiver: (constant) @class
(#eq? @class "I18n")
method: (identifier) @method
(#eq? @method "namespace")
)
arguments: (
argument_list (
(string) @namespace
)
)
)
)
`
);
});
}); |
|
further context: the lines are responsible. If I remove either or OR remove both predicates (the then the seg fault disappears. |
|
Thank you @cellog!
What is NAPI PR? |
ahlinc
added a commit
to ahlinc/node-tree-sitter
that referenced
this issue
Mar 7, 2021
ahlinc
added a commit
to ahlinc/node-tree-sitter
that referenced
this issue
Mar 8, 2021
|
I wanted to confirm that actually #81 solves this issue for me, and confirm closing this thread, but building the master branch I get this error: make: Entering directory '/home/ada/any-style-new/any-style/node_modules/tree-sitter/build'
make: *** No rule to make target 'Release/obj.target/tree_sitter/vendor/tree-sitter/lib/src/lib.o', needed by 'Release/obj.target/tree_sitter.a'. Stop.
make: Leaving directory '/home/ada/any-style-new/any-style/node_modules/tree-sitter/build'
gyp ERR! build error
gyp ERR! stack Error: `make` failed with exit code: 2
gyp ERR! stack at ChildProcess.onExit (/home/ada/any-style-new/any-style/node_modules/tree-sitter/node_modules/node-gyp/lib/build.js:194:23)
gyp ERR! stack at ChildProcess.emit (node:events:378:20)
gyp ERR! stack at Process.ChildProcess._handle.onexit (node:internal/child_process:290:12)
gyp ERR! System Linux 5.10.13-arch1-1
gyp ERR! command "/usr/bin/node" "/home/ada/any-style-new/any-style/node_modules/tree-sitter/node_modules/.bin/node-gyp" "rebuild"
gyp ERR! cwd /home/ada/any-style-new/any-style/node_modules/tree-sitter
gyp ERR! node -v v15.11.0
gyp ERR! node-gyp -v v6.1.0
gyp ERR! not ok This time again I am not sure if that is my fault or not. When would the #81 be released? |
|
FWIW and in case anyone else runs into this, make sure to clone with submodules: git clone --recurse-submodules |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I am no expert in nodejs - this happens to me when running larger queries. Is this a bug or this means I should do some sort special nodejs configuration?
The text was updated successfully, but these errors were encountered: