addr2line: 'v': No such file, instead of meaningful output, when build fails due to internal bug

[nkoehring]

Describe the bug

I have the given code, which triggers a known bug in V's ORM module. Running it, results in lots of addr2line errors, instead of meaningful output.

Reproduction Steps

module main

import db.sqlite

@[table: 'visits']
struct Visit {
        id   int    @[primary; sql: serial]
        site string
}

@[table: 'sites']
struct Site {
        hostname string  @[primary]
        owner    int
        visits   []Visit @[fkey: 'site']
}

@[table: 'users']
struct User {
        id    int    @[primary; sql: serial]
        name  string
        sites []Site @[fkey: 'owner']
}

fn init_db(path string) !sqlite.DB {
        mut db := sqlite.connect(path)!

        sql db {
                create table User
                create table Site
                create table Visit
        }!

        return db
}

Expected Behavior

I expect to see the line number and/or memory address, instead of an addr2line error.

Current Behavior

Running v src/dbtest.v results in following output on my machine:

signal 11: segmentation fault
addr2line: 'v': No such file
v() [0x6cff71]
addr2line: 'v': No such file
v() [0x6cfd12]
addr2line: 'v': No such file
v() [0x482afd]
addr2line: 'v': No such file
v() [0x47ddbb]
addr2line: 'v': No such file
v() [0x4a2d48]
addr2line: 'v': No such file
v() [0x475101]
addr2line: 'v': No such file
v() [0x6c83fe]
addr2line: 'v': No such file
v() [0x602a17]
addr2line: 'v': No such file
v() [0x60660a]
addr2line: 'v': No such file
v() [0x608dba]
addr2line: 'v': No such file
v() [0x40272a]
                                                        | 0x7ff3970cd14a | /lib64/libc.so.6(+0x2814a) 
                                                        | 0x7ff3970cd20b | /lib64/libc.so.6(__libc_start_main+0x8b) 
addr2line: 'v': No such file
v() [0x402765]

Possible Solution

No response

Additional Information/Context

The bug was discussed here, before I posted the issue: https://github.com/vlang/v/discussions/20765

V version

V 0.4.4 410bd9d

Environment details (OS name and version, etc.)

V full version: V 0.4.4 a79a9cb.410bd9d OS: linux, Linux version 6.6.13-200.fc39.x86_64 (mockbuild@72f11b2996ed4699b0f705186172808f) (gcc (GCC) 13.2.1 20231205 (Red Hat 13.2.1-6), GNU ld version 2.40-13.fc39) #1 SMP PREEMPT_DYNAMIC Sat Jan 20 18:03:28 UTC 2024 Processor: 16 cpus, 64bit, little endian, AMD Ryzen 7 PRO 5850U with Radeon Graphics

getwd: /home/n/src/pkg/gitly vexe: /home/n/pkg/v/v vexe mtime: 2024-02-08 18:03:24

vroot: OK, value: /home/n/pkg/v VMODULES: OK, value: /home/n/.vmodules VTMP: OK, value: /tmp/v_1000

Git version: git version 2.43.0 Git vroot status: weekly.2024.06-21-g410bd9db (4 commit(s) behind V master) .git/config present: true

CC version: cc (GCC) 13.2.1 20231205 (Red Hat 13.2.1-6) thirdparty/tcc status: thirdparty-linux-amd64 12f392c3

[!NOTE] You can use the 👍 reaction to increase the issue's priority for developers.

Please note that only the 👍 reaction to the issue itself counts as a vote. Other reactions and those to comments will not be taken into account.

[JalonSolov]

I do get a failure, but I don't see anything about addr2line on Manjaro Linux:

I get a failure here, but nothing about addr2line:

/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:47430: at v__ast__Scope_find: RUNTIME ERROR: invalid memory access
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:47469: by v__ast__Scope_find_var
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:11543: by v__checker__Checker_check_field_of_inserting_struct_is_uninitialized
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:11166: by v__checker__Checker_sql_stmt_line
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:11174: by v__checker__Checker_sql_stmt_line
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:11045: by v__checker__Checker_sql_stmt
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:62955: by v__checker__Checker_stmt
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:63403: by v__checker__Checker_stmts_ending_with_expression
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:63379: by v__checker__Checker_stmts
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:4835: by v__checker__Checker_fn_decl
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:62913: by v__checker__Checker_stmt
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:60965: by v__checker__Checker_check
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:61032: by v__checker__Checker_check_files
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:50880: by v__builder__Builder_middle_stages
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:50914: by v__builder__Builder_front_and_middle_stages
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:53360: by v__builder__cbuilder__gen_c
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:53343: by v__builder__cbuilder__build_c
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:53333: by v__builder__cbuilder__compile_c
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:53239: by v__builder__Builder_rebuild
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:52344: by v__builder__compile
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:53674: by main__rebuild
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:53622: by main__main
/tmp/v_1000/v2.01HP6XQK1AQ2T2CXFDNASBPEYW.tmp.c:54331: by main

[spytheman]

On Debian, addr2line is part of the binutils package. Try searching for it, and install it.

[spytheman]

https://github.com/vlang/v/pull/20774 may fix this specific compilation problem. (addr2line still has to be installed separately for others).

[nkoehring]

Fedora has addr2line installed as part of the binutils package.

[JalonSolov]

addr2line is installed. One of the error messages is coming from addr2line:

addr2line: 'v': No such file

This says addr2line was found, but v was not.

[prantlf]

addr2line is installed. One of the error messages is coming from addr2line:

addr2line: 'v': No such file

This says addr2line was found, but v was not.

This looks like a red herring. v was obviously installed, because it executed the tests:

v src/dbtest.v

Then something crashed. The error message "addr2line: 'v': No such file" appears to have nothing in common with it. It's pointing to v, which is being already executed, as the stack frame shows:

signal 11: segmentation fault
addr2line: 'v': No such file
v() [0x6cff71]

I started getting the same error on Linux (GitHub VM) some time ago, when executing vpopen, in a publishing tool, which uses git to commit and push and GitHub REST API to publish a new release. It also complained about the binary name, which was already being executed and which wasn't re-executed from within itself (vp here). I worked around it by running the publishing from the Mac VM instead of from the Linux VM. I get no segmentation faults on Mac.

prepared version 0.0.2 for pushing   <<< the beginning is logged
signal 11: segmentation fault
addr2line: 'vp': No such file
vp(+0xeb593) [0x55ca378d7593]
pushed version 0.0.2                 <<< no failure, although something crashed
addr2line: 'vp': No such file
vp(+0xf75e8) [0x55ca378e35e8]
addr2line: 'vp': No such file
vp(+0xd347) [0x55ca377f9347]
 | 0x7ff4fa829d90 | /lib/x86_64-linux-gnu/libc.so.6(+0x29d90) 
 | 0x7ff4fa829e40 | /lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0x80) 
addr2line: 'vp': No such file
vp(+0xd385) [0x55ca377f9385]
Error: Process completed with exit code 139.  <<< failure from a later system call

The code:

cmd := git push --atomic origin HEAD "v${ver}"

f := vpopen(cmd)
if isnil(f) { return error('executing "${cmd}" failed') }

fd := os.fileno(f)
out := read_all(fd, opts)

exit_code := vpclose(f)
if exit_code != 0 { return error('"${cmd}"" exited with ${exit_code}: ${first_line(out)}') }

But I'd need a local Linux VM and some time to debug it.

[JalonSolov]

What is looks like is that v symlink was never run.

When addr2line was called, it couldn't find v in the path, so gave the No such file message.

[prantlf]

What is looks like is that v symlink was never run.

When addr2line was called, it couldn't find v in the path, so gave the No such file message.

Well, v symlink wasn't run, because I installed v using brew and then continued upgrading it in-place by v up. Homebrew places symlinks to /usr/local/bin:

❯ ll `which v`
lrwxr-xr-x 1 prantlf admin 27 Mar 21 20:47 /usr/local/bin/v -> ../Cellar/vlang/0.4.5/bin/v*

The binary and symlink are installed like this in vlang.rb:

libexec.install "cmd", "thirdparty", "v", "v.mod", "vlib"
bin.install_symlink libexec/"v"

Do you see a problem with this setup?

[JalonSolov]

It's good that the symlink is there. Just saying that's what it looked like. When addr2line tried to run V, it couldn't find it, so it gave the message.

I don't know why it couldn't find it, with that link in place.