Repository navigation
Partially initialized multidimensional VLAs construct the wrong number of elements #557
lexx42xxel
started this conversation in
Issue Triage - Compilation
Replies: 2 comments
|
Found while implementing #561 (static/preallocated array construction without the allocating vector runtime). |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Issue Summary
E a[3][n] = {{E(1)}, {}};constructs more elements than the array has (stack buffer overflow under ASan) and destroys fewer; some shapes hit an assertion.Reproducing Source Code
Command-line Options
eccp --c++17 --g++ --exceptions t.cpp -o t && ./t
Type of Issue
[Back End] There is a problem with the generated C code.
Additional Details
Version: EDG C/C++ front end 7.0, main at commit
6a6c7f2d6bd8027f1b0c90877b4ba222fbff6642(2026-11-15).Actual output:
Expected: constructed 6, destroyed 6
The repeat count of the remaining elements is taken from the total element count of the array, and
the starting element is not scaled for rows of variable length.
The program exits nonzero when the bug is present. A fix with a regression
test is ready (branch
fix/vla-array-init-lowering); a PR will reference this issue.AddressSanitizer
AddressSanitizer detects the same bug in this variant, which has no counters or output of its own:
Report on main (6a6c7f2), addresses elided:
No report with the fix, nor with g++ (
-fsanitize=address).Reproducing (Standard) Configuration(s)
linux-gcc-debug
Reproducing (Non-standard) Configuration
No response
I acknowledge that:
LICENSE.txtfile in my.zipor.tar.gzupload. If I included aLICENSE.txtfile, any tests created using the reproducing source code will be governed by the terms of the license I provided instead.All reactions