[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Line-truncation bug in host_execute?
From: |
David Malcolm |
Subject: |
Line-truncation bug in host_execute? |
Date: |
Fri, 22 Jan 2016 16:53:00 -0500 |
I'm the maintainer of the "jit" component within gcc.
We're carrying a patched copy of dejagnu's host_execute,
"fixed_host_execute", for reasons given in:
http://lists.gnu.org/archive/html/dejagnu/2014-10/msg00007.html
though I think we may have drifted further.
I'm running into an issue with the patched copy that I think may also
affect the pristine one:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69435
My tests use dejagnu.h to write "PASSED" etc lines, some of which are
very long. They're sometimes being truncated by fixed_host_execute when
re-emitted as PASS results, so e.g.
PASSED: test-combination.c.exe iteration 1 of 5: make_calc_discriminant:
actual: "q->b * q->b - (double)4 * q->a * q->c" == expected: "q->b * q->b -
(double)4 * q->a * q->c"
is re-emitted as:
PASS: test-combination.c.exe iteration 1 of 5: make_calc_discriminant: actual:
"q-
(note the truncation in mid-expression).
The exact point of truncation varies from run to run, leading to churn
when comparing .sum files, so this is a pain.
I'm attaching the patch that I've attached to that gcc bug; does this
fix seem reasonable? (my Tcl and expect skills aren't great, sorry).
Thanks; hope this is helpful
Dave
0001-Fix-PR-jit-69435-truncated-lines-in-jit.log.patch
Description: Text Data
- Line-truncation bug in host_execute?,
David Malcolm <=