Opened 7 months ago

Closed 6 months ago

#5486 closed enhancement (fixed)

gcc-14.1

Reported by: Bruce Dubbs Owned by: lfs-book
Priority: normal Milestone: 12.2
Component: Book Version: git
Severity: normal Keywords:
Cc:

Description

New major version.

Change History (10)

comment:1 by Xi Ruoyao, 7 months ago

I've put some info at wiki/gcc14-todo.

in reply to:  1 ; comment:2 by Joe Locash, 7 months ago

Replying to Xi Ruoyao:

I've put some info at wiki/gcc14-todo.

Is it safe to assume that info is only for the BLFS side of things? I had expect and inetutils bomb in LFS.

I have more packages that are failing in BLFS. A consolidated list of what is failing to build and the fixes would be nice.

in reply to:  2 comment:3 by Xi Ruoyao, 7 months ago

Replying to jlocash:

Replying to Xi Ruoyao:

I've put some info at wiki/gcc14-todo.

Is it safe to assume that info is only for the BLFS side of things? I had expect and inetutils bomb in LFS.

Isn't the link to LFS wiki?

I have more packages that are failing in BLFS. A consolidated list of what is failing to build and the fixes would be nice.

For BLFS ​blfs:wiki/gcc14-todo.

Last edited 7 months ago by Xi Ruoyao (previous) (diff)

comment:4 by Bruce Dubbs, 7 months ago

I added --enable-host-pie to Chapter 8 and in the test section:

sed -i '/cpython/d' ../gcc/testsuite/gcc.dg/plugin/plugin.exp

The only failures now are: 

FAIL: gcc.target/i386/fentryname3.c scan-assembler call\\t\\\\*nop@GOTPCREL
FAIL: gcc.target/i386/pr113689-1.c (test for excess errors)
FAIL: libgomp.c++/../libgomp.c-c++-common/pr109062.c output pattern test
FAIL: libgomp.c/../libgomp.c-c++-common/pr109062.c output pattern test

I can easily fix pr113689-1.c and fentryname3.c with a seds. Not sure about pr109062.c, but if we can fix or remove that, we would get a completely clean gcc test, at least for x86_64.

At -j28 and SBU time=80, the time was about 14 minutes.

in reply to:  4 ; comment:5 by Xi Ruoyao, 7 months ago

Replying to Bruce Dubbs:

I added --enable-host-pie to Chapter 8 and in the test section:

sed -i '/cpython/d' ../gcc/testsuite/gcc.dg/plugin/plugin.exp

The only failures now are: 

FAIL: gcc.target/i386/fentryname3.c scan-assembler call\\t\\\\*nop@GOTPCREL
FAIL: gcc.target/i386/pr113689-1.c (test for excess errors)
FAIL: libgomp.c++/../libgomp.c-c++-common/pr109062.c output pattern test
FAIL: libgomp.c/../libgomp.c-c++-common/pr109062.c output pattern test

I can easily fix pr113689-1.c and fentryname3.c with a seds. Not sure about pr109062.c, but if we can fix or remove that, we would get a completely clean gcc test, at least for x86_64.

At -j28 and SBU time=80, the time was about 14 minutes.

I cannot reproduce the pr109062 failure. Can you compile it (using installed GCC 14) with -fopenmp and run the generated executable and paste the result?

Also check gcc/testsuite/gcc/gcc.log.

in reply to:  5 comment:6 by Xi Ruoyao, 7 months ago

Also check gcc/testsuite/gcc/gcc.log.

Oops, should be x86_64-pc-linux-gnu/libgomp/testsuite/libgomp.log.

comment:7 by Xi Ruoyao, 7 months ago

A bug may prevent us from building fmt in BLFS: ​https://gcc.gnu.org/PR114994

Maybe we need to patch GCC.

in reply to:  5 comment:8 by Xi Ruoyao, 7 months ago

The pr109062.c failure reported as ​https://gcc.gnu.org/PR115001.

in reply to:  7 comment:9 by Xi Ruoyao, 7 months ago

Replying to Xi Ruoyao:

A bug may prevent us from building fmt in BLFS: ​https://gcc.gnu.org/PR114994

Maybe we need to patch GCC.

Too pessimistic. It does not prevent building fmt (AFAIK) but it may break packages using fmt. In BLFS it's only paps, so if paps is fine we are "fine."

comment:10 by Bruce Dubbs, 6 months ago

Resolution: → fixed
Status: new → closed

Fixed at commit 340e17adc6.

Note: See TracTickets for help on using tickets.