Package Details: gcc5 5.5.0-3

Git Clone URL: https://aur.archlinux.org/gcc5.git (read-only, click to copy)
Package Base: gcc5
Description: The GNU Compiler Collection (5.x)
Upstream URL: https://gcc.gnu.org/gcc-5/
Licenses: GPL, custom, LGPL, FDL
Submitter: Svenstaro
Maintainer: severach
Last Packager: severach
Votes: 14
Popularity: 0.000302
First Submitted: 2018-05-07 18:23
Last Updated: 2021-09-28 17:20

Latest Comments

1 2 3 Next › Last »

severach commented on 2021-09-28 17:24

PKGEXT is from back when the default was the atrociously slow xz, completely unnecessary for large AUR packages. Now the default is zst which is no better than gz but tolerable.

bartus commented on 2021-09-26 10:23

@severach: Why PKGBUILD overrides the compression scheme defined in makepkg.conf any reason for preventing user selection they desired compression?

This sort of messes up my local repo configuration (╯°益°)╯彡┻━┻

Problematic declaration: https://aur.archlinux.org/cgit/aur.git/tree/PKGBUILD?h=gcc5#n36

Also, could you drop the autodetect of processor threads count and allow users to define their preferred multithreading scheme with MAKEFLAGS in makepkg.conf or as env var passed with makepkg MAKEFLAGS="-jxxx call.

gnaggnoyil commented on 2021-09-19 20:32

@yangwenbo99 You should set your MAKEFLAGS to -j1 and rerun makepkg to locate where the error is.

yangwenbo99 commented on 2021-09-08 20:50

I got the error during compilation. However, I cannot locate the exact source of the error.

...
mkdir -p -- .deps
make[1]: *** [Makefile:21181: stage1-bubble] Error 2
make: *** [Makefile:905: all] Error 2
==> ERROR: A failure occurred in build().
/usr/share/makepkg/util/message.sh: line 51: QUIET: unbound variable
/usr/bin/makepkg: line 131: logpipe: unbound variable
error making: gcc5

gnaggnoyil commented on 2021-05-30 16:04

Thanks. Confirmed this fix solves my issue.

BTW (probably) the same issue happens on gcc48 package too. Maybe we can apply the same fix to that one?

severach commented on 2021-05-28 20:48

That's not the best fix but it's the only one I could find.

gnaggnoyil commented on 2021-05-26 12:46

GCC 11 upgrades its default language standard mode from gnu++14 to gnu++17, meanwhile GCC 5 src contains code that is no longer valid since C++17.

Thus, building this package with GCC 11 reports the following error:

../../gcc/reload1.c: In function ‘void init_reload()’:
../../gcc/reload1.c:115:24: error: use of an operand of type ‘bool’ in ‘operator++’ is forbidden in C++17
  115 |   (this_target_reload->x_spill_indirect_levels)
      |   ~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~
../../gcc/reload1.c:470:7: note: in expansion of macro ‘spill_indirect_levels’
  470 |       spill_indirect_levels++;
      |       ^~~~~~~~~~~~~~~~~~~~~
../../gcc/reload1.c: In function ‘void elimination_effects(rtx, machine_mode)’:
../../gcc/reload1.c:3085:31: warning: this statement may fall through [-Wimplicit-fallthrough=]
 3085 |       if (code == POST_MODIFY || code == PRE_MODIFY)
      |           ~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~
../../gcc/reload1.c:3089:5: note: here
 3089 |     case STRICT_LOW_PART:
      |     ^~~~
make[3]: *** [Makefile:1066: reload1.o] Error 1

ghollisjr commented on 2020-04-30 08:56

I'm getting the same error as @asfjjv and @daniel_shub, but @8472's modified PKGBUILD and patch file do solve the problem.

daniel_shub commented on 2020-04-23 13:15

It does not build for me in a clean chroot. I get the same error as @adfjjv. It looks like there are some patches in the comments. Can someone more savvy than I update the PKGBUILD.

ecomaikgolf commented on 2020-03-22 23:52

Doesn't compile, needs to be patched