You are not logged in.

#1 2026-09-14 13:41:35

trakno
Member
Registered: 2019-11-05
Posts: 9

AUR guacamole-server 1.6.0-2

Hello,
Was unable to install this package, because warning are parsed as error during the compilation.
I had to download (clone) guacamole-server.git and modify the configuration line in the PKGBUILD file by adding `CFLAGS="$CFLAGS -Wno-error"` to get the installation to succeed.
Is it good as workaround? is there any better option? (was the first aur package with a problem since i use Archlinux - i didn't know what to do instead )


Rules changes with rolling updates !

Offline

#2 2026-09-14 14:51:36

seth
Member
From: Won't reply 2 private help req
Registered: 2012-09-03
Posts: 77,988

Re: AUR guacamole-server 1.6.0-2

Adjust the PKGBUILD, https://aur.archlinux.org/cgit/aur.git/ … server#n47 or https://aur.archlinux.org/cgit/aur.git/ … er-git#n50
Alternatively file a bug upstream but depending on the error they might just tell you to ignore it (recent gcc introduced kind of a pickle where c++20 will warn a lambda construct to be deprecated but c++17 requires that very construct - if you want to support both versions, you either add preprocessor branches or just look the other side…)

Online

#3 2026-09-14 15:43:44

loqs
Member
Registered: 2014-03-06
Posts: 19,087

Re: AUR guacamole-server 1.6.0-2

@seth different issues?
https://github.com/apache/guacamole-ser … ure.ac#L48

AC_DEFINE([_XOPEN_SOURCE], [700], [Uses X/Open and POSIX APIs])

From build log:

In file included from raw_encoder.h:23,
                 from audio.c:28:
../../config.h:290:9: error: '_XOPEN_SOURCE' redefined [-Werror]
  290 | #define _XOPEN_SOURCE 700
      |         ^~~~~~~~~~~~~
In file included from /usr/include/unistd.h:25,
                 from guacamole/socket-fntypes.h:31,
                 from guacamole/socket.h:31,
                 from guacamole/client-fntypes.h:33,
                 from guacamole/client.h:29,
                 from audio.c:24:
/usr/include/features.h:234:10: note: this is the location of the previous definition
  234 | # define _XOPEN_SOURCE  800
      |          ^~~~~~~~~~~~~
In file included from raw_encoder.h:23,
                 from raw_encoder.c:28:
../../config.h:290:9: error: '_XOPEN_SOURCE' redefined [-Werror]
  290 | #define _XOPEN_SOURCE 700
      |         ^~~~~~~~~~~~~
In file included from /usr/include/unistd.h:25,
                 from guacamole/socket-fntypes.h:31,
                 from guacamole/socket.h:31,
                 from guacamole/client-fntypes.h:33,
                 from guacamole/client.h:29,
                 from raw_encoder.c:24:
/usr/include/features.h:234:10: note: this is the location of the previous definition
  234 | # define _XOPEN_SOURCE  800
      |          ^~~~~~~~~~~~~
In file included from encode-png.h:23,
                 from client.c:23:
../../config.h:290:9: error: '_XOPEN_SOURCE' redefined [-Werror]
  290 | #define _XOPEN_SOURCE 700
      |         ^~~~~~~~~~~~~
In file included from /usr/include/unistd.h:25,
                 from guacamole/socket-fntypes.h:31,
                 from guacamole/socket.h:31,
                 from encode-jpeg.h:25,
                 from client.c:22:
/usr/include/features.h:234:10: note: this is the location of the previous definition
  234 | # define _XOPEN_SOURCE  800
      |          ^~~~~~~~~~~~~
cc1: all warnings being treated as errors

Edit:
https://github.com/apache/guacamole-ser … c233d25670
For next issue:
https://github.com/apache/guacamole-ser … df339280f3
For next issue:
https://github.com/apache/guacamole-ser … c41441116d which does not apply cleanly and introduces other issues.
Edit2:
This may be fixable with patching but adding the flag or cutting the flag from configure.ac and avoiding the wack a mole is much simpler.

Last edited by loqs (2026-09-14 16:00:54)

Offline

#4 2026-09-14 15:56:00

seth
Member
From: Won't reply 2 private help req
Registered: 2012-09-03
Posts: 77,988

Re: AUR guacamole-server 1.6.0-2

I just meant to provide an example why upstream might refuse to do sth. about the warning.
Not sure why the code wants to (re)define that but it will have to check (#ifndef) or #undef first.
That's not a fatal bug but certainly unexpected behavior, so upstream fault.

Online

Board footer

Powered by FluxBB