home bbs files messages ]

Forums before death by AOL, social media and spammers... "We can't have nice things"

   linux.debian.bugs.dist      Ohh some weird Debian bug report thing      28,835 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 28,168 of 28,835   
   Emanuele Rocca to All   
   Bug#1128400: dovecot: FTBFS with glibc 2   
   19 Feb 26 11:20:01   
   
   XPost: linux.debian.ports.arm   
   From: ema@debian.org   
      
   Source: dovecot   
   Version: 1:2.4.2+dfsg1-3   
   Severity: important   
   User: debian-arm@lists.debian.org   
   Usertags: arm64   
   X-Debbugs-Cc: debian-arm@lists.debian.org   
      
   Hi,   
      
   dovecot fails to build from source on arm64 when using glibc 2.43,   
   currently in experimental.   
      
   The build gets stuck in a endless loop. The failed assert below was   
   logged several million times on the system where I was building the   
   package when I noticed the multi-GB build log and killed the build:   
      
      make[4]: Entering directory '/build/reproducible-path/dovecot   
   2.4.2+dfsg1/src/lib-ssl-iostream'   
      for bin in test-iostream-ssl; do \   
        if !  ./$bin; then exit 1; fi; \   
      done   
      ssl: handshake ....................................................... : ok   
      test-iostream-ssl.c:448: Assert failed: o_stream_send(client->output, buf,   
   avail) == (ssize_t)avail   
      test-iostream-ssl.c:105: Assert failed: i_stream_read_bytes(ep->input,   
   &data, &size, 1) > -1 || ep->input->stream_errno == 0   
      test-iostream-ssl.c:105: Assert failed: i_stream_read_bytes(ep->input,   
   &data, &size, 1) > -1 || ep->input->stream_errno == 0   
      test-iostream-ssl.c:105: Assert failed: i_stream_read_bytes(ep->input,   
   &data, &size, 1) > -1 || ep->input->stream_errno == 0   
      [EMA: SNIP!]   
      
   Logs here: https://people.debian.org/~ema/glibc-2.43-rebuilds/ou   
   put-1/dovecot_arm64.build   
      
   The issue is triggered by glibc 2.43 on arm64 enabling 2MB THP by   
   default:   
   https://sourceware.org/git/?p=glibc.git;a=commit;h=321e1fc73f530   
   1d92ba357cdd48c56b79292020   
      
   The problem is however not architecture-specific. For example it can be   
   reproduced with glibc 2.42 on amd64 by building the package with the   
   GLIBC_TUNABLES environment variable set to glibc.malloc.hugetlb=2.   
      
   Cheers,   
     ema   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca