Mar2018 production compilation fails on Fedora 28
Compiling from Mar2018 branch, today's hg tip, with production settings: --disable-debug --disable-developer --disable-assert --enable-optimize Fails with this on Fedora 28: In file included from /usr/include/string.h:494, from /home/roberto/MonetDBServer/MonetDB.Spinque_Mar2018/src/gdk/gdk.h:316, from /home/roberto/MonetDBServer/MonetDB.Spinque_Mar2018/src/gdk/gdk_calc.c:10: In function ‘memcpy’, inlined from ‘addstr_loop’ at /home/roberto/MonetDBServer/MonetDB.Spinque_Mar2018/src/gdk/gdk_calc.c:3415:3: /usr/include/bits/string_fortified.h:34:10: error: ‘__builtin_memcpy’ forming offset [3, 65535] is out of the bounds [0, 2] of object ‘str_nil’ with type ‘const char[2]’ [-Werror=array-bounds] return __builtin___memcpy_chk (__dest, __src, __len, __bos0 (__dest)); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ (same error in several files) Compiling with debug settings succeeds: --enable-debug --enable-developer --enable-assert --disable-optimize Is this reproducible or has it something to do with my machine?
indeed reproducible on my Fedora 28 desktop in "production" mode (not in "development mode", though) with both parent: 66690:e96bb94b4c9c Mar2018_7 Mar2018_SP1_release Moved contents of ChangeLog.Mar2018 to MonetDB.spec, debian/changelog and ChangeLog-Archive. branch: Mar2018 and parent: 66829:557dfa22af73 tip make sure we use the local table of the replica table if possible branch: Mar2018 but not with parent: 66821:0f951a98884e Merge with Mar2018 branch. branch: default (the latter compiles without problems, both in "development" and "production" mode) Best, Stefan ----- On Jul 18, 2018, at 5:54 PM, Roberto Cornacchia roberto.cornacchia@gmail.com wrote:
Compiling from Mar2018 branch, today's hg tip, with production settings:
--disable-debug --disable-developer --disable-assert --enable-optimize
Fails with this on Fedora 28:
In file included from /usr/include/string.h:494, from /home/roberto/MonetDBServer/MonetDB.Spinque_Mar2018/src/gdk/gdk.h:316, from /home/roberto/MonetDBServer/MonetDB.Spinque_Mar2018/src/gdk/gdk_calc.c:10: In function ‘memcpy’, inlined from ‘addstr_loop’ at /home/roberto/MonetDBServer/MonetDB.Spinque_Mar2018/src/gdk/gdk_calc.c:3415:3: /usr/include/bits/string_fortified.h:34:10: error: ‘__builtin_memcpy’ forming offset [3, 65535] is out of the bounds [0, 2] of object ‘str_nil’ with type ‘const char[2]’ [-Werror=array-bounds] return __builtin___memcpy_chk (__dest, __src, __len, __bos0 (__dest)); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
(same error in several files)
Compiling with debug settings succeeds: --enable-debug --enable-developer --enable-assert --disable-optimize
Is this reproducible or has it something to do with my machine?
_______________________________________________ users-list mailing list users-list@monetdb.org https://www.monetdb.org/mailman/listinfo/users-list
-- | Stefan.Manegold@CWI.nl | DB Architectures (DA) | | www.CWI.nl/~manegold/ | Science Park 123 (L321) | | +31 (0)20 592-4212 | 1098 XG Amsterdam (NL) |
GCC 8 has added checks that fail when compiling the Mar2018 branch of MonetDB with optimizations enabled. They are in fact false positives (code that will never be executed), but the compiler doesn't know that. The problem has been fixed in the default branch. For now, use --disable-strict when configuring on Fedora 28. On 18/07/18 18:28, Stefan Manegold wrote:
indeed reproducible on my Fedora 28 desktop in "production" mode (not in "development mode", though) with both
parent: 66690:e96bb94b4c9c Mar2018_7 Mar2018_SP1_release Moved contents of ChangeLog.Mar2018 to MonetDB.spec, debian/changelog and ChangeLog-Archive. branch: Mar2018
and
parent: 66829:557dfa22af73 tip make sure we use the local table of the replica table if possible branch: Mar2018
but not with
parent: 66821:0f951a98884e Merge with Mar2018 branch. branch: default
(the latter compiles without problems, both in "development" and "production" mode)
Best, Stefan
----- On Jul 18, 2018, at 5:54 PM, Roberto Cornacchia roberto.cornacchia@gmail.com wrote:
Compiling from Mar2018 branch, today's hg tip, with production settings:
--disable-debug --disable-developer --disable-assert --enable-optimize
Fails with this on Fedora 28:
In file included from /usr/include/string.h:494, from /home/roberto/MonetDBServer/MonetDB.Spinque_Mar2018/src/gdk/gdk.h:316, from /home/roberto/MonetDBServer/MonetDB.Spinque_Mar2018/src/gdk/gdk_calc.c:10: In function ‘memcpy’, inlined from ‘addstr_loop’ at /home/roberto/MonetDBServer/MonetDB.Spinque_Mar2018/src/gdk/gdk_calc.c:3415:3: /usr/include/bits/string_fortified.h:34:10: error: ‘__builtin_memcpy’ forming offset [3, 65535] is out of the bounds [0, 2] of object ‘str_nil’ with type ‘const char[2]’ [-Werror=array-bounds] return __builtin___memcpy_chk (__dest, __src, __len, __bos0 (__dest)); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
(same error in several files)
Compiling with debug settings succeeds: --enable-debug --enable-developer --enable-assert --disable-optimize
Is this reproducible or has it something to do with my machine?
_______________________________________________ users-list mailing list users-list@monetdb.org https://www.monetdb.org/mailman/listinfo/users-list
-- Sjoerd Mullender
participants (3)
-
Roberto Cornacchia -
Sjoerd Mullender -
Stefan Manegold