Fix clang warnings #103

proxyles · 2013-10-15T13:05:15Z

No description provided.

When the offending code is reached, new_crr is either uninitalized or have been set to NULL. This patch removes the following warning: beam/erl_alloc_util.c:3510:6: warning: variable 'new_crr' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized] if (!(flags & CFLG_FORCE_MSEG)) { ^~~~~~~~~~~~~~~~~~~~~~~~~~ beam/erl_alloc_util.c:3567:23: note: uninitialized use occurs here DEBUG_SAVE_ALIGNMENT(new_crr); ^~~~~~~ beam/erl_alloc_util.c:674:51: note: expanded from macro 'DEBUG_SAVE_ALIGNMENT' UWord algnmnt__ = sizeof(Unit_t) - (((UWord) (C)) % sizeof(Unit_t));\ ^ beam/erl_alloc_util.c:3510:2: note: remove the 'if' if its condition is always true if (!(flags & CFLG_FORCE_MSEG)) { ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ beam/erl_alloc_util.c:3438:23: note: initialize the variable 'new_crr' to silence this warning Carrier_t *new_crr, *old_crr; ^ = NULL

The test needs to be false to fail, not true. This was noticed with the following warning, where marking erl_assert_error as non-returning didn't help: beam/erl_gc.c:2772:2: warning: variable 'refc' is used uninitialized whenever switch default is taken [-Wsometimes-uninitialized] default: ^~~~~~~ beam/erl_gc.c:2775:26: note: uninitialized use occurs here ERTS_CHK_OFFHEAP_ASSERT(refc >= 1); ^~~~ beam/erl_gc.c:2738:11: note: expanded from macro 'ERTS_CHK_OFFHEAP_ASSERT' if (!(EXP)) \ ^ beam/erl_gc.c:2759:18: note: initialize the variable 'refc' to silence this warning erts_aint_t refc; ^ = 0

The warning is: sys/common/erl_poll.c:2513:72: warning: for loop has empty body [-Wempty-body] for (prev_ps = pollsets; ps != prev_ps->next; prev_ps = prev_ps->next);

The warnings are: oe_ei_encode_string.c:26:3: warning: expression result unused [-Wunused-value] (int) ei_encode_string(0,&size,p); ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~ oe_ei_encode_port.c:26:3: warning: expression result unused [-Wunused-value] (int) ei_encode_port(NULL, &size, p); ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ oe_ei_encode_atom.c:26:3: warning: expression result unused [-Wunused-value] (int) ei_encode_atom(0,&size,p); ^ ~~~~~~~~~~~~~~~~~~~~~~~~~ oe_ei_encode_pid.c:26::3: warning: expression result unused [-Wunused-value] (int) ei_encode_pid(NULL, &size, p); ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ oe_ei_encode_ref.c:26:3: warning: expression result unused [-Wunused-value] (int) ei_encode_ref(NULL, &size, p); ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ oe_ei_encode_term.c:26:3: warning: expression result unused [-Wunused-value] (int) ei_encode_term(NULL, &size, t); ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

erl_memory.c:947:7: warning: passing 'int *' to parameter of type 'socklen_t *' (aka 'unsigned int *') converts between pointers to integer types with different sign [-Wpointer-sign] &saddr_size) != 0) ^~~~~~~~~~~ /usr/include/sys/socket.h:616:74: note: passing argument to parameter here int getsockname(int, struct sockaddr * __restrict, socklen_t * __restrict) ^ erl_memory.c:2613:57: warning: passing 'int *' to parameter of type 'socklen_t *' (aka 'unsigned int *') converts between pointers to integer types with different sign [-Wpointer-sign] sock = accept(lsock, (struct sockaddr *) &oth_addr, &oth_addr_len); ^~~~~~~~~~~~~ /usr/include/sys/socket.h:610:69: note: passing argument to parameter here int accept(int, struct sockaddr * __restrict, socklen_t * __restrict) ^

nox · 2013-10-15T13:22:31Z

I just amended some commits to fix some typos, please refetch.

OTP-Maintainer · 2013-10-15T18:18:37Z

Unable to build otp: xe_events.o: No such file or directory
g++: error: x86_64-unknown-linux-gnu/wxe_init.o: No such file or directory
make[3]: *** [../priv/x86_64-unknown-linux-gnu/wxe_driver.so] Error 1
make[3]: Leaving directory /ldisk/jenkins/workspace/pullrequests/otp/lib/wx/c_src' make[2]: *** [opt] Error 2 make[2]: Leaving directory/ldisk/jenkins/workspace/pullrequests/otp/lib/wx'
make[1]: *** [opt] Error 2
make[1]: Leaving directory `/ldisk/jenkins/workspace/pullrequests/otp/lib'
make: *** [libs] Error 2

nox · 2013-10-16T00:30:50Z

Forgot to include output.mk in the wx Makefile. This is fixed now.

Btw @OTP-Maintainer you should put the build output in a code block.

rimmius · 2013-10-16T07:40:05Z

@nox I should fix that, thanks for pointing it out. Also, maybe you should try to build your patches before submitting them ? ;)

nox · 2013-10-16T08:05:05Z

@gustehn Tried it, but I messed up at one point and updated Makefile instead of Makefile.in :)

nox · 2013-10-16T08:06:53Z

Also maybe you guys should properly use silent rules when tweaking Makefiles so that I don't have to check that you did not forget to support them :D

OTP-Maintainer · 2013-10-16T20:34:46Z

Unable to build otp: xe_events.o: No such file or directory
g++: error: x86_64-unknown-linux-gnu/wxe_init.o: No such file or directory
make[3]: *** [../priv/x86_64-unknown-linux-gnu/wxe_driver.so] Error 1
make[3]: Leaving directory /ldisk/jenkins/workspace/pullrequests/otp/lib/wx/c_src' make[2]: *** [opt] Error 2 make[2]: Leaving directory/ldisk/jenkins/workspace/pullrequests/otp/lib/wx'
make[1]: *** [opt] Error 2
make[1]: Leaving directory `/ldisk/jenkins/workspace/pullrequests/otp/lib'
make: *** [libs] Error 2

nox · 2013-10-17T08:37:16Z

And this time a badly set ERL_TOP is to blame. Cloned from scratch just to be sure that it works.

@gustehn: Btw there is something weird in the pasted output: there is no xe_events.o file built by Erlang/OTP, the first character seems stripped out.

OTP-Maintainer · 2013-10-17T19:56:59Z

Unable to build otp: iameter/src'
make[3]: Entering directory /ldisk/jenkins/workspace/pullrequests/otp/lib/diameter/doc/src' Makefile:144: *** commands commence before first target. Stop. make[3]: Leaving directory/ldisk/jenkins/workspace/pullrequests/otp/lib/diameter/doc/src'
make[2]: *** [opt] Error 2
make[2]: Leaving directory /ldisk/jenkins/workspace/pullrequests/otp/lib/diameter' make[1]: *** [opt] Error 2 make[1]: Leaving directory/ldisk/jenkins/workspace/pullrequests/otp/lib'
make: *** [libs] Error 2

fenollp · 2013-10-17T21:27:49Z

Are you sure you know what you are doing here?
nox@69c5861
Your code may not be portable as http://stackoverflow.com/a/75238/1418165 suggests.
Also you replaced every usgnd_int_8 * by a char *, I would have replaced those with uint8.

Am I off?

nox · 2013-10-17T21:31:24Z

@fenollp These should contain only ASCII characters so the signedness either shouldn't be a problem or we have one already because we pass those to printf functions. Hence the warnings. Most commits in that branch change the code to silence warnings to spark the discussion about why these warnings happen in the first place.

fenollp · 2013-10-18T08:46:06Z

Ok good. On a side note there are a couple explicit casts that seem superfluous eg: https://github.com/nox/otp/blob/69c5861feca44ac261bd63a2d053c87ba5ff241c/lib/tools/c_src/erl_memory.c#L922
Was clang really complaining about those or were you a little fast at silencing those warnings?

nox · 2013-10-18T08:48:05Z

Mmh I casted that one prior to changing em_area.ptr's type, thus making it useless. Will amend.

erl_memory.c:861:19: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] *p += sprintf(*p, "%*" USGND_INT_MAX_FSTR " ", fw, mi->size); ^~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:863:16: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] *p += sprintf(*p, ^~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:869:16: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] *p += sprintf(*p, "%*" USGND_INT_MAX_FSTR " ", fw, mi->no); ^~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:871:20: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] *p += sprintf(*p, ^~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:878:16: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] *p += sprintf(*p, ^~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:900:19: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] *p += sprintf(*p, "%*" USGND_INT_MAX_FSTR " ", fw, mi->max_ever_size); ^~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:902:16: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] *p += sprintf(*p, "%*s %*s ", fw, "", fw, ""); ^~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:904:16: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] *p += sprintf(*p, "%*" USGND_INT_MAX_FSTR " ", fw, mi->max_ever_no); ^~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:906:20: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] *p += sprintf(*p, "%*s %*s ", fw, "", fw, ""); ^~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:909:16: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] *p += sprintf(*p, "%*s %*s %*s ", fw, "", fw, "", fw, ""); ^~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:986:25: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] area.size = sprintf(area.ptr, format, carg); ^~~~~~~~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:1103:25: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] area.size = sprintf(area.ptr, ^~~~~~~~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:1109:23: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] area.size += sprintf(area.ptr + area.size, ^~~~~~~~~~~~~~~~~~~~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:1120:26: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] area.size += sprintf(area.ptr + area.size, ^~~~~~~~~~~~~~~~~~~~ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:1190:18: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] p += sprintf(p, "> %-*s", EM_TIME_FIELD_WIDTH - 2, "Maximum:"); ^ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:1223:18: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] p += sprintf(p, "\n"); ^ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:1227:15: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] p += sprintf(p, "%s", stop_str); ^ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:1230:15: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] p += sprintf(p, exit_str, state->info.exit_status); ^ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:1236:18: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] p += sprintf(p, format, tsz, bw); ^ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:1286:18: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] p += sprintf(p, "%*" USGND_INT_32_FSTR " ", EM_TIME_FIELD_WIDTH - 1, secs); ^ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:1311:18: warning: passing 'usgnd_int_8 *' (aka 'unsigned char *') to parameter of type 'char *' converts between pointers to integer types with different sign [-Wpointer-sign] p += sprintf(p, "\n"); ^ /usr/include/secure/_stdio.h:49:28: note: expanded from macro 'sprintf' __builtin___sprintf_chk (str, 0, __darwin_obsz(str), __VA_ARGS__) ^ erl_memory.c:2231:32: warning: passing 'char *' to parameter of type 'usgnd_int_8 *' (aka 'unsigned char *') converts between pointers to integer types with different sign [-Wpointer-sign] size = write_header(state, state->output.header, 1); ^~~~~~~~~~~~~~~~~~~~ erl_memory.c:736:44: note: passing argument to parameter 'ptr' here write_header(em_state *state, usgnd_int_8 *ptr, int trunc) ^

Why we are assigning a pointer to pointer to char to a pointer to char baffles me. The warning is: connect/ei_resolve.c:229:11: warning: incompatible pointer types assigning to 'char *' from 'char **'; dereference with * [-Wincompatible-pointer-types] *pptr = src_aliases; ^ ~~~~~~~~~~~ *

These are not needed when not on VXWORKS or without _REENTRANT. connect/ei_resolve.c:302:24: warning: unused function 'my_gethostbyname_r' [-Wunused-function] static struct hostent *my_gethostbyname_r(const char *name, ^ connect/ei_resolve.c:360:24: warning: unused function 'my_gethostbyaddr_r' [-Wunused-function] static struct hostent *my_gethostbyaddr_r(const char *addr, ^

encode/encode_ulonglong.c:55:25: warning: comparison of unsigned expression >= 0 is always true [-Wtautological-compare] if ((p < 256) && (p >= 0)) { ~ ^ ~ legacy/erl_marshal.c:293:24: warning: comparison of unsigned expression >= 0 is always true [-Wtautological-compare] if ((ul < 256) && (ul >= 0)) { ~~ ^ ~

nox · 2013-10-18T10:26:52Z

@fenollp Amended.

nox · 2013-10-18T10:28:21Z

I wonder if the data read by emtp_parse() is only ASCII too, if it is the memory trace parser should also be patched to just use char *.

fenollp · 2013-10-18T11:15:54Z

Nice, cheers. I wonder where this should be discussed as they are not really a lot of people that hang around pull request comments. The mailing list maybe?

nox · 2013-10-18T11:33:21Z

Well, you are preaching to the choir here. I loathe these pull requests' comments. But unfortunately it seems the Erlang/OTP team is hell-bent on using them now.

OTP-Maintainer · 2013-10-18T21:29:07Z

Patch has passed first testings and has been assigned to be reviewed

psyeugenic · 2013-10-21T10:23:41Z

We're using Pull Request because the community, i.e. you, asked for it. =) I rather like it because thoughts and code are located at the same place. We should also get rid of the ML so we don't split the information flow imho.

This patch has a lot of changes on very varying places in the code, i.e. different maintainers. Most of the changes seems to be in debug enabled code only, i.e. low priority.

I will get back to you with feedback once I've goon through it properly.

nox · 2013-10-21T10:32:16Z

Please note that the third commit is not just warning silencing, it's an actual bug. Most of the changes is not in debug-enabled code, no.

Otherwise while pull requests are good in theory, @OTP-Maintainer gives a good example of where it sucks big time, you have to take care of formatting your emails etc, and Markdown is underspecified. I would rather have plain text emails which I can browse offline. I don't even have my own comments received by email when using pull requests, that's splitting information.

We gave up on standard email protocols to use a proprietary forum, this sucks.

And I have never asked for pull requests.

nox · 2013-10-21T10:33:18Z

Only the first three commits and the last one change debug-only code.

nox · 2013-10-21T10:35:03Z

Also, while someone can't delete an email that I have already received, anyone can delete a comment he made on GitHub. I fail to see that as an improvement.

And people can also amend their comments whenever they want, this is insane.

psyeugenic · 2013-10-21T10:41:36Z

You should voice your opinion on Erlang Questions then. We certainly want feedback. GitHub has shortcomings and ML's certainly have also.

nox · 2013-10-21T10:56:12Z

Done. Now I am bracing for impact.

proxyles · 2014-01-23T14:14:42Z

Graduated

nox added 6 commits October 15, 2013 15:21

Properly mark erl_assert_error as non-returning

bb4d0cd

Silence a warning in erl_poll about an empty loop body

fad8c6e

The warning is: sys/common/erl_poll.c:2513:72: warning: for loop has empty body [-Wempty-body] for (prev_ps = pollsets; ps != prev_ps->next; prev_ps = prev_ps->next);

nox added 6 commits October 18, 2013 12:19

Compile in_heapfrag() only in debug mode

e308ec0

Fix some silent rules in diameter, eunit and wx

0817771

proxyles closed this Jan 23, 2014

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Fix clang warnings #103

Fix clang warnings #103

proxyles commented Oct 15, 2013

nox commented Oct 15, 2013

OTP-Maintainer commented Oct 15, 2013

nox commented Oct 16, 2013

rimmius commented Oct 16, 2013

nox commented Oct 16, 2013

nox commented Oct 16, 2013

OTP-Maintainer commented Oct 16, 2013

nox commented Oct 17, 2013

OTP-Maintainer commented Oct 17, 2013

fenollp commented Oct 17, 2013

nox commented Oct 17, 2013

fenollp commented Oct 18, 2013

nox commented Oct 18, 2013

nox commented Oct 18, 2013

nox commented Oct 18, 2013

fenollp commented Oct 18, 2013

nox commented Oct 18, 2013

OTP-Maintainer commented Oct 18, 2013

psyeugenic commented Oct 21, 2013

nox commented Oct 21, 2013

nox commented Oct 21, 2013

nox commented Oct 21, 2013

psyeugenic commented Oct 21, 2013

nox commented Oct 21, 2013

proxyles commented Jan 23, 2014

Fix clang warnings #103

Fix clang warnings #103

Conversation

proxyles commented Oct 15, 2013

nox commented Oct 15, 2013

OTP-Maintainer commented Oct 15, 2013

nox commented Oct 16, 2013

rimmius commented Oct 16, 2013

nox commented Oct 16, 2013

nox commented Oct 16, 2013

OTP-Maintainer commented Oct 16, 2013

nox commented Oct 17, 2013

OTP-Maintainer commented Oct 17, 2013

fenollp commented Oct 17, 2013

nox commented Oct 17, 2013

fenollp commented Oct 18, 2013

nox commented Oct 18, 2013

nox commented Oct 18, 2013

nox commented Oct 18, 2013

fenollp commented Oct 18, 2013

nox commented Oct 18, 2013

OTP-Maintainer commented Oct 18, 2013

psyeugenic commented Oct 21, 2013

nox commented Oct 21, 2013

nox commented Oct 21, 2013

nox commented Oct 21, 2013

psyeugenic commented Oct 21, 2013

nox commented Oct 21, 2013

proxyles commented Jan 23, 2014