This is GCC Bugzilla
This is GCC Bugzilla Version 2.20+
View Bug Activity | Format For Printing | Clone This Bug
A "gcc -Wextra -O" misses the 4 bugs in this test program: int main (int argc, char *argv[]) { if (argc != 1 || argc != 2) return 1; if (argc < 0 && argc > 10) return 1; if (argc || !argc) return 1; if (argc && !argc) return 1; return 0; } With a tweak to fold-const.c (below), they will be caught: t.c:4: warning: `or' of collectively exhaustive tests is always 1 t.c:5: warning: `and' of mutually exclusive tests is always 0 t.c:6: warning: `or' of collectively exhaustive tests is always 1 t.c:7: warning: `and' of mutually exclusive tests is always 0 A bootstrap of gcc-3.5-20040627 reports such a warning in fold-const.c itself, the patch below includes the obvious fix. ==== Comment: The -O is needed due to this code in fold-const.c: /* We only do these simplifications if we are optimizing. */ if (!optimize) return t; This is probably from when fold-const.c was much simpler, but I think is no longer appropriate and recommend it be deleted. ==== *** fold-const.c.orig Sun Jun 27 11:23:44 2004 --- fold-const.c Wed Jun 30 12:43:39 2004 *************** *** 4377,4379 **** in_p, low, high)))) ! return or_op ? invert_truthvalue (tem) : tem; --- 4377,4388 ---- in_p, low, high)))) ! { ! if (TREE_CODE (tem) == INTEGER_CST && extra_warnings) ! { ! if (or_op) ! warning ("`or' of collectively exhaustive tests is always 1"); ! else ! warning ("`and' of mutually exclusive tests is always 0"); ! } ! return or_op ? invert_truthvalue (tem) : tem; ! } *************** *** 9998,10000 **** || (TREE_CODE (op0) == REAL_CST ! && TREE_CODE (op0) != REAL_CST)) { --- 10007,10009 ---- || (TREE_CODE (op0) == REAL_CST ! && TREE_CODE (op1) != REAL_CST)) {
Confirmed, patches goto gcc-patches@ after reading http://gcc.gnu.org/contribute.html.
It is not practical for gcc outsiders to submit patches, the requirements are too exacting and complex. E.g. see http://gcc.gnu.org/ml/gcc-patches/2004-07/msg00072.html which demands a testcase (something which is IMO unclear in the "contributing" URL you gave). It is unlikely my first (or second) attempt at a testcase will be flawless. Similarly, I am asked to document this warning (which I find odd, since there are lots of undocumented -Wextra warnings). Do I need to fix the "broken" existing warnings too? The response was unclear. The only requirement I can confidently handle is fixing the warning strings. And are these *all* the requirements? I doubt it. On the other hand, such issues are trivial for veterans such as yourself. Could you please handle this for me? I would greatly appreciate it.
Subject: Re: gcc fails to warn about some common logic errors On Thu, 1 Jul 2004, trt at acm dot org wrote: > (which I find odd, since there are lots of undocumented -Wextra warnings). Please report these as separate bugs. Phil Edwards fixed the known such problems in <http://gcc.gnu.org/ml/gcc-patches/2003-01/msg01741.html>; any that have been added since then would be regressions.
I have submitted a revised patch in http://gcc.gnu.org/ml/gcc-patches/2004-07/msg00148.html
Confirmed.
Here are some reasons why I think this patch should be applied. 1. Inspection of the patch will show that it is "low risk", except for the potential risk of false positives. 2. I have been using this warning for about 4 years now, on a 35Mloc source code base involving 100s of developers, and it has triggered dozens of times with only a couple false positives. (The false positives do not happen with gcc 3.5, I think some changes to fold() have caused them to go away.) Sometimes, for fun, I use this compiler on other source code. E.g. I remember reporting two of these warnings (among others) in an early version of "valgrind". 3. It spots 3 bugs in gcc 3.5. One of them is fixed by this patch. If you want to see the other two, try applying the patch. As a downside, some versions of gcc/insn-attrtab.c trigger these warnings. Last I checked, the ia64 version had about 7 of them. They are technically valid, but probably indicate harmless inefficiencies. On the other hand, they might indicate something that should be cleaned up. I realize there is no urgent need for this patch, but I have seen much less effective (and sometimes counterproductive) warning messages added to gcc. Please consider one.
The patch given in comment #4 no successfully applies because the warning() function has new initial parameter. Also the fold_buildN() cleanup invalidated (and fixed) the "REAL_CST" part of the patch. Some "speculative folding" is now causing false positives, for which there is a simple fix: suppress warnings when doing such folds. For example *** tree-ssa-loop-niter.c.orig Wed Apr 27 13:48:21 2005 --- tree-ssa-loop-niter.c Wed Apr 27 14:28:59 2005 *************** *** 778,780 **** --- 778,782 ---- notcond = invert_truthvalue (cond); + inhibit_warnings++; e = fold_build2 (TRUTH_OR_EXPR, boolean_type_node, notcond, te); + inhibit_warnings--; if (nonzero_p (e)) People who think warnings belong only in the front-end might balk. I could submit an alternate indirect check for this situation in parser_build_binary_op. It would be tidy, but would no longer warn about non-C glitches such as: libjava/gnu/java/security/x509/X500DistinguishedName.java:447: if (sep != '+' || sep != ',') libjava/java/net/HttpURLConnection.java:555: if (((code / 100) != 4) || ((code / 100) != 5)) libjava/javax/swing/plaf/basic/BasicGraphicsUtils.java:401: if ((underlinedChar >= 0) || (underlinedChar <= 0xffff)) libjava/javax/security/auth/x500/X500Principal.java:380: if (sep != '+' || sep != ',') Please let me know if there is any interest in a C-only patch. (Or any interest in pursuing this PR.)
I wish to answer this because I am also a newbie and I understand your frustration. On the other hand, I also understand why the current situation is as it is now. (In reply to comment #2) > It is not practical for gcc outsiders to submit patches, > the requirements are too exacting and complex. E.g. see > http://gcc.gnu.org/ml/gcc-patches/2004-07/msg00072.html > which demands a testcase (something which is IMO unclear > in the "contributing" URL you gave). The requirements are documented. Unfortunately, the documentation is divided among many places and not in the best form. It would be superb and very welcome if someone took the responsibility of summarizing, re-writting and improving the different pieces, so they can be as clear and simple as possible. (Like a step by step guide on how to contribute, maybe especialised for different user-cases). Sadly, experienced people don't have time (and perhaps are not very well suited) for this job. Someone brave needs to step up and take the burden, perhaps seek collaborators among other frustrated newbies. Another problem is that as soon as a newbie becomes experienced with the procedure, then the interest in improving the docs fades away :( > It is unlikely my first (or second) attempt at a testcase > will be flawless. It doesn't matter, reviewers should catch any flaw and patches are regularly submitted many times. It is part of the learning proccess. Flawless patches are accepted earlier, though. ;-) > Similarly, I am asked to document this warning > (which I find odd, since there are lots of undocumented -Wextra warnings). > Do I need to fix the "broken" existing warnings too? The response was unclear. That is not an excuse. There are things that are historically wrong. New mistakes cannot be justified based on past mistakes. New warnings should be documented. Normally, you don't need to fix anything that is broken already as long as it is not closely related to your patch. If the response was unclear, you have to insist. No all mails are neccessarily read by all people and people are busy busy busy (or on holidays), so sometimes your mail will get lost. Wait one week and insist again. Wait another week and keep insisting. Eventually someone will answer. Short and polite emails get more answers than angry, long rants. This also applies to patches. If you patch is not reviewed: wait one week and insist. Repeat until reviewed. > The only requirement I can confidently handle is fixing the warning strings. > And are these *all* the requirements? I doubt it. It is very likely that if a reviewer finds an obvious mistake in your patch, he/she will not review in deep the rest of it and just move on to the next patch. You will need to send it again and wait for the answer. At first you will make a lot of mistakes but that is normal. My first patch was rejected several times just for the Changelog format. In a few iterations your patches will be almost perfect. > On the other hand, such issues are trivial for veterans such as yourself. > Could you please handle this for me? I would greatly appreciate it. That is a sad misconception. Bootstrapping + testing take time even if there are no failures at all. Adding testcases takes time. Having patches around takes thinking time. Keeping track of submitted patches and pinging reviewers take time. Fixing minor issues in patches takes time. Committing patches takes also a little amount of time. Of course, the total time could be trivial if you had nothing else to do. But if the issue is minor and you are busy (or you are not paid for fixing it and just prefer to invest your time in something else like your family), then it is already too much time. Also there is the issue (I guess) that if sloppy patches are regularly fixed by reviewers, that will advocate more sloppy patches which in turn will make reviewers more unwilling to review patches. So please, reconsiderate your stance, and resubmit your patch. I am sure your contribution will be appreciated. Thanks in advance.
Subject: Bug 16302 Author: manu Date: Fri May 15 20:08:21 2009 New Revision: 147596 URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=147596 Log: 2009-05-15 Manuel López-Ibáñez <manu@gcc.gnu.org> PR 16302 * fold-const.c (make_range,build_range_check,merge_ranges): Move declaration to... (merge_ranges): Returns bool. * tree.h (make_range): .. to here. (build_range_check): Likewise. (merge_ranges): Likewise. Renamed from merge_ranges. * c-typeck.c (parser_build_binary_op): Update calls to warn_logical_operator. * c-common.c (warn_logical_operator): Add new warning. * c-common.h (warn_logical_operator): Update declaration. cp/ * call.c (build_new_op): Update calls to warn_logical_operator. testsuite/ * gcc.dg/pr16302.c: New. * g++.dg/warn/pr16302.C: New. Added: trunk/gcc/testsuite/g++.dg/warn/pr16302.C trunk/gcc/testsuite/gcc.dg/pr16302.c Modified: trunk/gcc/ChangeLog trunk/gcc/c-common.c trunk/gcc/c-common.h trunk/gcc/c-typeck.c trunk/gcc/cp/ChangeLog trunk/gcc/cp/call.c trunk/gcc/fold-const.c trunk/gcc/testsuite/ChangeLog trunk/gcc/tree.h
FIXED for GCC 4.5