From mboxrd@z Thu Jan 1 00:00:00 1970 X-Msuck: nntp://news.gmane.org/gmane.linux.lib.musl.general/290 Path: news.gmane.org!not-for-mail From: =?UTF-8?B?THVrYSBNYXLEjWV0acSH?= Newsgroups: gmane.linux.lib.musl.general Subject: Re: Simple testing task - string functions Date: Mon, 18 Apr 2011 14:20:31 +0200 Message-ID: <4DAC2C8F.8060102@gmail.com> References: <20110410044515.GB13185@brightrain.aerifal.cx> <4DA73605.9080107@gmail.com> <20110414231140.GO13185@brightrain.aerifal.cx> NNTP-Posting-Host: lo.gmane.org Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit X-Trace: dough.gmane.org 1312595706 11468 80.91.229.12 (6 Aug 2011 01:55:06 GMT) X-Complaints-To: usenet@dough.gmane.org NNTP-Posting-Date: Sat, 6 Aug 2011 01:55:06 +0000 (UTC) To: musl@lists.openwall.com Original-X-From: envelope-from@hidden Mon Apr 18 16:23:44 2011 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; DomainKey-Signature: a=rsa-sha1; c=nofws; User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.15) Gecko/20110402 Icedove/3.1.9 In-Reply-To: <20110414231140.GO13185@brightrain.aerifal.cx> Xref: news.gmane.org gmane.linux.lib.musl.general:290 Archived-At: On 04/15/2011 01:11 AM, Rich Felker wrote: > On Thu, Apr 14, 2011 at 07:59:33PM +0200, Luka Marčetić wrote: >> Hello again. >> Attached is the solution to the task. The program seems to preform >> as expected, but may still need double-checking. >> The tests therein fail where expected when linked with various old >> versions of musl. Note that although the program is designed to >> allow tests to fail gracefully as suggested, this does not happen >> due to bugs in function implementations in said old versions that >> the program depends on. Rich and Chris have confirmed a bug in 0.7.6 >> that causes a segfault in the siglongjmp (longjmp to be exact). I'm >> still waiting for confirmation of inability of version 0.7.5 to >> dispose the same signal to a specified handler properly. > Here's an idea for avoiding the sigsetjmp/siglongjmp bug: just use > setjmp/longjmp, and use sigaction with SA_NODEFER to install the > signal handler. Actually sigaction should always be used to install > signal handlers, since signal is under-specified and you can't be sure > how it will behave. > > Rich Hey. I've been away (I'll explain why in #musl). Anyway, thanks for the critique. I'll take your word and use sigaction further on. I may correct these tests too for GSoC if you think they'll prove useful. Anyway, we'll talk. It's now up to you guys to accept/reject my proposal if I'm not mistaken. Thanks -Luka