X-Original-To: alpine-devel@lists.alpinelinux.org Received: from mail.ovgu.de (mail.ovgu.de [141.44.1.66]) by lists.alpinelinux.org (Postfix) with ESMTP id 3BCD45C4C9B for ; Mon, 28 Aug 2017 08:34:00 +0000 (GMT) Received: from mail.ovgu.de (localhost [127.0.0.1]) by localhost (Postfix) with SMTP id 1465740065; Mon, 28 Aug 2017 10:33:59 +0200 (CEST) Received: from notebook-marian (euk-gateway.cs.uni-magdeburg.de [141.44.25.67]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.ovgu.de (Postfix) with ESMTPSA id C177F40062; Mon, 28 Aug 2017 10:33:58 +0200 (CEST) Date: Mon, 28 Aug 2017 10:33:58 +0200 From: Marian Buschsieweke To: Timo Teras Cc: alpine-dev , Natanael Copa Subject: [alpine-devel] Re: Cross-compilers [was: Re: [alpine-aports] [PATCH 2/3] testing/gcc-arm-none-eabi: New aport] Message-ID: <20170828103358.5a82363f@notebook-marian> In-Reply-To: <20170828100950.636fd65a@vostro.util.wtbts.net> References: <20170826140729.2822-1-marian.buschsieweke@ovgu.de> <20170826140729.2822-2-marian.buschsieweke@ovgu.de> <20170828100950.636fd65a@vostro.util.wtbts.net> Organization: =?UTF-8?B?T3R0by12b24tR3Vlcmlja2UtVW5pdmVyc2l0w6R0?= Magdeburg X-Mailer: Claws Mail 3.15.0-dirty (GTK+ 2.24.31; x86_64-alpine-linux-musl) X-Mailinglist: alpine-devel Precedence: list List-Id: Alpine Development List-Unsubscribe: List-Post: List-Help: List-Subscribe: MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-PMX-Version: 6.3.3.2656215, Antispam-Engine: 2.7.2.2107409, Antispam-Data: 2017.8.28.82416, AntiVirus-Engine: 5.40.0, AntiVirus-Data: 2017.8.28.5400001 X-PMX-Spam: Gauge=IIIIIIII, Probability=8%, Report=' MULTIPLE_RCPTS 0.1, HTML_00_01 0.05, HTML_00_10 0.05, BODY_ENDS_IN_URL 0, BODY_SIZE_3000_3999 0, BODY_SIZE_5000_LESS 0, BODY_SIZE_7000_LESS 0, INVALID_MSGID_NO_FQDN 0, IN_REP_TO 0, LEGITIMATE_SIGNS 0, MSG_THREAD 0, MULTIPLE_REAL_RCPTS 0, NO_URI_HTTPS 0, REFERENCES 0, SINGLE_URI_IN_BODY 0, __ANY_URI 0, __BOUNCE_CHALLENGE_SUBJ 0, __BOUNCE_NDR_SUBJ_EXEMPT 0, __C230066_P1_5 0, __CC_NAME 0, __CC_NAME_DIFF_FROM_ACC 0, __CC_REAL_NAMES 0, __CP_URI_IN_BODY 0, __CT 0, __CTE 0, __CT_TEXT_PLAIN 0, __FORWARDED_MSG 0, __HAS_CC_HDR 0, __HAS_FROM 0, __HAS_MSGID 0, __HAS_X_MAILER 0, __IN_REP_TO 0, __MIME_TEXT_ONLY 0, __MIME_TEXT_P 0, __MIME_TEXT_P1 0, __MIME_VERSION 0, __MULTIPLE_RCPTS_CC_X2 0, __NO_HTML_TAG_RAW 0, __REFERENCES 0, __SANE_MSGID 0, __SINGLE_URI_TEXT 0, __SUBJ_ALPHA_NEGATE 0, __TO_MALFORMED_2 0, __TO_NAME 0, __TO_NAME_DIFF_FROM_ACC 0, __TO_REAL_NAMES 0, __URI_IN_BODY 0, __URI_NOT_IMG 0, __URI_NO_PATH 0, __URI_NO_WWW 0, __URI_NS , __URI_WITHOUT_PATH 0' X-Spam-Score: Gauge=IIIIIIII X-PMX-consideredAsSpam: no Hi, I need a ARM cross compiler for my work as PHD student, so I will maintain the APKBUILDs for cross compiler toolchain anyway. Thus, it would be little afford for me to keep seperate APKBUILDs in testing up to date until cross toolchains can be created from a single APKBUILD. Btw: The APKBUILDs were created to be reuseable for different bare metal targets, simply changing the "_target" variable to "xtensa-lx106-elf" would create a bare metal toolchain for ESP8266, sadly newlib does not provide support for the xtensa ISA in the official source. I bet changing "_target" to avr would work just as well. > Also, arm-none-eabi is also targetting bare-metal. So I wonder if this > should be separately added anyway. @ncopa, any thoughts if this > should/could reference the main 'gcc' or be separate like suggested in > the patch? I'd like to suggest to keep toolchains targetting bare metal architecture in a separate APKBUILD from those targetting a full OS. E.g. java, go, Fortran, objective C, Ada and (at least in my opinion, others see this different) C++ is not really interesting for bare metal devices. Most languages (except C++, Ada and Fortran) will not produce machine code that could be run on bare metal devices and all but C++ will be a huge pain to even get packaged (if possible at all). Kind regards, Marian PS: In order to fuse arm embedded with AVR, enabling C++ would be required, as a lot of users prefer Arduino to develop for the AVR platform. Maybe I should add C++ before the APKBUILD gets accepted, if you choose to accept separate APKBUILDs for different platforms for now. On Mon, 28 Aug 2017 10:09:50 +0300 Timo Teras wrote: > Hi, > > We've occasionally received requests for cross-compilers. And in fact > we do keep the *-avr variant already because we can't have native > building for it really nor run alpine on these microcontrollers. > > Now ARM embedded is the second most popular, and many would benefit > from a cross compiler, probably me including. > > I was earlier hopeful on getting an 'abuild build flavors' feature > where abuild could be used to automatically build several different > flavors of same APKBUILD - in cross-compiler case involving building > binutils & gcc for several different CTARGETs. > > But as this feature has not been implemented yet, I'm writing to check > if anyone is working on it, or willing to work on it very soon. Or > should we open up to accept cross-compilers as separate APKBUILDs until > this feature is done? > > Also, arm-none-eabi is also targetting bare-metal. So I wonder if this > should be separately added anyway. @ncopa, any thoughts if this > should/could reference the main 'gcc' or be separate like suggested in > the patch? > > Thanks > Timo > > > On Sat, 26 Aug 2017 16:07:28 +0200 > Marian Buschsieweke wrote: > > > The GNU Compiler Collection for embedded ARM targets (e.g. Cortex > > M0/M3/M4/...) http://gcc.gnu.org --- Unsubscribe: alpine-devel+unsubscribe@lists.alpinelinux.org Help: alpine-devel+help@lists.alpinelinux.org ---