From mboxrd@z Thu Jan 1 00:00:00 1970 X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me X-Spam-Level: X-Spam-Status: No, score=-1.3 required=5.0 tests=BAYES_00,INVALID_MSGID autolearn=no autolearn_force=no version=3.4.4 X-Google-Language: ENGLISH,ASCII-7-bit X-Google-Thread: fac41,9a0ff0bffdf63657 X-Google-Attributes: gidfac41,public X-Google-Thread: 1108a1,9a0ff0bffdf63657 X-Google-Attributes: gid1108a1,public X-Google-Thread: 103376,4b06f8f15f01a568 X-Google-Attributes: gid103376,public X-Google-Thread: f43e6,9a0ff0bffdf63657 X-Google-Attributes: gidf43e6,public From: Gerhard Menzl Subject: Re: Software landmines (loops) Date: 1998/09/01 Message-ID: <35EB9ECE.6737310@sea.ericsson.se>#1/1 X-Deja-AN: 386705849 Content-Transfer-Encoding: 7bit References: <902934874.2099.0.nnrp-10.c246a717@news.demon.co.uk> <6r1glm$bvh$1@nnrp1.dejanews.com> <6r9f8h$jtm$1@nnrp1.dejanews.com> <6renh8$ga7$1@nnrp1.dejanews.com> <6rf59b$2ud$1@nnrp1.dejanews.com> <6rfra4$rul$1@nnrp1.dejanews.com> <35DBDD24.D003404D@calfp.co.uk> <6sbuod$fra$1@hirame.wwa.com> <35f51e53.48044143@ <904556531.666222@miso.it.uq.edu.au> <6sf87j$47n$1@hirame.wwa.com> <6sfrok$h3i$1@hirame.wwa.com> Content-Type: text/plain; charset=us-ascii Organization: Ericsson Austria AG Mime-Version: 1.0 Newsgroups: comp.lang.eiffel,comp.object,comp.software-eng,comp.lang.ada Date: 1998-09-01T00:00:00+00:00 List-Id: Robert Martin wrote: > >> Yes. Imagine that you had to change the function to make it thread safe; > >> and that the way to do that was to sieze and release a mutex while the > >> function was executing. As written you would have to add the release in > >> three separate places. But if you had avoided the multiple returns, you > >> would have had a single release. > > > >Well... The proper way to use a mutex is to wrap it in a controlled > >type, so that release is called automatically as a result of subprogram > >exit, no matter what the reason. (Controlled types in Ada have > >operations that are roughly analagous to constructors and deconstructors > >in C++.) > > In a language that supports such things, using controlled types is *a* way > (not necessarily the "proper" way). (it happens to be the way that I choose > in many cases). But this has nothing really to do with the issue at hand. > Yes, it may be feasible to put some resource management code into a > controlled type and avoid the issues of maintenance that I raised earlier; > but that doesn't eliminate the problem of structure. In the end, if you can > make the structure of the software solve a problem, that is better than > using a special language feature to do it. In the presence of structured exception handling, as in C++, controlled types are the proper way; everything else will lead to code duplication and unnecessary tests. This does not lead to the conclusion that multiple returns are a good thing, but neither is resource allocation and deallocation a good argument *against* them as far as object-oriented languages are concerned. With purely procedural languages, I concede that you have a point. Gerhard Menzl