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=-1.9 required=5.0 tests=BAYES_00 autolearn=unavailable autolearn_force=no version=3.4.4 X-Received: by 10.66.199.71 with SMTP id ji7mr35196318pac.23.1437430350727; Mon, 20 Jul 2015 15:12:30 -0700 (PDT) X-Received: by 10.140.104.147 with SMTP id a19mr84757qgf.2.1437430350478; Mon, 20 Jul 2015 15:12:30 -0700 (PDT) Path: eternal-september.org!reader01.eternal-september.org!reader02.eternal-september.org!news.eternal-september.org!mx02.eternal-september.org!feeder.eternal-september.org!usenet.blueworldhosting.com!feeder01.blueworldhosting.com!border2.nntp.dca1.giganews.com!nntp.giganews.com!f3no589028igg.0!news-out.google.com!b31ni224qge.0!nntp.google.com!69no929230qgl.1!postnews.google.com!glegroupsg2000goo.googlegroups.com!not-for-mail Newsgroups: comp.lang.ada Date: Mon, 20 Jul 2015 15:12:30 -0700 (PDT) In-Reply-To: Complaints-To: groups-abuse@google.com Injection-Info: glegroupsg2000goo.googlegroups.com; posting-host=73.138.95.40; posting-account=wEPvUgoAAABrLeiz_LRhQ3jeEhyfWVMH NNTP-Posting-Host: 73.138.95.40 References: <084c0dc7-ae74-4cb8-b1fe-78f42de94291@googlegroups.com> <5IidnWtEWowowjHInZ2dnUVZ5v6dnZ2d@giganews.com> <61db6272-72a4-47d7-b0e5-e870ac8a0432@googlegroups.com> User-Agent: G2/1.0 MIME-Version: 1.0 Message-ID: <9078db18-03ef-4724-9818-e9ae37e3d7a8@googlegroups.com> Subject: Re: Weakness of Ada is expensive / security / etc. ? Anything else? From: NiGHTS Injection-Date: Mon, 20 Jul 2015 22:12:30 +0000 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable Xref: news.eternal-september.org comp.lang.ada:26917 Date: 2015-07-20T15:12:30-07:00 List-Id: On Monday, July 20, 2015 at 1:35:37 PM UTC-4, Shark8 wrote: > On Monday, July 20, 2015 at 3:52:03 AM UTC-6, Serge Robyns wrote: > >=20 > > I'm surprised that the AdaCore GNATCOLL SQL library has no "standard" s= upport for Oracle, although it was already there in GNADE from which GNATCO= LL SQL was build. I'm ok to work and contribute back, but this removes me = from my own project. Moreover, the code is quite "shaky" and not resilient= at all, which came as a bad surprise to me or a language like ADA. >=20 > GNATCOLL is terrible -- I've *never* gotten it to build w/ DB support and= properly link... and that's even after the massive inconvenience of cygwin= (needed because the blasted project needs .configure and make). >=20 > What's doubly disappointing about that is that GNATCOLL is presented by A= daCore (which presents itself as the public face of Ada) as one of the core= useful-libraries and DOESN'T take advantage of Ada's nice portability inst= ead relying on the crap-pile that is make and .configure even when they hav= e/use GPR-files to represent a project themselves. (i.e. not using your own= tools is a vote *against* your own product.) >=20 > If you adopt the "eat your own dogfood" position, you have an incentive t= o make that a quality system, fix bugs, and make it easier to use. Here are a few comments on my personal experience with GNATCOLL and the gre= ater AdaCore code base. So I learn the language, fall in love (of course), then try to dive in and = make something useful as a way to get experience in this great human accomp= lishment of a language. So I go to AdaCore (the main google result to most = queries involving installation of Ada). Then you are presented with a page = with a "Build your own package" button. A laymen would likely not understan= d what this means. Compilers, Linkers, Tools, IDEs all have a clear single = point installation or at the very least some guide on what to do to get sta= rted. In AdaCore's world the most you will get is a huge list of checkboxes= and "On Windows, you can use the Extract facility of the Winzip utility." = Thats literally all you get, that and "If you have problems downloading or = installing GNAT GPL or SPARK GPL Edition contact us at gnat-gpl@adacore.com= ". Oh really? I can ask installation questions to this email address for GN= AT GPL?? Been there, done that. Here is the response I got: "We do not provide support for the use of GNAT GPL, and do not provide help on rebuilding from sources which requires some specific expertise." -- AdaC= ore Support. I love the irony. I pointed out the irony. They did not respond to it as ex= pected. Now any average person would run away as soon as they see the long meaningl= ess checklist of out-of-context packages, or perhaps after downloading all = of them become even more intimidated by the massive downloaded archive and = its convoluted rats-nest of archives and the occasional executable (If this= is a Windows installation). So a Windows user would be excited to see an executable and run it to get t= his installation started. Nope. You need a compile environment. If you are = like me trying to get the GTK installed as well as GNATCOLL, you need to in= stall MinGW with MSYS first (select all ada packages!), then install all ta= rgets (3 installers) in a very specific order changing all the default dire= ctory to the MSYS (NOT MinGW!) root folder, then use a command prompt to fi= nalize the installation of the GNATCOLL library only after ensuring that th= e environment paths are correctly pointing to the right root (it defaults t= o MinGW).=20 To get these packages installed it took me 2 weeks of trial and error. The = Windows installation took me nearly as long as the Linux installation. Each= had their own unique challenges, and each required that the order of insta= llation be perfect. But what drives me crazy is how many dependencies they = REQUIRE which are not initially stated by any documentation on the matter. = Its up to you to figure this complexity out. Now in my limited experience working with the AdaCore GDK (GUI) library I w= as disappointed with the behavioural differences when compiling between Lin= ux and Windows. I am forced to switch between both platforms and compile & = test on both when adding new GUI elements because some components behave wi= ldly differently between the two platforms. Additionally other components i= n the standard library are completely missing yet fully documented as if it= was expected to be there (For instance Listbox). Combine that with the str= angeness I sometimes encounter working in the library (the messagebox packa= ge has a mysterious undocumented "Arg5" System.Address parameter mentioned = no where in the entire world) and the lack of resources to help with usage = (my only friend was the example code, C-based documentation, and the GDK so= urce) and what I ended up with is a hugely unproductive and soured opinion = on the Ada path as a solution to serious application development. GNATCOLL got me excited when I read about how they leverage the features of= Ada's strong typing mechanism to protect SQL queries. But when you want to= actually use it the documentation is literally non-existent. The only plac= es you can look is in the GNATCOLL source code and a few places here and th= ere on the internet. I had posted a question recently on this newsgroup reg= arding not being able to DELETE anything in GNATCOLL. It went unanswered, a= nd I'm not surprised. This is very depressing to me. The library has so muc= h raw potential. It can be a market killer -- completely dominate the world= of RDBMS, and yet it's difficult to install, nearly undocumented, and seem= ingly waiting to die alone. I can't speak for the other libraries, but I have been creating my own libr= ary interfaces to compensate for a lack of out-of-the-box library support. = I'm so thankful that Ada is so easy to interface with C libraries and that = GprBuild is another shining human accomplishment. Well that's it for my rant on the AdaCore's non-insightful way to get newco= mers to use their products.