Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    salotz
    @salotz:matrix.org
    [m]
    I also don't use LSP nor do I want to. I'd rather not have anything at all...

    However, there are a few others that may make you Emacs experience better.

    Any particulars?

    salotz
    @salotz:matrix.org
    [m]
    my gut feeling is that ada is quite simple syntactically. I have no idea though. Probably they tried to add all kinds of IDE like features. This is kind of a problem... I will try to install it manually and see if I can get it to work, could just be the recipe on elpa is out of date
    I also am thinking about starting a yasnippet collection for ada. There is none in the "official" listing. Perhaps that would alleviate some verbosity issues newcomers have
    salotz
    @salotz:matrix.org
    [m]
    Orzech: When you tried to install ada-mode was it 'wisi' giving issues?
    1 reply
    salotz
    @salotz:matrix.org
    [m]
    yes I just found that. Usually emacs packages have hooks to run the compilers. But looks like you have to run build and install script. I tried running this but I failed...
    its probably something stupid with my toolchain setup since I don't really know how to do that yet
    figured it out. Just needed to install some libraries :)
    1 reply
    salotz
    @salotz:matrix.org
    [m]
    sudo apt install -y \
         gnat \
         gprbuild \
         libgnatcoll18 \
         libgnatcoll18-dev \
         libgnatcoll-xref19 \
         libgnatcoll-xref19-dev \
         libgnatcoll-sqlite19 \
         libgnatcoll-sqlite18-dev \
         libgnatcoll-sqlite-bin
    on Ubuntu 20.04
    salotz
    @salotz:matrix.org
    [m]
    no...
    1 reply
    its ada-mode 7.1.4, perhaps 7.1.3 will work as that what that site says to use?
    I get Symbol’s function definition is void: wisi--lexer-error
    1 reply
    ..
    1 reply
    salotz
    @salotz:matrix.org
    [m]
    with 7.1.4 right?
    2 replies
    salotz
    @salotz:matrix.org
    [m]
    I failed to compile from the ELPA tarball...
    🥲
    ohenley
    @ohenley

    @ohenley what should we do with these PRs that want to add links to GH repos not containing any source code (and instead just throw code over the wall by adding binary .7z files)?

    @onox. yeah... I am somewhat procrastinating to formulate to the OP why those should not, IMO, be included in the list. I hate having to push back on someone's 'efforts'. What do you think? Should we just go along (I already did with some previous PRs)? ... but then he might become 'productive' and start pushing for more and more. Man open source and QA ... sometimes :facepalm:

    ohenley
    @ohenley

    @salotz:matrix.org @irvise:matrix.org Rust is pop culture, it's Mozilla, it's all the old school C guys with a social/cult following giving their approbation for a new kid on the block. It's also flexible, brings in a lot of loved high-level idioms from 'toy' languages, and adds to the table core 'new' features of real value. Now is it geared to make engineering; with big teams, rolling staff, lasting and maintainable architectures, uniformity of practices? In my opinion, it is not.

    I once flamethrowed a junior guy on Rust forum because he was musing over an indecipherable if/else equivalent (My post got 3 stars but then was ban for rudeness. This whole incident just nuked my Rust interest. I am not interested in abstraction pissing contests. I just want to work). You cannot reinvent the if/else concept in Ada and that is a great, basic, engineering provision. Any good senior knows why. But some code libertarians do not like this fact, it's boring, maybe it will impede their future, it takes out wizardry.

    Rust is pop, so pop that even someone like Jonathan Blow, who talks like he is second to God does not know better. He never talked about Ada in his design for Jai, yet he will reinvent or not what Ada demonstrated to be of essential value to make great 'high rise' code constructions.

    Now if someone does not understand/does not want to know/is not curious enough to find out that using a readily available 'battle tested' technology like the Ada Adacore stuff ( from gpr to gnat to gnat libs to ASIS to libadalang to gnatstudio to ...) can and will make their retail endeavor better over all aspects of software engineering, well, their credibility should stay where their knowledge is at; pop culture. The quality Ada offers is valid for any context and failing to see this is just one's limitations; trends and preferences do not win over rational arguments of professional concerns.

    0rzech
    @0rzech:matrix.org
    [m]
    @ohenley What do you mean by "'toy' languages" and "indecipherable if/else statement"?
    ohenley
    @ohenley
    @0rzech:matrix.org 'toy' languages: eg Python, Js ... most script languages that have ambitions where they then can then be perceived as toy, eg. Attempting large system programming using Python.
    salotz
    @salotz:matrix.org
    [m]
    @ohenley: really like what you said. I agree about John Blow /et al./ but at least they have their own specific requirements they are designing against. The one point I'm wondering comparing the Jai and other newer C++ contenders to Ada is compile times.
    Does Ada have good compile speeds? I understand that the package system helps with this a lot since incremental builds are more well defined and robust than in C/C++ but as far as raw compiler speed for real projects. This is something difficult for me to test and discern for myself
    ohenley
    @ohenley
    @0rzech:matrix.org indecipherable if/else statement: using fancy 'generics' patterns and what not, the OP developed an original expression that was equivalent to a simple if/else for a particular context. IIRC, he even pushed to arguing 'I prefer the expressiveness of my way'. I just explained that in a company such liberties are unwelcomed because nobody is interested in decipherering one long gone colleague when we need to go back in its now breaking code. I did such error in the past, I also suffered such situation in the past and therefore feeled entitled to raise a concern.
    Orzech
    @0rzech:matrix.org
    [m]
    @ohenley Ok, thanks!
    salotz
    @salotz:matrix.org
    [m]

    thindil:

    As far I see, the last released version on the official page is 7.1.3, but GNU ELPA shows 7.1.4 🤔. Maybe try the older version 7.1.3? You could also try contact with Stephen Leake, the author of the ada-mode: or on comp.lang.ada or on ada-mode mailing list.

    Its a bit confusing in terms of which version is right... on the ada-mode downloads page there is different versions than on elpa. The readme has one version the package has another. I think I will just contact Leake and see what is up

    Orzech
    @0rzech:matrix.org
    [m]
    salotz I think there was an error in versioning and 7.1.4 is labeled as 7.1.3.
    salotz
    @salotz:matrix.org
    [m]
    in any case, if I install with emacs package install and then run build scripts it succeeds to build the ada components, but if I download the tarball myself it fails..
    I wonder if libadalang is going to replace the semantic analyzer stuff in ada mode? There is a check for it in the ada-mode build.sh script. I could actually do something with that since I know python :P
    salotz
    @salotz:matrix.org
    [m]
    "code libertarians" is 👍️ lmao. The sad thing is that if your going to be a real code libertarian and do syntactic extensions & redefinitions, live systems updates etc. you should really go to a language like scheme/common lisp where they take that stuff seriously and actually give you useful tools to make that kind of approach more viable
    Orzech
    @0rzech:matrix.org
    [m]
    @ohenley salotz While I'd phrase my opinion different way, I think software industry is very susceptible to trends.
    ohenley
    @ohenley
    @salotz:matrix.org It's true, J.blow has ambitions for game programming, that's it. Concerning compiling speed I do not know really... (like you said the package system helps and the whole thing is robust, definitely) but for me, it's not that much of a deal. I far prefer the FACT that Ada's compile/link/etc messages are miles clearer than C++. They just make sense. Also, IIRC, distributed compilation is part of GPR (never used it myself): https://docs.adacore.com/gprbuild-docs/html/gprbuild_ug/building_with_gprbuild.html#distributed-compilation. That is what I mean by Ada being geared for engineering, we may not like the whole ecosystem, but it is really geared toward serious endeavors. Good stuff is all there already.
    salotz
    @salotz:matrix.org
    [m]
    well if its never been an issue for you I guess its not that bad, unless your expectations were too low already :)
    Has anyone ever seen a project that is like building a toy compiler for some core subset of Ada features for pedagogical purposes?
    1 reply
    ohenley
    @ohenley
    Bah, my expectations are no less than Ubisoft C++ multi-million lines .sln. When you touch something core, you just go take a coffee. I know Jai looks just fantastic, but there is more to the whole picture. I do Ada as a hobby, so compile time...
    salotz
    @salotz:matrix.org
    [m]
    ok forgot about your awesome page: https://github.com/ohenley/awesome-ada#compilers
    salotz
    @salotz:matrix.org
    [m]

    Rust is pop culture, it's Mozilla

    whats funny about this is that Mozilla is basically dropping their bull push for Rust, that slack is being picked up by the other big tech giants after they did their major layoffs. My suspicion is that costs skyrocketed after trying to use rust for browser components

    ohenley
    @ohenley
    But to anyone interested in the future of langs in general, if I may, I really think one should give a serious look at ParaSail. It's the clean slate project by the main architect of Ada 95 onward, Tucker Taft. If you want something incredible, check there. IIRC, everything is an interface, oo, strongly typed, parallel constructs are first-class (eg. map reduce expressions), containers are first-class, no pointers, region-based, var-const-ref, contracts, compiled but also can be interpreted, reflection, natively interface to Ada and C for anything 'wilder' (io, OS libs etc). https://adacore.github.io/ParaSail/images/parasail_ref_manual.pdf
    1 reply
    salotz
    @salotz:matrix.org
    [m]
    definitely
    ohenley
    @ohenley

    whats funny about this is that Mozilla is basically dropping their bull push for Rust, that slack is being picked up by the other big tech giants after they did their major layoffs. My suspicion is that costs skyrocketed after trying to use rust for browser components

    Interesting, did not know apart from Mozilla somewhat dropping Rust.

    ohenley
    @ohenley

    "code libertarians" is 👍️ lmao.

    :joy: thx!

    salotz
    @salotz:matrix.org
    [m]
    well not officially I think
    Irvise
    @irvise:matrix.org
    [m]
    I really like ParaSail in its premise. Bring the "philosophy" of Ada to newer techniques. I hope that it takes off.
    ohenley
    @ohenley
    Me too. I plan to invest a lot more time in ParaSail.
    salotz
    @salotz:matrix.org
    [m]
    From what I understand ParaSail is a new design for "automatic" memory management, like the scenarios in which C/C++ and Ada are used like where you precise control of memory layouts and such (for games, numerical calculations etc.) ParaSail isn't so useful there. Perhaps I'm mistaken
    thats one thing I was trying to figure out about it. I was also trying to compare it to the approach of Chapel for distributed/parallel languages
    6 replies
    ohenley
    @ohenley
    BTW, just created https://gitter.im/parasail-lang/community for further ParaSail chats.
    onox
    @onox
    @ohenley I think they should not be added to the list
    ohenley
    @ohenley
    @onox Ok I will comment to the OP asking for "code" not bundled binaries and what not.
    onox
    @onox
    :+1: