Jump to content

Stefan Glienke

Members
  • Content Count

    342
  • Joined

  • Last visited

  • Days Won

    27

Stefan Glienke last won the day on October 9

Stefan Glienke had the most liked content!

Community Reputation

391 Excellent

6 Followers

Technical Information

  • Delphi-Version
    Delphi 10.1 Berlin

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Stefan Glienke

    Unit testing cross platform code

    DUnit runs on all platforms that Delphi supports - it just does not have the dcus shipped for each which requires adding the source directory of DUnit for those platforms (such as Linux). I mean the core framework code, not the GUI runner of course - but I think there are some FMX out there and even if not writing your own is easy enough. Running tests on Windows and Linux (and I am referring to pure unit tests) is as easy as it can be using TestInsight. Mobile can be a bit more tricky as you cannot simply shoot the app there and run it without interaction to collect the results (not 100% sure as I don't do mobile or mac development).
  2. Stefan Glienke

    Quick debug variables?

    Writeln is a joke compared to Pythons print. You can pass any object to print and it then depends on which methods the object implements what gets printed (see https://stackoverflow.com/q/1535327/587106) And yes, almost all library types implement such methods so you can simply pass arrays, lists or dictionaries to print and get their content printed out.
  3. Stefan Glienke

    Modern C++ and Delphi

    FWIW in Delphi that fear is often caused by the fact that new features often don't work for years and are poorly designed so they affect some important aspects negatively such as compile speed or runtime performance which often you encounter way later when you already extensively use them (latest example: inline variables). Many other languages have a plethora of incredibly smart people designing things for years with enough time to thoroughly test and ensure they are zero or minimal overhead.
  4. You do realize that the -x options are provided by the fixpack and thus are only via some hackery done by Andreas and might not be stable, yes? @jbg FYI
  5. Stefan Glienke

    Changes in Parallel Library

    I do and it will be subject of one of my talks I will be giving at some conferences in the coming months so I will not spoil that before.
  6. Stefan Glienke

    Modern C++ and Delphi

    IOW: no, they still don't work
  7. Stefan Glienke

    Changes in Parallel Library

    Which also was an RTL thing that was very late to the party and thus has to compete with solutions that already existed before it. And yes if there have been other solutions in need of some multi-threading they either had to roll their own or only be compatible or rely on other third party components themselves which they usually rather refrain from. So what we have achieved in the past few years in the RTL is closing the gap on the feature matrix to give the impression that for nowadays standard usecases like http, json or easy parallel processing you don't need any specific third party components or libraries but once you start using them you quickly run into numerous issues that range from "little annoying" to "absolute showstopper". And often fixes either take multiple iterations or major versions and range from "now you broke something else" to "dirty hack because of keeping broken behavior because of backwards compatibility" Even using the smallest RTL feature these days causes me to look into the implementation before recommending it to my coworkers, telling them to stop using them or not giving green light to upgrade the newest Delphi version (RSP-19439 for example) - sad but true. If it weren't for licensing reasons and the impossibility to fix many bugs that interest me by just recompiling the RTL (depending on what units you change) I would have started an "enhanced RTL" project already to fix the most annoying issues (like the ridiculous code bloat caused by System.Generics.Collections/Defaults and many other annoyances). At least the latter was solved by banishing those units from our code entirely and exclusively using spring4d collections. Using any third party code and the RTL and VCL itself though still is affected by that.
  8. Stefan Glienke

    Changes in Parallel Library

    We have a catch 22 here. If stuff does not work I don't use it - especially if it tries to solve a task that has already been solved years ago. So if people don't use it less people are affected by bugs and less likely it is they get fixed - however even when issues are being reported by many people it takes time to fix them and only with an upgrade you get those fixes. With a library like OTL you don't need to upgrade your entire IDE to get some bugfix. This is the fundamental flaw you have when IDE, compiler and runtime are so tightly coupled that they cannot be upgraded individually. That plus a lack of quality control.
  9. Stefan Glienke

    Changes in Parallel Library

    I doubted that already
  10. Don't tell me - but this the almighty excuse brought up by the decision makers almost every time some issue like this is brought up - "but what about existing code" - like for that dreaded "directly passing result of ctor call to interface parameter issue" - what about existing code? Well existing code is most likely suffering from a defect. The stupidity of the Delphi compiler in almost every aspect and the unwillingness to change anything significant on that is making me furious!
  11. Stefan Glienke

    How best to update from 10.3.1 to 10.3.2?

    Updating a single application even if complex is not a hard task, almost every other major IDE out there can do it. Only Delphi developers have to manually apply hotfixes by unzipping some archive and unblock the contained binaries...
  12. Changing that would raise some warnings in existing potentially defect code and some people don't want that Similar to being able to call non virtual class methods on instances
  13. TL'DR 😛 Btw you can utilize the xmldoc for that - we are doing that for some of our classes and pump that into confluence. Does not need a single line of executable code or RTTI inside the to be documented to produce this. The only disadvantage this has is that implementation and what's in the xmldoc might differ.
  14. Keep voting: https://quality.embarcadero.com/browse/RSP-13290 And given that the compiler already knows line numbers and the filepath it's working on for Assert it should be almost no effort to implement something like this as well: https://docs.microsoft.com/en-gb/dotnet/api/system.runtime.compilerservices.callermembernameattribute?view=netframework-4.7.1
×