Fr0sT.Brutal

VCS branches are good when a feature is tied to project. When it's a new concept or some new class, I prefer to create and test it in new standalone project thus not carrying the burden of a whole big project (build time, long code, excess actions etc). My old-school colleague is testing all features on a main app which takes ~1 min to build and ~5-6 mouse actions to reach functionality he is testing. I wonder how he hadn't go crazy yet %-)

You could look at https://github.com/Fr0sT-Brutal/Delphi_BuildScripts for inspiration

Git + Tortoise is my choice. I tried Fossil as well and it's pretty nice to start using VCS with. It even has web-based issue tracker! There's Mercurial as well but Git seems to become a "VCS Google search" currently. One will likely use Git for 3rd party projects so why not use it for own projects as well. And, for OSS purists, it is developed by Legendary Linus!

You also can use 'absolute'

I use msbuild for anything more than D7. dproj files contain tons of additional necessary stuff like lib subfolders, compiler options, defines etc. In fact, with preliminary configuring it allows generating f.ex. all combinations of platforms and build modes in several lines.

Another +1 for throwing an error. You can't tell whether it's OK to replace some value from "foo" to "0" and will take responsibility for this replacement f.ex., if someone's bank balance would become zero when a corrupted input is given to you app 🙂 In the area you described specs are your Bible, your weapon and defense. Don't do anything more or less

FastMM can dump the whole heap, this could help in those "pseudo-leaking" objects that grow in number at runtime but are correctly disposed on exit

Thanks! Plz could you fill an issue at Github so that it won't be forgotten. Meanwhile you could refer to OSM.MapControl.pas TMapControl.DrawMapMark for ideas of how to draw and try OnDrawMapMark callback to do what you need.

https://www.clevercomponents.com/products/datapump/index.asp + SQLite ODBC driver

In your scheme with tokens FTP is useless IMHO. Do everything with HTTP. You'd have to use custom clients anyway (to utilize token stuff) so no sense in FTP at all. Alternatively, you can really deal with temporary logins and folders personal for each user. This way FTP is fine and any standard client could be used. Expiration could be done by the moment of last activity

Hoorray, pretty short post from David :)))) I think whether divide apps or not depends on 0) Architecture If user = admin sometimes then any other factors are unrelated, just do what you like. 1) Security If a user doesn't possess admin interface binary, it couldn't inspect what you probably want to hide 2) Use-case How the admin app is to be used. Whether it could be run from the same PC as user app? How frequently? Whether an admin could occasionally copy the app to machine to do some things or it will be very unhandy? Anyway it's very generic subject. Without an architecture of a system the questions will be just a set of preferences...

If the subject is just about setup ISOs then they are likely remain in torrents. Emba could setup its own initial seeder if they're worrying about bandwidth but still care of old customers. If not, old ISOs will still live in torrents but without Emba.

Surprisingly, ARC <> GC

Honestly didn't read thoroughly - too much text. If you've got to store some user-related data then you have to use accounts. Or you may save things in local storage but it has drawbacks

I'd prefer using nil as a sign of taking some previously configured encoding. Or, if you dare, you could use code page numbers.

AFAIK Jedi has this feature as well

Okay let's go deeper. You intend to use custom TCP. So you have to implement both sides (client & server) of it, develop message format, likely including auth & encryption, different request types (get list, put file, ...) and so on. Of course that's possible and not that hard but seem useless in your case. Sure TCP won't go away I was talking on FTP only. Well, it has some design problems (two channels f.ex.) and for some reason more and more servers drop it in favor of HTTP. HTTP is implemented with system-native components which means you won't bother with custom SSL libs and it's pretty powerful protocol that is able to do all things you want. If you mean completely serverless solution then probably any TCP-based protocol won't fit your needs, maybe you'll need UDP multicasting.

Why inventing custom solutions (in TCP you'll have to develop whole protocol and FTP is very old and likely will become obsolete in several years) and using 3rd party components when you have full-featured HTTP with system-native HTTP request component.

It could be even simpler. Just "eval expression" `IntToStr(GetTickCount)` and log the result.

Flags or removing event handlers are OK. It's the only option to get things done. VCL is very dumb in some places. F.ex., some controls do not distinguish between user-made and app-made change calling handlers in both cases and some see the difference so you have to run control.OnChange() manually

Interesing idea but too harsh solution IMHO. I'd rather consider spoiling some app output if it's the main app's goal. Or the behavior of xplorer2 could be examined. It's not only the app with the strongest protection I've ever seen (made by a single author, uses offline license key and still couldn't be cracked reliably!), it fails on invalid key harmlessly and randomly. On some random moment it just closes and removes all its settings from registry. For file explorer software it really hurts.

https://github.com/Fr0sT-Brutal/awesome-pascal#code-checkreview-debug

You can create hand-drawn-like visual style and apply it to Delphi app

I always used MaxInt to copy til the very end but wasn't aware I can omit Count 🙂

There are plenty of object wrappers that allow using ref-counting with any object, and everyone could write his own one in 10 minutes (I did, and use it for temp datasets)