Lars Fosdal

Congratulations, @Attila Kovacs 🙂

First rule of fight club - you never talk about fight club. Same thing for beta versions.

If the initial description is accurate - you don't really have a record - but a stream? I.e. the length of the string is variable, not fixed and #0 padded? [int flags][int dummy][shortish string#0][int padding][int flags][int dummy][this is a lot longer string#0][int padding] Decoding the stream would have to be done element by element, and it would be necessary to know the format of the string - is it ASCII, ANSI (if so, what codepage) or UTF-8. To put it into a more manageable format, you could use your initial record - but you would have to read it element by element to stuff that record or object structure. This can be done in a number of ways - such as a TFileStream or by going old school BlockRead. A well designed stream would have had the string length as an int before the string - but now you instead have to scan for that #0 termination.

Self-repairing problems are the best, Totte 🙂

There is a workaround in the following thread.

We've replaced a lot of expensive industrial PCs with WinMate and Surfaces tablets. They are so much cheaper that the economy of possibly having to replace them more often is a no-brainer. The only exception are trucks that need to operate in -27°C - that is when you need those industrial solutions. Having those tablets on vanilla Windows 10 is a bonus.

@Hans♫ - How performant is the M1 compared to the i5?

There is one way that can avoid duplicate "hashes" or in this case - ensure a unique key - but it is not super performant, since lookup a database is expensive - but it will give you a fixed length unique identity for a signature. Create a signature table in a DB and add a text field for the signature and a constraint for it to be unique. Add an identity column as either an auto incremented int field, or an auto-generated GUID field. The identity column will be your unique "hash". Edit: If the "hash" will be exposed to third party - the GUID is recommended - since you cannot easily inject randoms GUID to access data - as compared to a sequential id which are predictable.

Even though they may refer to the same physical address? In a warehouse management system, unique client identities and their addresses is a continuous source of "entertainment".

A major difference here is that a coordinate in What3Words is strictly defined. A signature built from combined data from external sources is far more difficult. Consider '1688 High St.' vs '1688 High Street' as an example.

Signature - a unique footprint of data Hash - algorithmic calculation on a signature

It looks like he wants to use a hash as a signature for an address - but that signature is only as good as the code doing signature building and the quality of the supplied data. The algorithm would need to understand ordering, abbreviations, perhaps even common typos.

I just did an inventory of our warehouse hardware and discovered an Intermec industrial Truck PC running XP that had been overlooked. Hopefully, it will be brought up to ... drumroll ... Win 7 SP 2 soon. I guess it is time to start shopping for some newer truck PCs. Luckily, our truck PC software doesn't seem to have any Win10 specific code in it.

https://blog.mindorks.com/how-to-programmatically-take-a-screenshot-on-android explains how to do it in Java - but the question remains - how to get hold of the parent handle/coordinates in Delphi?

Doh! The risks of reading the Unread Content list and not paying attention which subforum the post is in. My bad.

http://forum.codecall.net/topic/60613-how-to-capture-screen-with-delphi-code/

@TiGü I do get Thai, Arabic, Chinese - even when using a font that does not have such characters in it. The reason seems to be that there is trickery going on under the hood for glyph substitution. https://docs.microsoft.com/en-us/globalization/input/font-technology It does look good and behave well in VS Code - and it does look like shit in the Delphi IDE. In VS Code - even navigation with right arrow goes right to left within the Arabic parts of the text. Did you create a report in Quality Central?

Yes. Same Consolas font in two different editors. The new VS Code, and the old IDE. I get it. I wonder - do you and I actually have the same Consolas font file? Mine has Greek and Cyrillic, but not Thai and Arabic. Edit: Original spec doesn't seem to include these either: https://docs.microsoft.com/en-us/typography/font-list/consolas Could you attach that .pas file, please - so that I can see how it looks in my IDE?

Or, Åström (U+00C5 U+0073 U+0074 U+0072 U+00F6 U+006D) which is using the correct Unicode chars, i.e. not built from composites - which we know is not handled for monospaced fonts in the IDE.

@TiGü The IDE editor supports Unicode, but most fonts do not have every Unicode character for every language defined, hence which font is being used, really matters. The editor is definitively left to right, and made for monospaced fonts, which does not work well with national character sets that uses ligatures and variable kerning. VS Code is many years younger than the Delphi Editor, which started supporting Unicode source in Delphi 2009 - so yes, it could use modernization. Your example is definitively a good one - and should be reported as an issue in https://quality.embarcadero.com/ If you do, make sure you attach the original .pas file as well as the screen shots.

Good question. I am afraid I am not of any help.

@Sonjli - Did you read this article? https://blog.grijjy.com/2020/10/07/an-xml-dom-with-just-8-bytes-per-node/ Not sure how that lib would fare with regards to CPU usage - but it should be fast enough. Edit: Doh... I didn't realize Neslib was the lib in the article 😛 Need more coffee.

I guess you've read this: https://stackoverflow.com/questions/39190582/error-itms-90023-missing-required-icon-file Edit: Never mind - seems to be irrelevant. Are you using the most recent version of XCode?

Another solution to the font predicament is to use a translation tool for the actual text to be displayed, keeping the string constant in the code in English. In translation tools, you are usually not bound to monospaced fonts. This would also allow programmers not fluent in Thai to understand the text, and it would make the application ready for translation to a multitude of languages.

For future reference: Which font was it and where did you find it? Edit: An idea for a future enhancement of development IDEs would be support for custom fonts (that not necessarily are monospaced) for string constants in the source code. I won't be holding my breath for that to arrive, though - since it opens a hornet's nest of questions such as "Why does the cursor jump many characters to the right when I navigate up or down from this string constant?"