dummzeuch

Which doesn't help, since he is using Lazarus, not Delphi: On the other hand, the source code for the LCL is also available, but it's much more convoluted than the VCL, because it's supposed to be cross platform.

Due the the COVID19 pandemic I am currently working from home, using Putty + ssh + Remote Desktop to log into and work on my office PC. For this to work, the office PC must be turned on and booted. So far I have let it running 24h which is really a waste of energy but since sometimes nobody is in the office at all, that was the most fool proof way. Today I have had some time at my hands waiting for an Ubuntu server to finish installing, so I thought about alternatives. ... read on in my blog post.

Apparently not. At least I have not seen any spontaneous awakenings 😉 of our office computers. There are three check boxes on the "Power Management" tab of the network controller properties (on my computer): * Allow the computer to turn off this device to save power * Allow this device to wake the computer * Only allow a magic packet to wake the computer On my computer, all of them are set. On a colleagues' computer the last one was not set. My colleague said that he didn't touch this (but on the other hand I don't remember doing that either on my computer), so I don't know what the default is. I can only assume that it depends on some circumstances during the installation. Maybe different hardware, or maybe whether the computer belongs to a domain or not (mine doesn't, the colleague's does).

Aparently many modern computers are already configured to wake on any LAN activity, not just the WOL magic package. So e.g. trying to RemoteDesktop into a computer might be sufficient to wake it up. I tested this with two of our computers, both of which had a special setting to restrict WOL to the "magic package", which seems to be off by default. But I am not sure from which state this will work. WOL didn't work at the one time I actually shut down the computer rather than hibernating it. Another one didn't wake up from sleep. The only state for which this has always worked for me was hibernate.

No, that's perfectly valid. But why don't you decleare RecPointer as ^TIniRec and save yourself the typecasting? Or are there pointers to different types? If the latter, why do you use records instead of classes? Actually I don't think this should compile: TComparedData = record PointerRecType: TRecType; // this identifies RecPointer as TINIRec RecPointer: Pointer; // Pointer to TINIRec ... function IsSectionHeaderLine: boolean; end; function TComparedData.IsSectionHeaderLine: boolean; begin Result := false; if RecPointer <> nil then Result := TINIRec(RecPointer^).IsSectionHeader; // <-- this should not compile as RecPointer is an untyped pointer and therefore RecPointer^ is undefined. end;

No, that one is sent by the computer which runs the ssh server.

It does, but unfortunately only on a daily basis which would mean that it starts on weekends (and doesn't shut down so it runs all weekend). WOL is probably the easiest option for me (now that I know that it works) since I log on to an ssh server first thing anyway, so I can simply call a script that wakes up my PC. I could even automate via .profile it but that would mean that PC starts every time I log on even if I don't need it.

Only one for ssh (but not the default port). Everything is tunnelled through an ssh connection.

I can't look at my office PC now (it's in the office, I am not and I can't connect to it because it's hibernated 😉 ), but on my home PC there is indeed such a setting: EDIT: Wonders of remote access: Yes, you are right, it's supported. I just ssh'd into the LAN installed etherwake and woke up my PC using the wakeonlan command. Apparently the option just doesn't show up in the BIOS. Thank you very much for this hint!

Recently I came across a problem with INI files: Some editors (including recent versions of Windows Notepad) add a byte order mark (BOM) to the files they save. In particular the BOM for UTF-8 kept appearing in INI files which then were read incorrectly by the Delphi 2007 implementation of TMemIniFile (I guess the same applies to all pre Unicode versions of Delphi). In particular this was a problem with programs that used TJvAppIniStorage for streaming application settings to disk. (TJvAppIniStorage internally uses TMemIniFile.) So I tried to fix this ... read on in my blog post

I just got a very odd error message from a TJvSaveDialog: M.A1R.SigBrightness_2020-03-25 The filename is not valid. As you can see in the screenshot, there are already several files with the same name scheme in that directory: The dialog seems to come from the Windows Dialog itself, at least the call stack doesn't show any JVCL or RTL code after the call to Execute. I admit that the file name is a bit unconventional, but it should still be valid. If I save the file as e.g. 'bla.dat' I can later on simply rename it to the name I originally intended. Also, if the file already exists and I simply select it, I get an overwrite prompt and then can save the file. To be sure that there are'nt any invisible characters there, I type the whole thing. The error occurs anyway. I also tried different directories, on the server and on a local drive, even in the root of drive d:. No difference. WTF is going on here?

OK, found it: The problem was not the file name but the filter (Save as type): If I change it to *.* it works. Seems that Windows for whatever reason doesn't like the filter to contain a * at the end of the extension (in my case "*.SigBrightness*". If I switch to the filter "*.SigBrightness" (without the trailing *), Windows appends ".SigBrightness" to the file name, which I can then manually remove and save the file with the intended name. Now, how do I get Windows to show the selection I want but still let me save the file?

I think that some functionality in GExperts (and some of my other tools) could profit from switching from sorted TStringLists to a dictionary. Unfortunately I can't just use TDictionary<string, [whatever]> because I want to continue supporting Delphi versions that didn't have generics (in particular Delphi 2007). Does anybody know of a free dictionary implementation that stores pointers indexed by strings? I know of THashedStringList but it has too many gotchas to simply use it (e.g. the hashes get recalculated every time a new item was inserted and IndexOf is called, which is exactly what I would need to do). Yes, I could write one myself, but why reinvent the wheel if there already is one?

I just downloaded the 30868_bde_installer_for_rad_studio_delphi_c_builder_10.3_rio.ZIP from Embarcadero and tried to install the BDE on a computer that I want to run some programs that still depend on the BDE. And I got the following error: --------------------------- No Rad Studio 10.3 found --------------------------- BDE install can not continue because Rad Studio 10.3 could not be found on the system. Please install Rad Studio 10.3 and try this install again. --------------------------- OK --------------------------- WTF? Does anybody know how I can install the BDE on a computer that does not have (and will never have) Rad Studio installed? Is there a separate installer for that? (And please refrain from telling me that we should get rid of the BDE already. We're getting there but we're not there yet.) twm

If you only ever handle files created with Windows programs, that's probably fine. Unfortunately I also get files from other systems, where the BOM might be useful.

The parser still has bugs. It could be as simple as that. If you can produce a small excerpt of the unit that produces this behavior I will have a look.

IT'S A PROBLEM IN THE BUILD PROCESS THAT ONLY AFFECTS THOSE WHO USE IT. I DON'T CARE IF YOU CALL IT A BUG, FOR ME IT'S AN ANNOYANCE AND THERE IS A SIMPLE WORKAROUND THAT TAKES ALL OF 30 SECONDS TO APPLY. FIXING THIS SO CALLED BUG WOULD TAKE ME AT LEAST HALF AN HOUR. AND THAT'S MY TIME, NOT YOURS. (OK, I will stop using caps now.) I won't take the time to fix this, it doesn't affect me. If you want it fixed, fix it yourself.

No, I won't. It's just an annoyance and can be worked around easily. But if you think it is worth your time: Go ahead.

I was just about to say that Alt+I is already taken by some other function, because I distinctly remember not setting it for some reason, but apparently it isn't and my memory is wrong. So: Done in revision #3063

OK, I know what the problem is: The GExperts build scripts do not support paths that include a space character. (Which is actually documented.) Rename your "GExperts Sourcecode" directory so it does not contain a space character and it should work.

I just tried to reproduce this problem: * Deleted all files in the Projects\DelphiXx103 that are not under source control (including the GExpertsRS103_Version.RES file you mention * Deleted all files in the dcu subdirectory * run the _Build_Project.cmd file I got no error messages.

There was once: Look up STColl (part of TurboPower's tpsystools): - STCOLL theoretically allows up to 2 billion elements. The collection is "sparse" in the sense that most of the memory is allocated only when a value is assigned to an element in the collection. - STCOLL is implemented as a linked list of pointers to pages. Each page can hold a fixed number of collection elements, the size being specified when the TStCollection is created. Only when an element with a given index is written to is a page descriptor and a page allocated for it. However, the first page is allocated when the collection is created. Unfortunately TurboPower went out of business back in - hm, the article on slashdot says 08/01/03, so was it 2008 or was it 2003? Damn stupid American date format - I think it was 2003. Their libraries were open sourced though.

There are other structures that don't actually need contiguous areas of memory, e.g. linked lists or some tree structures. But of course these have other weaknesses.

Hm, looks like somewhere a relative path is resolved wrongly. GExpertsRS103.res is not the same file as GExpertsRS103_version.res. The former is automatically created by the IDE and should not be referenced in the project. The latter is supposed to be generated by the pre-build script, which reads the version number from the GExperts_version.ini file, creates an .rc file from it and then compiles it to the .res file. Apparently that process is broken.

And also: Responding to any questions that may follow. (Unfortunately I am guilty of not answering questions to some of my QC(?) reports: When I'm on holiday I simply can't. Guess what? They closed the report.)