Re: Elements for inactive additional selections

On Wednesday, March 27, 2024 at 8:58:14 PM UTC Neil Hodgson wrote:



I'm having trouble working out what bug could possibly cause these symptoms 
on some systems. It's the same code drawing the active and inactive states 
and it should only differ in colours and translucency. Perhaps the drawing 
surfaces are different configurations on different machines (high-colour 
versus shallow-colour, for example) and the platform drawing code didn't 
implement some drawing operations on particular configurations or 
implemented them poorly.

They didn't test any of their XCB code. Someone was Hell bent on adding 
Raspberry Pi support and someone else was chasing Wayland and

*Agile with TDD is ___NOT___ Software Engineering*
 
When one is relying 100% on CI/CD TDD they are 100% of the time shipping a 
hand polished turd.

Qt went down this path and added insult to injury adding the completely 
useless QML which is why the industry is abandoning it . . . Well that and 
the audacity of charging $650,000 for 3 developer seats.

For the "latest" Scintilla that I merged there were some XCB changes that 
simply weren't needed on CopperSpice. I thought about playing with putting 
them back in but . . .

QXcbConnection: XCB error: 3 (BadWindow), sequence: 12768, resource id: 
11337400, major code: 40 (TranslateCoords), minor code: 0 QXcbConnection: 
XCB error: 3 (BadWindow), sequence: 18651, resource id: 11329781, major 
code: 40 (TranslateCoords), minor code: 0 QXcbConnection: XCB error: 3 
(BadWindow), sequence: 60011, resource id: 11592169, major code: 40 
(TranslateCoords), minor code: 0 QXcbConnection: XCB error: 3 (BadWindow), 
sequence: 1776, resource id: 11667167, major code: 40 (TranslateCoords), 
minor code: 0 QXcbConnection: XCB error: 3 (BadWindow), sequence: 3369, 
resource id: 11692114, major code: 40 (TranslateCoords), minor code: 0 
QXcbConnection: XCB error: 3 (BadWindow), sequence: 5456, resource id: 
11691685, major code: 40 (TranslateCoords), minor code: 0 QXcbConnection: 
XCB error: 3 (BadWindow), sequence: 13873, resource id: 11789353, major 
code: 40 (TranslateCoords), minor code: 0 QXcbConnection: XCB error: 3 
(BadWindow), sequence: 22011, resource id: 10903244, major code: 40 
(TranslateCoords), minor code: 0 QXcbConnection: XCB error: 3 (BadWindow), 
sequence: 27434, resource id: 11289833, major code: 40 (TranslateCoords), 
minor code: 0 QXcbConnection: XCB error: 3 (BadWindow), sequence: 39643, 
resource id: 11103538, major code: 40 (TranslateCoords), minor code: 0 
That is coming out of the CopperSpice code for almost every selection left 
on screen for a few seconds. Someone knew they broke it, put a debug 
statement in, then checked in the code.

It's start of spring spraying season here (tomorrow) so I don't know when I 
will get back to it, but . . . The plan going forward is to get RedDiamond 
to build with of the previous CopperSpice Debian (I changed something about 
CMake and have to remember what it was) and just lock it down. I've already 
told them I'm only going to merge their code once per year. I have spent an 
entire year working in the same branch while they contemplated their naval 
lint about merging in my RPM and Debian code *which only had to change one 
line in one file of their code*. All the rest were scripts for building and 
packaging both Debian and RPM.

When I come up for air I'm going to officially fork under new name because 
I need to be able to Flatpak RedDiamond in the coming year and their 
academic pursuits are getting in the way.

I got really corked off in the forum today and they probably won't ever 
allow that post to be seen by the general public. They mandate a C++17 
compiler and are chomping at the bit to force C++20 on the world yet they 
have pre-C++11 enums everwhere.

Scintilla has to straddle C and C++ so they old school "use an int for 
everything" has to exist. CopperSpice does not. The claim they made about 
"breaking the installed base" was the spark that lit the fuse on 27 kegs of 
powder. They haven't been around long enough to have an installed base that 
matters. What they have is a tiny handful of customers they have done 
billable work for, that's it. Honestly, I have more code in production at 
more clients than all of them combined . . . so do you Neil. 

Sorry, been popping in and out of the office and a few kegs were slow to 
detonate.

So far the latest version of CopperSpice has this problem on every system 
I've tried. The AppImage (which accidentally built with the previous 
CopperSpice due to a rename issue) seems to work everywhere. 

Last night I spun CsScintilla back to the previous debian package, built on 
this machine, and still have the problem. Since RedDiamond code doesn't 
actually do this, there is only one place left for the problem to exist.

I can make the exact same thing happen with the CsScintilla example 
programs. Those have basically been unchanged since I wrote them over a 
year ago. They didn't have the problem then.

Take deep breath. Easter arrives Sunday. I can drink again then.

-- 
You received this message because you are subscribed to the Google Groups "scintilla-interest" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [email protected].
To view this discussion on the web visit https://groups.google.com/d/msgid/scintilla-interest/48ebee98-2636-4004-a973-a9cbb16f8516n%40googlegroups.com.