myScite / Notes: Recent posts

Hello everyone. found the sources and decided to push them to git.
started playing around with the Chatgpt api via lua. its super easy and as i already compiled luasec linked with openssl some time ago.
Here's how you do it - Have Fun!
~~~
local https = require("https") --ssl.https
local ltn12 = require("ltn12")
local json = require("dkjson")
local api_key = "get an api key from https://platform.openai.com/"
local prompt = "Hello world from SciTE'S lua interpreter, do you know Scintilla ?"
local response_body = {}
local request_body = json.encode({
model = "gpt-3.5-turbo",
messages = {
{role = "user", content = prompt}
}
})... read more

Im loosly on the task fixing some bugs now and porting patches for the next release but im slow because of lack on concentration.

Things done by now:
directory restructuring. The user folder contains all props and api files needed to start scite like virgin mainline., the rest resides in the opt(ional) directory. If you have problems starting scite, just rename the folder.
the menus are white and much smaller. Also, similar property files, for example xml,json and yaml have been joined so the option menu is shorter now.
scite has its lua interpreter back
a save crash, has been fixed
theres a beautiful new theme white in sciteuser.properties
some bugs with ctags based colouring have been fixed. its compatible with mainline now
build batches have been enhanced and store logfiles
scite and scintilla are buildt with cpp14... read more

Hey Folks !
sadly for anyone, the pandemic continues and so does the Lockdown. Like many of you, im also often sittin here in my room. As i cannot afford to do some physical renovating work here, i got in touch with myScite again. Im using "autocomplete" from lexikos since some time now and noticed, that it sometimes bugges out of function for no immediate visible reason. Some testing shows, that it has problems with newer Scites luaextension. SciTe provides a per buffer lua table useful to keep buffer specific data. As it turned out, that table isnt fully initialized sometimes and accessing it then doesnt work so protected calls to it with a silly type(buffer)=="table" as it can be either nil or string in some cases. Edit: Fixed in current Git
Then, i wanted to use one runtime packer for exe/dlls either 32 or 64 bit and found petite which can be run from commandline and doesent add much to loading times. Currently its used to compress some of the more space demanding lualibs in myscite with so far no problems. i link it in here in case someone else searches for one.
LbnL i was wondering if currend manlines SciTe could handle Steve Donovans SciTEDebug written quite a while in the past now and was able to implement some smaller fixes to it as well as for extman.lua and put it up at https://github.com/arjunae/SciTE.debug Since some peops put a good effort in it and its capabilities ( Lua and a gdb interface ) , i think its nice seein it in a functional shape again. So far, so good keep yourself safe out there in the wilds and lets hope all the chaos out there comes to an end soon.
Have a nice day,
Thorsten... read more

Hey there - Sometime ago, i was given some VCR-Tapes including X-Files,Horror Aligator, Evolution and other cult movie stuff.
Since then i was searchin for a recorder so i could enjoy all the Boomer stuff. A bit later, i found an 6Head VCR, but it failed to play them. So i had a look at it and was able to repair it by changing the rubberband which drove the mechanics. Also, i carefully cleansed the Video and Audio Heads.Took the time to write a Blog post about the topic. So if you are interested seeing some Pics come here 😎 https://arjunae.wordpress.com/2020/12/23/seg-5301/
Hope you had a happy Xmas - Healthy new Year everyone !

Yesterday . Germany within another Lockdown. On Halloween. Which meant for me -sitting at home alone. No Halloween Parties, and all Cinemas closed. Gn.... So -what to do ? While greppin through my usual set of www sites, i found out that, for quite a while now, Windows Dos CMDs scripts are finally able to do simple dereferencing of Variables using >>setlocal DelayedExpansion. Since Dos Batch scripts used to be quite limited in functionality - that Info catched my attention. And since it was still Halloween time - i started playing with the stone Age old "Interpreter" and came up with a little cmd script.
It has some knowledge about the format of possible commandline arguments and so it understands Params in Quotes, the most common option prefixes "/" and "--" and values at the beginning and the end of the args string. Felt quite strange doin it. Anyway -Decided to publish it in case someone else wants to play with it :)) Edit 12.11.2020 Fix case calling the script with no parameters / properly handle args spaces. 30.08.2022 cleaned and simplified

@ECHO OFF
@chcp 65001 >NUL 
setlocal enableextensions enabledelayedexpansion
REM !Halloweeen%Jahr%! erstellt 31.10.2020 - 4.11.2020
REM Beispiel für Dereferenzierung von sowie Stringfunktionen mit BatchVariablen  :)
REM Parse cmd Parameter in LIFO Art
REM Damit kann der Parameter mit den Daten entweder am Anfang oder am Ende sicher identifiziert werden.
REM Versteht Anführungszeichen und Optionen mit den Vorzeichen / oder - oder --.   
REM
set /a phNr=0 & set ph= & set UserParam=0 
:start
if ?"%1" equ ?"" goto psHilfe
if %UserParam%==1  (set /p params="Parameter? ")  else (set params=%*)
REM lese alle Parameter in nummerierte stackvariable %ph% ein und speichere die Gesamtzahl in phNr
REM Die Stackvariable wird danach mit !! aufgerufen.  
FOR %%A IN (%params%) do (
set /a phNr+=1
set ph!phNr!="%%A"
)
Echo !phNr! Parameter gefunden
set /a phCnt=!phNr!
For /L %%K IN ( %phNr% ,-1, 1) do (
set /a phCnt-=1
set psOpt=
REM echo Behandle Parameternummer %%K !ph_%%K!
REM Entferne Anführungszeichen, suche nach Kenzeichner '/' oder '--' an position 1
set ph%%K=!ph%%K:"=!
if "!ph%%K:~0,1!" equ "/" set psOpt=!ph%%K:/=!&&call :psOpt
if "!ph%%K:~0,1!" equ "-" set psOpt=!ph%%K:-=!&&call :psOpt
if "!psOpt!" equ "" echo Lese (Wert ^!ph%%K!^)  
if !phCnt! equ 0 goto :psFrage
)
goto :k
:psHilfe
echo Parameterformat: Programpfad /Option1 --Option2 Wert /Option3 ^(...^) )
goto :k
:psOpt
if "%psOpt%" neq "" Echo Lese (Option %psOpt%)
if "h" equ "%psOpt%" call :psHilfe %psOpt%
if "help" equ "%psOpt%" call :psHilfe %psOpt%
goto k
:psFrage
CHOICE /C JN /M "Weiter ?"
If %ERRORLEVEL% EQU 1 (set UserParam=1 && goto :start)
:k

Hello Peops - found a new "to be repaired" Gadget which was dumped from a neighbour.
Its a HP Officejet 6500A and its fixed - after working half a month on it. It had a lot of scratches on one side and so it obviously felt onto the floor, which the mechanic in it didnt quite liked at all. A gear dropped out of its hole and even if i was finally able to put it back where it belongs, im not sure it suffered in more ways, as some plastic in it broke too. But - anyway it prints again, at least for the time beeing. Made a small Blogentry so interested Persons can see some pictures. Have a nice Time:
https://arjunae.wordpress.com/2020/07/14/hp-officejet-6500a/

@webgrep Stanford engineers have taught MARTY, their driverless DeLorean, to drift through a kilometer-long autocross course with the agility and precision of a human driver. comment - Powerful, plain c magic. Merry Xmas. https://www.youtube.com/watch?v=3x3SqeSdrAE.

Hello readers -
Switched my isp 4 Month ago and wondered about their network management policy. Initially their DSLAM was configured to provide 90d/40up. Then, it dropped to 30d/12u. At first i thought that my local installation introduced too much attenuation and fixed the cableing, which resulted in attenuations values dropping from 15db (ok thats really not okay) to round about 9db, which should be fine.
But since then - their DSLAM Configuration stayed stale at that reduced rate. Maybe they do configure uplinks - manually, depending on error rates ? Hey - we live in the 21Century ? nah -.. dont get salty, as their service isnt expensive compared to the competition.... Edit: [3 Weeks later] Finally, the link runs stable at acceptable rates for over a week now.
Anyway. Doing another small project. A shell script for my FritzBox, that will allow it to use OpenWrt mips packages. Oh. and yes- correct - in shell, not in bash or zsh, so it can be used with busybox environments. Its my first bigger work in shell scripting - so dont exept wonders.
See https://gist.github.com/arjunae/b77b6eb1857c78ce14af40b38f081131
Basically, it replicates some opkg functionality and uses that knowledge to provide a bootstrap environment which supports using opkg to download other packages later.
Have a nice Time,
Marcedo

Hey Folks,
my stuff Box keeps requiring my attention. This Time, because of a broken, cheap eCigarette, whichs Battery refused to function after a Year.
Since im always trying to fix those Situations before im willing to throw stuff away - i finally managd to open its Case in an effort to measure the Bats Voltage and the Cigs current requirements. It showed to need 4Amps at 4.2Volts with the bat having only 1500mAh. Wondered why that even worked that long.
Luckily mentioned Stuff box had hardware with a 12 trafo in. So i planned to use that one instead of a battery.
My limited electronic skills told me that i need a "protective resistor" (really - correct translation Google ?) to get a suitable Voltage Divider reducing the Voltage to 4.2Volts. As the eCigs Heater Coil had a 1,5Ohms Resistance , math stated the 2.Resistor should have 0.8Ohms.
Thanks. And where to conjure that littl beast ? Ha! Dismounted an older eCig-heater and gained the heater Coil. Then shortend it so it had a lower Resitance and finally woundet it around a match for physical stability. GOTCHA. Now i can smoke stuff again.
Add: Inspired by the success, i also revived another eCigaret, which had a broken Battery. Instead - i used initially developed for a smartphone which i had in spare.
Pictures will be posted later.... read more

Dear Readers!
So - Following a full 2 Month timeout because of a heavy tendonitis, im happy to report back that "Im Good again".
Im convinced that, even it was a forced timeout - it finally suited me well. So i strongly recommend doing such IT-fasting sometimes.
recently i started to play with stuff again. A Net provider change resulted in strongly enhanced Uplink Perfomance but my router showed to have regular problems. Fortunatly it provides enhanced Support logfiles. Those included a stacktrace showing it suffered from page faults - memory allocation issues.
And those ones are indications to Hardware problems. After some fiddeling i remembered the hot german weather phase with temps around 40Degrees.
Bought the Router more then 7Years ago - so i checked the Power supply. It felt really hot and had a brown colouring. I found another one lying around in my stuff Box with doubled Power capacity written on it. After switching that against the previous one - Situation got better, as the reboots seemed to be fixed by now - but some connectivity problems were currently not completely resolved. So im inspecting the issue further. Upd: Just Switched from AVMs to a freetz image. If your interested - i noted down the steps needed to do German- here: gist
Next Job -- I wanted to use a Linux VM again with Win10 1903. But Hello -Compatibility ? Its Device/Credential Guard has to be deactivated so external provided Hypervisors can be used. But- all linked Solutions inclusive a MS provided Script failed. /Hrmpf. After a long search it was revealed that its necessary to manually ensure deactivating Win10 internal hypervisor role with: bcdedit /set hypervisorlaunchtype off also some manual modifications to Microsofts script were needed.... read more

Heyo... As my tendonitis annoyes me - i do avoid to code much at the moment an so searched for other stuff to do.
And so - i found a little bit dated Lumia735 within my rooms - [Stuff Box] -
Well I think most geeks have , or should have ;) , such a box containing Hardware to be played_with_now_or_later but that point might be left to decide to the reader. Then - yesterday- i finished updating the Device to Win10 Rev 1703 using a registry tweak. Wrote a little howTo (German) here if your interested.... read more

Hey there...
1) Did some changes to the releases Dir.
Before - we had a all-mixed situation within the Release Folder which makes up 4 Files per Pack and even 6 with (planned) linux builds.
Seemed to be a bit clumsy and so I thought about a neater Release folder structure. No big change- just that subDirs within the Release folder were named following an [OS][ARCH] scheme and im quite happy with it :)
2) In reallife - I just dont really like that "just trash and buy new" mentality.
So Im thinking about (edit) update packs for myScite.190.xx. At least for lpeg.lexers, languages and lua-mods which shouldnt be much work.
But i like to keep it flat - just thinkin about once per season or so.. At least currently seems to be a good point in time as the devel branch has reached a state where im -again quite happy with its contents. ... Will check that in a few days. Edit: For that idea to work - some stuff has to be unified between versions. So there will be an update which prepares all 190.xx Packs. Planned for within June. Edit2; done within local trees scite-lua and lua dirs, still some further testing required.
Tho

Last Weekend was filled with adding a new Lua Extension addition. Started with a boring Friday evening with nothin either goin on TV or Twitch.
Before. myscite was able to do some simple http requests using luasocket. But - hey, nowdays you won't come far with that one because most dudes use SSL. Okay- one could get around by using a simple vbscript wrapper, at least thats done by myscites update checker, but like any other Nerd around - i just wanted to spend my time doing Nerd stuff that evening.
At first i learned that getting precompiled, current and working static libs for Openssl (mingw / vc / 32/64bit) can be challanging. And - fetching and compiling OpenSSL isnt really a problem. (which i thought it would be) For Mingw i just used the MSYS2- and for Windows Perl/NASM/Nmake chain.
The Pitfall was to write a nmake makefile for LuaSec - something i didnt do till now - and a rewrite of the gnu makefile so it works when called outside of MSYS2.
finally learned, that 1) invoking LD via GCC is different from invoking LD directly. - Reason: MSYS Gcc search path knows about Windows Pathes, whilst LD alone only knows MSYS internal ones. 2) The luasec build with openssl MT based libs requires ws2_32.lib and - user32.lib + advapi32.lib .
and now ? i can do in lua

status= https.request("https://www.google.com/search?q=myscite") 
print(status) 
results in : https://www.google.com/search?q=myscite [HTTP/1.1 200 OK]

For what ? i dont know. but Hey- its cool and working -hahaha.
ED: Moved 'enhanced' LuaExts to github - See here... read more

ola !
Did i already tell you, that im currently employed at a local circus helping with janitorial tasks?
Its nice work, with nice peops helping young people learning artistic stuff so their self esteem improves when they play for a bigger audience later.
For me, its helpful in reconvalescense and im planning to visit them sometimes, when i quit in some month.
Sadly, i do think that i wont be able to dedicate much time working on myScite, at least for a period of some month.
So, while porting the patches to mainline-scite-4.0.4 im also fixing some rough edges within the tree, so everything has a good shape before the "timeout". In case you are interested: see gitHub Repository) .

Update: Last week, i focussed on revisiting the makefile lexer. I didnt touch it for about some Month and tought it might be enough time to read the code with some healthy distance. Seemed to be a good tought :) . Fixed a possible Assertion that had a chance to occure, when code could try to write one byte behind the current styling blocks end. To fix that - we now have a central function responsible for communicating styling Info with Scintilla protected by some sanity checks.
after that, i did some regression testing and further optimizing. Mostly for user and automatic Variables which shoud behave much better in "convoluted" situations now. Finally, LexMake will now style the "Key" part in a "Key=Value" Snippet as an Identifier, given a clearer View on Assignements in makefiles now.
Also, its performance will be better now, because it will only ask Scintilla to begin styling the current Block at an earlier Place, when the line effectively is a continuated and not a "normal" one.
Since the changes made are sort of Important for stability and usability - mySciTE.190.01 & mySciTE.190.01.02 will be updated to contain them within the next days.
Update2: Decided to also do a CodeCleanup. Also found that it might be nice to improve Number Styling to also support simple Version schemes as 1.0.1. Think about commitin upcoming weekend. Also Fixed Lua5.1 VC Build. ->Done... read more

Hey Folks...

just finished rebasing the patches to Mainlines SciTE 4.0.4 which happens to be the first Version depending on C17 Features - particular in String::View.
For now i just reverted the relevant lines to use 4.0.3 Versions of them so everything compiles again. Later, those Functions will be reinterpreted so they compile using C++14. Yay - MySciTE.200 will require at least a C++14 compliant Compiler with 190.02 beeing the last Version compatible with c++11.
Im quite sure that wont be much of a problem and will ease further rebase works. As soon as stuff matures locally - we will see it arrive at githubs repository.
till then -
bye, Marcedo

Hello there.. slowly movin torwards finishing the todo list forthe next Release.
MartyMcFly.190.02 will be another but also the last planned incremental Update.
1) Finally we have an early versioned update Checker.
As for now, its simply fetching githubs readme.md to compare the contained version Information with either scilexer.dlls CRC32 Hash or the Release Name and Platform Identification. Behind the scenes, some logic within common.lua will trigger when the last changed timestamp of the last fetched readme.md was at least 4 Days ago. If that version has been marked for an update, it will print a friendly message within the output pane containing a link to sf.nets project release page.
So far so simply. But seems to do the task it was designed for.
2) Also cleaned all remaing compile warns for win32 and win64.
3) lua5.1 support for SciTE.3.7.5 / Foicica Scintilla 3.8.0. Backport.
4) Noticed and fixed a bug within lexMake preventing it from styling the first line.
5) Mentioned in aprevious post - php 7.2 keywords 1 2 3 4 (and also lua reserved Keywords) have been updated. Also recreated stdC Keywords up to c11.
If everything works as excepted, we will see a new wscite.cookies3.7.5 release followed by another one for 3.7.5/lua5.1 in some weeks.
So - seems that most of the goals have been accomplished -
Preview; mySciTE.120 which will be based on SciTE 4.0.4 and might recieve the Name "Mystique" ed: or "Torchwood" depends on my mood
Happy WinterTimes - Marcedo

Just stumbled upon that rather strange 90's Song... https://www.youtube.com/watch?v=68ugkg9RePc

Last week, i was thinkin about automatically wordWrap SciTE's Calltips. Currently, one can define calltip.abc.useescapes.
After that, newlines \n entered within an ApiFiles entry leads to a newline within that Calltips content.
Would'nt it be far more nicely to optionally let scite decide automatically where to put such a newline, based on the function declarations length ?
First - I started a Websearch - because sometimes such generally nice Ideas are already implemented in other SciTE Forks, givin hints on what has to be done. At first sight - SciTE.ru seemed to allow such a Feature, so i took a closer Look at their way on handling stuff which effectively adds 3 new Scintilla messages, one of them handling a fixed position calltip-wrap position. (not included, but here just for a reference ).
What i came up with now can be found here It does what its designed for, but as im hardly a C-Novice - dont expect perfectness and - it might be improved in the future. Its build around a wordWrap Function published by Sean Hubbart at c-programming.com. It wraps a Functions Documentation Entry based on the size of that functions Declaration. Adds; 3 SciTe Environment Vars:
calltip.*.use.manual.escapes has been added to preserve current behaviour.
calltip.*.max.wrappos allows us to set a Maximum Width that a Calltips function Documentations Text can extend to.
calltip.*.show.funcdocs can be set to Zero, effectively showing only the functions declaration and ommitin the Documentation.
License: BSD-3Clause
As always - feel free to use the code, but dont sue me if it bites your stuff :=)
Happy Autums-, Marcedo

Did another Fun Project, which might will recieve the name Pumuckl
Its a youtube video / playlist fetcher, allowing ad free contiguos playing via vlc. written mostly in php, just to try the language.
As a side result, i did another vbscript work. Not very amusing, but it finally grabbed php,net's function reference for the core, bundled,external and pecl api sets. Currently i only use the functions (not the classes) but its quality is far better then what was in previously. see: php_api
Add; Also refined vbs_11_18.api and theme.grey... read more

Hey Folks .. Within the past half Year there had been numerous changes , additions and fixes within myScites Language Support and Theme systems.
So decided to schedule a maintenance Update for MartyMcFly. Just Uploaded the updated Sources which functionallity will be tested during the next days.....Okay, nice One - Committed
cya, Tho

holla - as theres already an RubyEmbeddedScite project - https://github.com/sdottaka/mruby-bin-scite-mruby The patches still apply with some offset :)) happyHacking if your interested....

Yo... Since last Weekends Work with Project_2 resulted in the ability to log into the Games Main screen again - im quite pleased with its current State.
So only some minor Stuff will be done that Weekend.like extending the test class in a way that allows me to gain more knowledge on how the MVC Components play with each other. Oh - just clarifying - its really far from beeing "playable" as that would need one ore two full weekends of tedious fiddeling. So -Its done as its at a point what it was ment for -beeing a "relaxing" little dive-sight-seeing adventure within the matter.
Update: Anyway - who really would want to learn Actionscript these Days... such dropped project2.
Happy Grey Clouds - Tho

Just finished the Plan to use SciTEs Lua state with LexLPeg. Themeing had some Issues there, because of some predefined Styles with non-neutral Backgrounds. Noticed that when using 'style.type' having a greyisch bg which doesnt fit in a dark Theme. Now it can be overridden via:
property['style.type']= props['colour.keyword'] ..","..props['colour.background']
For some Reasons, it didnt work with the LexLPeg version contained within the Scintilla Backport, but using 'Stock' Scintilluas Version did fine. Also, since the lua Project doesnt recommend statically linking lua into an executable - and as SciLexer.dll contains Lua now - I decoupled Lua form SciTEs binary. Now it uses the dynamically linked Lua from SciLexer. Some stuff is yet to be done, namely VCSupport and relinking all Lualibs. That stuff will be done over Time - specifically when Temperatures arent that high around here. So Raaiinn pleaase! ... read more

sooo.. at first: When talking about embedding flashmovies -forget nearly everything you find in Forums or even StackOverflow about the AS3 embed tag. Flex depreceated it ages ago and since then it only works via ByteArrays and a LoaderEvent. That book helped me to finally diggin through that fact - i was able to jump over some stones. Fixed;
:: loader (part one and two) up and working as they should with external loaded SWF-Animations.
:: the startscreen teaser works - also with scripted external Animations.
:: the status indikator , login and register Panels also show up. - textfield managing needs work.
-> (Add) Next Weekend plans: use a Singleton to dynamically load all SWFs in a Folder. Change the static asset scripts to instanciate that class. Goodie: try to dynamixally define the asset class properties using the singleton.
At the Moment i do like my sightseeing-round-Trip travel through Actionscripts World. It has good Docs and a nice,logical Structure. We will see how far i will get just with that hahaha:) Btw : im using another great Scintilla based IDE: FlashDevelop - with FlexSDK4.6 - and raised java.args=-Xmx1024m in jvm.config.
Sunday.. Chilled for nearly the whole day. then downloaded ApacheFlex 4.16.0 .Hmm. Seems the guys at apache didnt notice Adobe's switch to https. So, as by now (3.Juni.2018) both the Airbased and the cmdline installer fails. A quick Solution to the Situation:
-> Use the ant based installer once - and after that - find and comment the following statement in "installer.xml"
"<get src="${xml.properties}" dest="${basedir}/sdk-installer-config-4.0.xml"> "
-> within sdk-installer-config-4.0.xml - replace http://airdownload.adobe.com with https://airdownload.adobe.com
-> ... same for https://sourceforge.net/ within en_US.properties.
-> restart the installer, (ant -f installer.xml -Dair.sdk.version=2.6)
-> youll also need tools.jar</get>

Last Year, i started an effort to "test" FFDECs disassembling capabilities with a flashGame i used to play that times. It took some days, but ended when the fxed disassembly was capable to bring up the games bitmap engine rendering logged in ingame content. (Of course i never put and never will release that stuff online). --> That Projects names simply "Project_2" <-- Now, while temporarly searching for some distance to myScite and C(pp). I picked up working on "Project_2" again. (Using recently released FFDEC 11.1).
Seems that they have refactored games Code to make reversing it more difficult. (for whichever reasons) So after a complete 'headbanging Saturday' - it compiles again but im stuck in getting the MFC controller to initialize the "game" further at the moment. So @first - Thanks to the Producers of that littl Flash game. Always nice to "play" with your Content. Another big kudos flow to FFDEC11.1 Devs for their quite_hot_Stuff.
Most notably automatic reversing errors are:
-> messed up Select Case Statements with multiple "Defaults"
-> incorrectly declared functions (eg private vs overrides)
-> local-loop scoped - untyped variable Names
-> random wrong named stuff aka §§pushd §§popd and strange §§gotos
and some more - more specific stuff.
:: Movin further quite slowly now, but have managed to export Frames and Sprites in an "re"importable format. rollEyes
:: The "exported" Assets display properties are somewhat wrong, but repositioning them after the add works. Ok.
-> testcase
So - Quite Cool- but after that weekend - i was happy to , at my "1-Euro Job", doing just "simple" gardening and cleansing stuff....
Have a nice Day - Tho