Bug fixes & changes in InvokeCall.inc #225

Vizit0r · 2020-03-02T20:12:56Z

added type "Variant" to params types list
added Interface to result types list
Changed flag IsStatic for x64 for "String" type, so Strings will be passed and returned to\from Delphi methods correctly in x64.
added notice about using types-synonyms

2) added Interface to result types list 3) Changed flag IsStatic for x64 for "String" type, so Strings will be passed and returned to\from Delphi methods correctly in x64. 4) added notice about using types-synonyms

Vizit0r · 2020-03-06T07:50:08Z

@martijnlaan

carlokok · 2020-09-30T11:47:49Z

@Vizit0r Can you check the conflict? I'll try to merge it next week.

Vizit0r · 2020-09-30T11:55:41Z

@Vizit0r Can you check the conflict? I'll try to merge it next week.

@carlokok

Negative, atm i'm at work with only smartphone on hand. Will back home after 3-4 week only.

The best will be to wait, if possible. I will check it all again once will be near home pc, and report to you.

pult · 2020-10-31T15:30:15Z

Some ideas for "openarray,record,set,fpc,xe3"
https://github.com/pult/pascalscript/blob/pult/Source/InvokeCall.inc

Vizit0r · 2020-11-06T09:58:36Z

@carlokok i remember 1 thing, whick cause changes all over the PS files.
in Delphi in x64 POSIX "LongInt" type will have 8 bytes size. All other systems - 4 bytes size.
In FPC this type have constant size 4b anywhere.
same for Integer - in FPC in x64 it has 8b size, in Delphi anywhere - 4b.
My decision was to make define

{$IFDEF FPC}
LongInt4b = LongInt;
{$ELSE}
LongInt4b = Integer;
{$ENDIF}

but in this case ALL "LongInt" in ALL PS files must be changed to LongInt4b. Looks not so good.
Without this changes - all is dying in POSIX x64.

All other changes still checking. Also, will look @pult InvokeCall.inc

2) types fixes and other common changes for POSIX

Vizit0r · 2020-11-16T09:33:08Z

Discord not work for me :(
Telegram: ....

Joined.

Vizit0r · 2020-11-16T10:43:32Z

@pult i cant write you, thats not a chat, but channel.

…Ptr in x64, should be fully rewritten @MyAllMethodsHandler & MyAllMethodsHandler2) Tests for Android will be done in 1-2 days, after finish will update + attach tests.

1) added support of all 64bits platform on Delphi (not only intel x64). Also Delphi 10.4 added. 2) InvokeCall.inc fully tested (about tests later on) in Delphi on Win x86 + x64, MacOs x64, Linux x64, Android 32b + 64. IOS not tested (i have no iphone etc), and simulator is very slow). All PS types supported, in params and result. 3) Included folder "Invoke Tests", which contains tests for Delphi VCL (Win) + Delphi FMX(all paltforms) + FPC. 4) FPC failed almost all tests. I have not enough FPC skills and power to find a right way with ugly debugger in FPC Invocation, who is strong enough - finish it pls. 5) about calling scripts methods from Delphi for non-Windows systems - its still not completed, i have no time at the moment for this - but i know the way, and will do this later on. It's not the most important and popular thing, really.

Vizit0r · 2020-12-15T08:49:50Z

@carlokok i have finished all.
As described in last commit, big problems with FPC there. Need someone, who have good skills in FPC, to finish its part in InvokeCall.inc. In Delphi all works fine. Tests for checking included in folder "Invoke Tests"
All tested in Delphi, all systems, including Android 64.
Also read remark in commit about calling script methods from Delphi in non-Win systems.

martijnlaan · 2020-12-28T21:50:52Z

This is pull request shouldn't be merged as-is in my opinion: the commits contain lots of undocumented and therefore unexplained changes.

I still believe #205 and #207 should be reverted to bring ROPS back to a usable state.

Vizit0r · 2020-12-28T22:50:18Z

okey guys, revert changes, as you want.
If nobody interested in multiplatform - no problem. Sit more on Win only, with existed documented and explained (really, no) code.
For me and two hundreds of my users it works fine on Linux&Windows&Macos, thats was my goal. I try to publish my code for community, to make a pleasure for other developers, but if the problem is "undocumented code" - okey, forgot it.

Tests - its also nothing, code is undocumented.

Normally, who wants to see the explanation - read a commits comments. Have no idea, why you dont want to go this way.

Good bye, and happy new year.

martijnlaan · 2020-12-28T23:25:58Z

Just giving my opinion, it's up to Carlo what happens.

It's a fact though that your commits contain changes which don't match the commit descriptions like 8ef297c which has an additional debugger related change and 3b2cabe which has an additional change that looks like a memory leak fix. These are just 2 examples. These changes might be good and intentional but that can't be seen from your descriptions.

I've pointed this out in earlier commits by you which also contained unexplained changes.

One of the commits has a change to uPSRuntime so large that Github can't even display it.

This isn't how you're supposed to make pull requests. Here's a good explainer if you're interested: https://www.jamescroft.co.uk/a-guide-to-making-a-good-pull-request/ The biggest problem of a pull request like this one that it's way too big and doesn't focus on a specific area, making it impossible to properly review.

Vizit0r · 2020-12-28T23:52:36Z

i've not comment EVERY small change, if its not critical or important. F.e. in 8ef297c adding one more condition for checking in "LineInfo" - thats for preventing error, which my user have, very rare error, but its exists. Thats much more explanation, what its done for, than real code. Others same.

I have a lot of changes, which i dont include in my commits, because i'm not sure it can be useful for someone else.

About the rules of commiting - i've never read it, works only with repos alone or in small commands, there were no any problems with my commitment style.

Anyway, @carlokok should decide. If question only in commits styling, i can spent a lot of time and divide them again, but size of some of them cant be decreased.

zedxxx · 2020-12-29T19:06:30Z

Resolves #241

Vizit0r · 2021-01-07T16:40:39Z

Additionally, I beleive Data^[l] and Items[l] point to the same thing so this change doesn't do anything except making the 2 places inconsistent.

Thats jut to to access field in same manner, as all over the PS code.

@Vizit0r The old code where it uses Data is still present in TPSExec.Destroy. Shouldn't it be changed there too?

i dont make an aim to change it anywhere, what i see - i change to common manner.

… - decided to use RTTI types prior to "TypeByKindSize" (with caching RTTI types for corresponding script types names) 2) in uPSRuntime.pas added "Generics.Collections," to uses + fixed invokecall define in FPC 3) fixed bug in uPSPreProcessor.pas for correct parsing of non-Win linebreaks (remobjects#10 instead of remobjects#13#10 in Win). 4) Added POSIX define to uPSPreProcessor.pas

Nothing so, just some small fixes & changes. P.S. InvokeCall works like a charm.

zedxxx · 2021-11-12T07:44:22Z

Source/uPSUtils.pas

-  {$IFDEF DELPHI2009UP}
-    IPointer = NativeUInt;
+  {$IFDEF VER140UP}
+    IPointer = NativeInt;


Why do you change this definition? #233

fixed, have no idea, why VER140 was written there.
P.S. Native(U)Int coming from XE2, not from 2009.

Please, look into #233 - you should change all definition block for IPointer.

Versions prior XE2 do not have support Win64 at all, so you don't need to check for CPU64 etc.

@zedxxx

…t version only, so mess it. Thanks @zedxxx for notice.

1) added type "Variant" to params types list

63f4612

2) added Interface to result types list 3) Changed flag IsStatic for x64 for "String" type, so Strings will be passed and returned to\from Delphi methods correctly in x64. 4) added notice about using types-synonyms

Vizit0r changed the title ~~1) added type "Variant" to params types list~~ Bug fixes & changes in InvokeCall.inc Mar 2, 2020

few small changes, for compatibility InvokeCall.inc with Delphi XE8

2353005

This comment has been minimized.

Sign in to view

Vizit0r added 2 commits November 14, 2020 20:43

first part of types & defines corrections

8ef297c

1) added types to DateUtils import

1fd7061

2) types fixes and other common changes for POSIX

Vizit0r and others added 2 commits December 8, 2020 21:23

all tests in Win (x86&x64), Linux x64 and MacOS passed (except btProc…

4ed89c3

…Ptr in x64, should be fully rewritten @MyAllMethodsHandler & MyAllMethodsHandler2) Tests for Android will be done in 1-2 days, after finish will update + attach tests.

Vizit0r and others added 2 commits December 15, 2020 09:56

small changes in tests.

feaed07

Merge branch 'master' into master

12fe951

Vizit0r mentioned this pull request Dec 28, 2020

Param with type "array of const" doesn't work #241

Closed

Vizit0r added 5 commits February 10, 2021 23:59

added IsConstructor implementation, for correct work in Linux

f4e5562

Merge branch 'master' of https://github.com/Vizit0r/pascalscript

a8067ac

change in IsConstr detection.

6a7acee

fixed potential leakage on 64b systems

711384e

cumulative commit for long period.

af56107

Nothing so, just some small fixes & changes. P.S. InvokeCall works like a charm.

zedxxx reviewed Nov 12, 2021

View reviewed changes

Vizit0r added 2 commits November 12, 2021 11:39

incorrect version check, have no idea why it was changed. I use lates…

1ae3b14

…t version only, so mess it. Thanks @zedxxx for notice.

IPointer definition taken from @zedxxx code.

f9bee1b

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Bug fixes & changes in InvokeCall.inc #225

Bug fixes & changes in InvokeCall.inc #225

Vizit0r commented Mar 2, 2020 •

edited

Loading

Vizit0r commented Mar 6, 2020

carlokok commented Sep 30, 2020

Vizit0r commented Sep 30, 2020 •

edited

Loading

pult commented Oct 31, 2020

Vizit0r commented Nov 6, 2020

This comment has been minimized.

Vizit0r commented Nov 16, 2020

Vizit0r commented Nov 16, 2020

Vizit0r commented Dec 15, 2020

martijnlaan commented Dec 28, 2020 •

edited

Loading

Vizit0r commented Dec 28, 2020 •

edited

Loading

martijnlaan commented Dec 28, 2020

Vizit0r commented Dec 28, 2020 •

edited

Loading

zedxxx commented Dec 29, 2020

Vizit0r commented Jan 7, 2021 •

edited

Loading

zedxxx Nov 12, 2021

Vizit0r Nov 12, 2021 •

edited

Loading

zedxxx Nov 12, 2021

Bug fixes & changes in InvokeCall.inc #225

Are you sure you want to change the base?

Bug fixes & changes in InvokeCall.inc #225

Conversation

Vizit0r commented Mar 2, 2020 • edited Loading

Vizit0r commented Mar 6, 2020

carlokok commented Sep 30, 2020

Vizit0r commented Sep 30, 2020 • edited Loading

pult commented Oct 31, 2020

Vizit0r commented Nov 6, 2020

This comment has been minimized.

Vizit0r commented Nov 16, 2020

Vizit0r commented Nov 16, 2020

Vizit0r commented Dec 15, 2020

martijnlaan commented Dec 28, 2020 • edited Loading

Vizit0r commented Dec 28, 2020 • edited Loading

martijnlaan commented Dec 28, 2020

Vizit0r commented Dec 28, 2020 • edited Loading

zedxxx commented Dec 29, 2020

Vizit0r commented Jan 7, 2021 • edited Loading

zedxxx Nov 12, 2021

Choose a reason for hiding this comment

Vizit0r Nov 12, 2021 • edited Loading

Choose a reason for hiding this comment

zedxxx Nov 12, 2021

Choose a reason for hiding this comment

Vizit0r commented Mar 2, 2020 •

edited

Loading

Vizit0r commented Sep 30, 2020 •

edited

Loading

martijnlaan commented Dec 28, 2020 •

edited

Loading

Vizit0r commented Dec 28, 2020 •

edited

Loading

Vizit0r commented Dec 28, 2020 •

edited

Loading

Vizit0r commented Jan 7, 2021 •

edited

Loading

Vizit0r Nov 12, 2021 •

edited

Loading