Recent

Author Topic: Laz_synapse - several sources and alternatives  (Read 6569 times)

Thaddy

  • Hero Member
  • *****
  • Posts: 14210
  • Probably until I exterminate Putin.
Re: Laz_synapse - several sources and alternatives
« Reply #15 on: August 01, 2020, 09:28:29 am »
Again: good code does not need much maintenance.
( About the pending issue: ssl protocol negotiation should be from highest to low and ssl2/3/tls1.0 should be dropped )
« Last Edit: August 01, 2020, 09:30:58 am by Thaddy »
Specialize a type, not a var.

process_1

  • Guest
Re: Laz_synapse - several sources and alternatives
« Reply #16 on: August 01, 2020, 09:30:57 am »
Again: good code does not need much maintenance.

Again, synapse code is far from well wrote, period. It is way too complicated and nightmare for maintenace. Perhaps basic socket parts are well wrote, other is plain disaster. Look at ICS code and compare it! They cannot be comparable as ICS code is way better in any aspect.
« Last Edit: August 01, 2020, 09:44:33 am by process_1 »

Thaddy

  • Hero Member
  • *****
  • Posts: 14210
  • Probably until I exterminate Putin.
Re: Laz_synapse - several sources and alternatives
« Reply #17 on: August 01, 2020, 09:32:31 am »
ICS is not cross-platform. I used that too. And synapse is not complex. Period.
Specialize a type, not a var.

process_1

  • Guest
Re: Laz_synapse - several sources and alternatives
« Reply #18 on: August 01, 2020, 09:38:19 am »
ICS is not cross-platform. I used that too. And synapse is not complex. Period.

Pontless childish argumentation - no point to continue.

As a professional progammer for more than 35 years with strict professional habbits and skils during designing, implementation and optimization, I believe I know what I'm talking about...
« Last Edit: August 01, 2020, 09:45:54 am by process_1 »

Jurassic Pork

  • Hero Member
  • *****
  • Posts: 1228
Re: Laz_synapse - several sources and alternatives
« Reply #19 on: August 01, 2020, 10:31:01 am »
hello,
why not lnet ?
lnet is cross platform 

Friendly, J.P
Jurassic computer : Sinclair ZX81 - Zilog Z80A à 3,25 MHz - RAM 1 Ko - ROM 8 Ko

PascalDragon

  • Hero Member
  • *****
  • Posts: 5446
  • Compiler Developer
Re: Laz_synapse - several sources and alternatives
« Reply #20 on: August 01, 2020, 10:38:35 am »
ICS is not cross-platform. I used that too. And synapse is not complex. Period.

I have just looked at their most current code that's available for download (their 2018 release): it supports Delphi's POSIX based targets (Linux, macOS), so your point regarding cross-platform is definitely wrong.

What is the case however is that it does not support FPC and thus would need quite some adjustments to make it work.

rvk

  • Hero Member
  • *****
  • Posts: 6112
Re: Laz_synapse - several sources and alternatives
« Reply #21 on: August 01, 2020, 10:42:45 am »
I have just looked at their most current code that's available for download (their 2018 release):
So, according to the argument... ICS it's not maintained either.

process_1

  • Guest
Re: Laz_synapse - several sources and alternatives
« Reply #22 on: August 01, 2020, 10:43:28 am »
hello,
why not lnet ?

Thanks, I have noted your previous post about lnet aleady and I will take a look.

But if TRon is correct, I would apriori not want to play with abandoned libs, as I do not really have a time no more to inspect the code in details and maintaining it myself.

PascalDragon

  • Hero Member
  • *****
  • Posts: 5446
  • Compiler Developer
Re: Laz_synapse - several sources and alternatives
« Reply #23 on: August 01, 2020, 10:49:12 am »
I have just looked at their most current code that's available for download (their 2018 release):
So, according to the argument... ICS it's not maintained either.

If you look at their SVN repository you'll see that their latest change is from July 23rd and in their changelog for V8 they also mention Delphi 10.4 support, so it's definitely maintained. It's just that their webpage isn't up to date (and it's rather annoying with all its popup windows >:( ).

TRon

  • Hero Member
  • *****
  • Posts: 2435
Re: Laz_synapse - several sources and alternatives
« Reply #24 on: August 01, 2020, 10:54:00 am »
@process_1:
Let me get into detail a bit more.

It seems to have a similar faith. However, it also seems to be used more often (existing projects etc) and as such is able to count on more support.

At least I just noticed some commits where done this year and a overhaul on the tls last year. (https://github.com/almindor/lnet)

The wiki however.... sad https://wiki.lazarus.freepascal.org/lNet#News

edit: ah, PascalDragon was quicker  :)


process_1

  • Guest
Re: Laz_synapse - several sources and alternatives
« Reply #25 on: August 01, 2020, 11:15:29 am »
So, according to the argument... ICS it's not maintained either.

According to the argument...

Good designed and implemented code doesn't require maintenance often. On the contrary, badly designed and implemented code does. But if the author itself abandone the project and really no one to continued maintenance and good work, I would really not be interested in it.

But then again, non of code is perfect (there is no such thing as perfect product) and from time to time need to be modified or upgraded, expecialy such complex lib. Except if the author itself think it is perfect (even it is far from that in reality), that is another story...

Look just Mike Lischke's TVirtualTreeView - that was great component for Delphi I used almost all the time. When he stopped to maintained it, some fork appears and they ruin it. With Lazarus fork, I had no issue to convert old apps code. And now, TVTV is essential part of Lazarus, which is more than great decision!
« Last Edit: August 01, 2020, 01:17:04 pm by process_1 »

avra

  • Hero Member
  • *****
  • Posts: 2514
    • Additional info
Re: Laz_synapse - several sources and alternatives
« Reply #26 on: August 02, 2020, 08:50:44 pm »
I use Synapse and Synaser for about last 10 years and I am extremely happy with it. I have it in simple GUI applications with and without communication threads, and I have it running 24/7 in several industrial applications without any leaks, hangs or any other kind of problems which I have seen with many other libraries. Sockets, serial communications and included protocols have not changed much in last 30 years so I do not find a need for frequent updates.

Synapse is not for everyone so I can get it that someone does not like it. That's fine. Use another lib.

I am sorry if someone finds it buggy after using it for years. I would expect such person to either file many bug reports so we can all benefit, or simply switch to another lib. I do not find it logical that someone insists using buggy lib for years.

There is no library without bugs, but I find Synaser rock solid, stable (for my use case at least), and very simple to use.
ct2laz - Conversion between Lazarus and CodeTyphon
bithelpers - Bit manipulation for standard types
pasettimino - Siemens S7 PLC lib

TRon

  • Hero Member
  • *****
  • Posts: 2435
Re: Laz_synapse - several sources and alternatives
« Reply #27 on: August 02, 2020, 09:16:08 pm »
I am sorry if someone finds it buggy after using it for years. I would expect such person to either file many bug reports so we can all benefit, or simply switch to another lib. I do not find it logical that someone insists using buggy lib for years.
Another missed target audience  :'(

There is no point in doing so if bug-reports are not taken into account or around the time when pigs start falling out of the sky. That is/was the whole freakin' point of complaining. I bet you do not even noticed that OPM using sources that do not match original authors repo. So much for FOSS and supplying fixes.

Please stop using the guilt-trip approach there, as there is literally no point in doing so. Those who truly know simply understand how this really works in practise.
« Last Edit: August 02, 2020, 09:27:28 pm by TRon »

avra

  • Hero Member
  • *****
  • Posts: 2514
    • Additional info
Re: Laz_synapse - several sources and alternatives
« Reply #28 on: August 02, 2020, 11:12:18 pm »
I am sorry if someone finds it buggy after using it for years. I would expect such person to either file many bug reports so we can all benefit, or simply switch to another lib. I do not find it logical that someone insists using buggy lib for years.
Another missed target audience  :'(

There is no point in doing so if bug-reports are not taken into account or around the time when pigs start falling out of the sky. That is/was the whole freakin' point of complaining.
Do you mind pointing to URLs of such bug reports so that we can judge by ourselves? Saying that something is buggy is simply not enough in a developer forum. You should give us arguments to consider. Or you think that we should believe without questioning? Do you consider your self such an authority?

Anyway, I still did not get the answer why would anyone use buggy lib for years? That simply does not make sense to me.

I bet you do not even noticed that OPM using sources that do not match original authors repo.
You lost that bet. If you search the forum you will find that original synapse lpk did not register ssl units by default, there were SVN and official version in OPM, and there was pl_synapse which I maintained for a while to allow users to use the same project in both Lazarus and CodeTyphon. It was a mess and there was a forum discussion and desire to simplify things with a single Synapse version. If I remember well, author didn't want to make such change (which is his right and I do not have a problem with that), so LPK change was made by the community and SVN Synapse version was chosen. That change was deployed by GetMem - the author of OPM, and now we have a single Synapse version in OPM instead of 3 versions. Lazarus IDE was also much happier to see only 1 instead of 3 versions of units having the same name.

Please stop using the guilt-trip approach there, as there is literally no point in doing so.
I have no problem when someone says that Synapse has bugs. I just want to see argumentation for that and concrete examples or bug reports since I am personally interested in elimination of such bugs. On the contrary - it looks like you have a problem when someone defends Synapse. Like defending Synapse is forbidden, since you are not happy with it.

Those who truly know simply understand how this really works in practise.
Such a silly and disrespectful argumentation. Please, the enlighten ones, tell us mere mortals how this works in your practice, so that we can improve ours!
ct2laz - Conversion between Lazarus and CodeTyphon
bithelpers - Bit manipulation for standard types
pasettimino - Siemens S7 PLC lib

process_1

  • Guest
Re: Laz_synapse - several sources and alternatives
« Reply #29 on: August 03, 2020, 06:45:26 am »
avra,

Please stop this. Who claimed that Synapse is buggy? I asked "is it buggy", that some forks appears or simply they brings new features, or lack of maintenance from the author itself. Read my first post carefully.

I do claim that the code is awful and far from optimal, that no body can dispute, anyone can inspect itself. Some decisions the author made are at least "questionable", to be "politically correct" (not sharp in words), making synapse code as it is.

I have maintainined it for several years for my internal use, as the author itself refused to do anything about improving, considering it's current code "perfect". I did considered to make my own lib, but lack of free time is something anyone here know how it goes. Then I switched to ICS.

Now, with Lazarus, I simply don't want to back to the nightmare again, thus I asking here for available suitable alternatives. And that is it.
« Last Edit: August 03, 2020, 07:01:27 am by process_1 »

 

TinyPortal © 2005-2018