Archive: What's the fate of NSIS?


What's the fate of NSIS?
Is there a posibility for NSIS to become a better language much like J2SE?

I realize that the only strange thing with NSIS is the way we write scripts. It could have been easy to use it had it been like the standard coding of programs like PHP, ASP, Java or even ActioScript, JavaScript and etc.

Is it possilbe for this nice tool to progress into that direction?

Many would be very greatful if such a major release would be implemented. NSIS would be more famous than it is today.

I am sure that since it is open source, in as much as PHP made its way to what it is, I believe it is entirely possible to help this wonderful tool make its progress to such a better fate.

Bob


Like you said, its open source, there is nothing stopping you from "fixing" the syntax and submitting a patch


I see. But I am not proficient in any programming language. I just know few things about certain langguages.

Computer Science students would love to do that. Although I love to, I dont think it is an easy task. As Ive said. I dont know much.

It is impossible for a newbie to do that. I'll have to eat more and learn more before I can submit a patch.

Thanks for your kind reply Anders.


Originally posted by oliverbob

It is impossible for a newbie to do that. I'll have to eat more and learn more before I can submit a patch.
Just a few questions:
a) What "language" you don't have to learn it to use it?
b) What a "newbie" would be doing such task as making patches for a thing that doesn't know? Because is "new" in that area, right?
I realize that the only strange thing with NSIS is the way we write scripts. It could have been easy to use it had it been like the standard coding of programs like PHP, ASP, Java or even ActioScript, JavaScript and etc.
Nsis is based on scripts, ins't those languages also based...on scripts? Which each one has their own style code; some of them aren't standard, like ActionScript, because you can't port old version code into the new one.

It's like inno....but without inno script makers ;) try that :)

I would love to see Python be used as the language of choice... :)

But I know it was already answered awhile ago that a major goal of NSIS is the small installer size, and having to ship the python layer will increase the installer size way too much.
Too bad though!


Care for the world
All I'm saying is,

Support for NSIS such as the HM NIS (for instance). Well, it has a good graphilcal user interface, and it is not bundled along with the package. If we are concerned with the size of the installer, then surely tools can be done as a separate package such as this.

Its something like someone can choose the type of package that suits his need(s). There can be one package that might include all the necessary developer's tools, there may also be another that can include fewer (the most useful tools), and there is (as always) the standalone bundle that does not depend on the rest - NSIS as it is.

This way, developers that are newbies in NSIS can entegrate their applications easily with all the needed tools, plus the great need of packaged tutorials in CHM or HTML format.

By the way, I am yet learning C and C++. I am new to Java myself, and I go as far as database integration. I am quite proficient so little in PHP, knew intermediate ActionScript 2.0 long ago and few intermediate to advance pure actionscript 3.0 so far nowadays, and knew so little javascript, and 45-50% html. But above all, Im not yet one week with NSIS.

But I am getting some progress. I am honored to have NSIS, and I would like to participate the little that I know but may be indirectly such as training others to know the shortest possible time to take to know the basics and then the intermediate and so on.

As to patches, I had no plans yet, I dont know the language it was built from/of/with yet. But I see that its possible for some, and may be especially true for a team of developers. To the best that I can judge, that is how I see things here with this project at a glance.

Bob


You don't need to be an expert in C or have a CS degree to create something. All you need is to want it and to try. Experience will come with time and success will come with perspiration. Good programs don't have to be written in C. GUI applications can be written in Java (like EclipseNSIS) or even Python. You don't have to be an expert to start anything. If you put enough time and will into it, it'll work. NSIS was my first real C project and look at where it's at today. If you really want to create a better and easier development environment for NSIS, or even tweak up an existing one, it's not about your experience. If you set your mind to it, you'll succeed. Waiting for someone else to do it will just not do.


Originally posted by kichik
You don't need to be an expert in C or have a CS degree to create something. All you need is to want it and to try. Experience will come with time and success will come with perspiration. Good programs don't have to be written in C. GUI applications can be written in Java (like EclipseNSIS) or even Python. You don't have to be an expert to start anything. If you put enough time and will into it, it'll work. NSIS was my first real C project and look at where it's at today. If you really want to create a better and easier development environment for NSIS, or even tweak up an existing one, it's not about your experience. If you set your mind to it, you'll succeed. Waiting for someone else to do it will just not do.
Well said Kichik. Its true that we shouldn't wait for someone as if we all think in that way then who will come up with the solution. And oliverbob please do not treat it as too personal. I'm very much comfortable with the current scripting style but there may be others who would prefer different scripting style. Someday if we go for a massive change....then I may feel at the first instance out of the NSIS...but then I've to learn it. And believe me on my 1st day with NSIS- was zero output. I had been assigned to create an installer for our software..but had zero exp. in installer building softwares leave apart that I had never heard of NSIS. But today I love this so much and this community of NSIS.

Present is very pleasant...hope future will be also bright for NSIS.

Regards,

Originally posted by Sheik
I would love to see Python be used as the language of choice... :)

But I know it was already answered awhile ago that a major goal of NSIS is the small installer size, and having to ship the python layer will increase the installer size way too much.
Too bad though!
I think that Lua (www.lua.org) would be much better candidate. It is lightweight but powerful (in other words it fits NSIS philosophy "...that doesn't suck and isn't huge"). It's the best option for embedded scripting and even major software players start using it.

what? NSIS + nsExec + python, doesn't work for you all? ;)


I think kichik's post (above), should be added in a sticky thread on the top of the forum, e.g. "Welcome to NSIS".


It's Welcome to Life, if you ask me.


Maybe this could be a project in University lol. I'm starting Computer Science in September.

Edit: As for kichik's post, I am quite a good example :)

Stu


Live and learn, mates ;)


Well using NSIS in an automated fashion (over 2100 targets now ,) I can tell you that if I were having to generate LUA, or Python, or PHP it would make my job quite a bit harder. I in fact use Python to do all the generation of the NSIS scripts which are built.
This works remarkably well.

While in general NSIS has been an incredible asset, I've been looking for alternative solutions which enable to construction of packages, instead of full self contained installers.
Honestly there doesn't appear to be a such a tool, as a result it appears I may even write my own. I don't really look forward to this, as NSIS does a very good job.

The only concern my group has with NSIS is error detection/reporting when we ship to customers, we've had some problems with determining why particular customer cannot complete particular installations. Also our automation spends a great deal of time doing compression over and over, when it doesn't need to. (Which we call the bane of nullsoft.) You can't precompress files, and while at first this was not a major issue, it has become such.

So this leads me to my biggest point, whatever language you decide to put on top of NSIS is pointless, ***UNLESS*** you make it debuggable with a debugger. The nice thing about NSIS is that errors in code tend to be pretty easy to find by walking through it in your head. That is not so with languages which hide code being called from the developer at compile time.