1. Is Windows testing of Eu 40B2 adequate?

Forked from Re: Eu4 win32lib, IDE

I suspect that Euphoria v4 beta 2 has not been tested adequately on Windows systems because of the lack of a suitable win32lib on general release. I use win32lib version 0.70.4a minimally modified to rename identifiers which clash with the new list of keywords. Programs using this take much longer to load on Euphoria v4b2 than on Euphoria v3.1. Consequently, I still use v3.1 unless I want to try something out on v4. I have seen references to more recent versions of win32lib, but those are not on general release in the Euphoria archive.

How many people are sticking to v3.1 or v3.1.1 because a suitable win32lib is not easily available? Surely this must be delaying testing of Euphoria on Windows systems. When a suitable win32lib is finally released, I would not be surprised to see a lot of bugs reported which only apply to windows systems.

Arthur Crump

new topic     » topic index » view message » categorize

2. Re: Is Windows testing of Eu 40B2 adequate?

ArthurCrump said...

Forked from Re: Eu4 win32lib, IDE

I suspect that Euphoria v4 beta 2 has not been tested adequately on Windows systems because of the lack of a suitable win32lib on general release. I use win32lib version 0.70.4a minimally modified to rename identifiers which clash with the new list of keywords. Programs using this take much longer to load on Euphoria v4b2 than on Euphoria v3.1. Consequently, I still use v3.1 unless I want to try something out on v4. I have seen references to more recent versions of win32lib, but those are not on general release in the Euphoria archive.

How many people are sticking to v3.1 or v3.1.1 because a suitable win32lib is not easily available? Surely this must be delaying testing of Euphoria on Windows systems. When a suitable win32lib is finally released, I would not be surprised to see a lot of bugs reported which only apply to windows systems.

Arthur Crump

A big sticking point here is that many bugs have been found since beta2 and fixed, but no one has had time to polish up the current code to make an RC1 release (or even a suitable beta3 release). Likely for similar reasons that the win32lib devs (or dev) hasn't found time to polish up and release the 4.0 compatible win32lib yet.

new topic     » goto parent     » topic index » view message » categorize

3. Re: Is Windows testing of Eu 40B2 adequate?

ArthurCrump said...

Forked from Re: Eu4 win32lib, IDE

I suspect that Euphoria v4 beta 2 has not been tested adequately on Windows systems because of the lack of a suitable Win32Lib on general release.

I'm not so sure that's right, because all the testing we do is done on both Linux and Windows, and Win32Lib just uses the same Euphoria functionality that all the other testing uses.

The one thing that is significant about Win32Lib is that it is very large; bigger than most Euphoria applications I suspect.

ArthurCrump said...

I use Win32Lib version 0.70.4a minimally modified to rename identifiers which clash with the new list of keywords. Programs using this take much longer to load on Euphoria v4b2 than on Euphoria v3.1. Consequently, I still use v3.1 unless I want to try something out on v4. I have seen references to more recent versions of Win32Lib, but those are not on general release in the Euphoria archive.

The loading time is related to the size of the library, the way that include files were being used in the library, and the different way that Eu4 does initializes things now.

ArthurCrump said...

How many people are sticking to v3.1 or v3.1.1 because a suitable Win32Lib is not easily available? Surely this must be delaying testing of Euphoria on Windows systems. When a suitable Win32Lib is finally released, I would not be surprised to see a lot of bugs reported which only apply to windows systems.

The testing on Windows systems is not being delayed. Euphoria is simultaneously being tested on Linux, OSX, and Windows. The testing on Windows is not dependant on any GUI library, and specifically not dependant on Win32Lib.

As for bugs, I doubt whether there will be many bugs discovered in Euphoria due to using Win32Lib.

However I do expect that there will be a number of errors in Win32Lib for Eu4 when I (beta) release it. This is because the changes I'm making to its architecture involve creating a number of new files in the library and I'm moving a lot of routines around to put them in more appropriate files. The new win32lib.ew file will not be very large at all, but there will be many more sub-files that make up the library.

The work to make the library a pure Eu4 library is huge as I have to do small changes at a time and retest everything. One of the things I'm working on is speeding up the library's initialization time.

new topic     » goto parent     » topic index » view message » categorize

4. Re: Is Windows testing of Eu 40B2 adequate?

Hi Derek, I disagree. In practice, because win32lib for EU4 is nor working yet, I moved back to EU3.1.1. So I don't try EU4 anymore, therefore I am not testing it am I? Win32lib is *very* popular - remember the votes/money it attracted! How many Euphoria users rely on win32lib, versus other "nice to haves"? Sorry to point this out, and I am not trying to pick on anyone, but eg a new database access method is going to be used by how many, compared to win32lib? Worth pointing out too, is that the new database etc may never be tested by me, simply because I am not on EU4! IMHO, I am not the only one!

Yes, I know there are other wrappers - cross platform too. Bottom line though, what are most users using, and are most users, actual users or have the patience for testing ? Biggest bang for the buck (euphoria progress) is a EU4 win32lib.

new topic     » goto parent     » topic index » view message » categorize

5. Re: Is Windows testing of Eu 40B2 adequate?

alanjohnoxley said...

Hi Derek, I disagree. In practice, because win32lib for EU4 is nor working yet, I moved back to EU3.1.1. So I don't try EU4 anymore, therefore I am not testing it am I? Win32lib is *very* popular - remember the votes/money it attracted! How many Euphoria users rely on win32lib, versus other "nice to haves"? Sorry to point this out, and I am not trying to pick on anyone, but eg a new database access method is going to be used by how many, compared to win32lib? Worth pointing out too, is that the new database etc may never be tested by me, simply because I am not on EU4! IMHO, I am not the only one!

Yes, I know there are other wrappers - cross platform too. Bottom line though, what are most users using, and are most users, actual users or have the patience for testing ? Biggest bang for the buck (euphoria progress) is a EU4 win32lib.

Hi

I actually partly agree with this, and partly disagree. The 'biggest bang for the buck' is going to be euphoria + win32 lib + IDE - this is what is going to attract the vast majority of users, and its a shame that the initialisation speed is so slow, and that there are one or two problems with namespaces, and routine positionings, and token name changes.

I was originally against all the sweeping changes that were made (speed was my greatest concern), but I have since upgraded my veterinary database (about 6 months ago now) to eu4 now, with no performance degradation, and many 'ease of coding improvements' - for me this is the biggest plus.

Personally I would not go back. I would urge you (and other frustrated users like yourself) to have patience with the sterling work the developers are doing to advance our language, because once you start to to use the 'nice to haves' (I wouldn't do without eu.cfg now), you will start to realise that these are what made euphoria desirable in the first place.

Incidentally, I use the precompiled ide (compiled 3.1.1 I think) to write the small programs that I occasionally do, with eu4 and win32lib 0.70.4a, and (touch wood) this setup works fine. I still use the old standard includes, and the new incudes (horror), but I'm happy.

Chris

new topic     » goto parent     » topic index » view message » categorize

6. Re: Is Windows testing of Eu 40B2 adequate?

OK, I have tried v40b2 again, and so far so good.

This is what I changed to overcome namespace issues:
a) Remove the EUINC environmental variable, rely on eu.cfg instead.
b) Replaced my win32lib v0704d with v0704a, change all non-comment references of "routine" with "routine_name" in win32lib.ew
c) Put the following eu.cfg in my app's directory

-- eu.cfg file for my app (arch2vts v400) 
-- saved in the working directory of my app 
c:\euphoria\include\custom   -- my custom stuff here 
c:\euphoria\win0704a\include -- modified "routine" to "routine_name" in win32lib.ew 
c:\euphoria\include\euphoria 


HTH for others in a similar position :)
Chris, have I forgotten anything?

new topic     » goto parent     » topic index » view message » categorize

7. Re: Is Windows testing of Eu 40B2 adequate?

Hi

If it works, then no!

Cheers,

Chris

new topic     » goto parent     » topic index » view message » categorize

Search



Quick Links

User menu

Not signed in.

Misc Menu