Ini++ v1.5 Released

Hello everybody,

See this thread to download Ini++ v1.5

Ini++ v1.5 is compatible with Ini++ in so far as if you rename it from "Ini++15.mfx" to "Ini++.mfx" it will replace the original and there will be no problems. However, as the behaviour has changed (for instance, groups no longer automatically sort) it may break programs. This is why the file name has been changed.

The changes are those that people have wanted most of all.

Please report any bugs here.

I do not have a full list of bugs which have been fixed between this version and the previous, however I do know that many of them were never reported. One of them relates to using "." in search strings to mean any letter (it now works) and another is to do with receiving floating point values (that now works too).

Any comments are welcome too.

One question will it be available for Java Mobile?

Also examples for Shuffling and High Score tables aren't working from MMF2 (they crash application) but from EXE file they are working.

Newest MMF2 Dev and Windows 7 Pro.

I'm afraid it won't be, it's written in C++.

I have PMed you about the crash.

Jax, thanks for the nice update.

Marv

Just to warn people that there are a few bugs being reported that didn't go through Beta stage. I will try and fix them and update a release sometime later next week, so any less adventurous users might want to stay away.

On the other hand, if you like the latest stuff please still feel free and report your bugs =)

Thank you for the update :) I just PM'd you an issue I found.

PM'd you the issues i've found and the very, very long and complicated source code. I'd be impressed if you really do have all the extensions required to load it >.<


Oh yeah, in short- the most critical problem I've found so far is that there seems to be at least some issue with *creating* groups in the INI file, at least in my case. So its not like its going to eat your data, but its sure as heck not letting me enter new data!

But from what I remember Francois was showing how to convert extensions from C++ to Java with ease. Also greyhill is convertering some and having such advanced extension for multiplatforming instead of traditional Ini would be awesome!

Also question what is it using for encryption and decryption, compression and decompression of files? Is INI at last secure by itself?

Oh okay if that is possible then maybe. Most of the C++ code uses standard libraries anyway. Can you find any links to that?

Compression is Huffman and encryption is something which is not so standard but fairly secure. Blowfish, etc., is probably better though.

It start at 6:40 of http://www.youtube.com/watch?v=xUnrE9t3b24 and goes through this video http://www.youtube.com/watch?v=RhBy9mxto-M&feature=related ... Hope that it will help Jax.

So I can understand that it's more secure than traditional INI and there is basic coding included, if yes it's great!

It would be really cool to have this Java supported!

A few of bugs:

If you go to build, it asks for "data/runtime/ini++.mfx" which doesn't exist. I had to copy and rename "ini++15.mfx" to "ini++.mfx" for it to work.

If you try to update an existing file, it doesn't write any new data, instead creating a group called "[]" with nothing in it.

If you try to open a second file after opening a first file, then write to it, it just rewrites what it wrote to the first file, which is basically a duplication, no new data can be added to the second file.

This has caused me great frustration, and I've resorted to using the default ini extension in the mean time.

I'm looking forward to that fix.

I'm still using the previous version of the INI object until I can locate what is causing the 1.5 version to fail to update the INI in memory (and thus not allow saving new groups).

It is definitely a bug that "Change group" with "Ensure new group" does not create a new group at all.

Jax, this version do have some sort of Checksum verifier?

I've just got back from being away after being ill, so sorry for the delays everybody.

It turns out that, as Ryan said, the installer is incorrect. It installs it as INI++.mfx in one directory and INI++15.mfx in another directory. This is one of the things causing problems for people. If you installed it probably the best thing to do is uninstall, redownload the original Ini++ and wait for the update in a few days time (hopefully).

Bruto: As mentioned in a PM, that was for Ini++ v2.0 which I decided against writing for now (I probably will for MMF3 but under a different name). I'll look into it after I fixed bugs though.

Pixelthief: "Change group" with "Ensure new group" does not create a new group at all -- if I understand it, that is not actually a bug. See the blog.

See my blog post about the changes such

Ok but uh, how do we create a group now? :S

My issue is that I need to instantiate new groups to save data at runtime, and I cannot seem to figure out a way to do that. It appears that in the previous version, taking either the action
"Change group (ensure new group)"
or
"Set item in group" (nonexistant group)

would create a new group of that kind. With this update, neither of those are working, and for the life of me I cannot really figure out how to make a new group

My code is clearly using both the "Change Group" *and* "Set value of group" functions and neither is working :( But hey, besides that one small glitch your update looks awesome! Would be nice to get that hammered out. But uh, considering those change group / set item functions previously created the groups, wouldn't it make sense to restore that functionality? It would make it more backwards compatible, and I think it makes sense in each case.

But yarr, good work!

One question, the ini++ doesn't overwrite the file it is compressing?

What do you mean Bruto?

(Just a note: The new version has been sent off and should be released in a few days time)

Sent via PM.. looking forward to the next version!

I'm hoping this is fixed, but if you have initial data in the ini+ object, and no ini file exists, using the load ini command creates a blank ini file. I thought this created a new ini file with the default data in it (please correct me if I'm wrong).

Has this been fixed in the new version?

Mobichan

Hey, this was actually by design, but perhaps you are right that it is not the best design.

I'll have a think about what to do.

I personally think the best thing to do would be to not change anyting if the requested file was not found, but just keep whatever data the extension is currently holding. I'm somewhat under the impression that MMF2 is meant to work like "if an action can not be performed, don't do anything". If the user would want to load a file but revert to the default data if that file doesn't exist, (s)he could just put an action before the "load file" action that reverts to the default data (there is one, right?).

I guess an alternative that would make sense is to have an expression what to do if the file was not found (0=keep current data, 1=revert to default data, 2=clear everything).

Niffas, the problem is "Load file" should change the current file to whatever is chosen, shouldn't it? Hence all writing afterwards writes to that file, even if it doesn't exist.

Here is what the two actions do:

Code:

---

New (now "Change current file")

        Changes the current file path
        If keeping the old data it will then autosave. Otherwise it will not
        IT NEVER LOADS THE OLD DATA

Load
        Changes the current file path
        Clears all the data
        Loads the new file with the current settings
        DOES NOT SAVE THE FILE

---

What do you think it should change to?

Maybe a new one action, "Load INI into object"? (Internally this function is called 'Load In Addition') It might be more what people want (as in it wouldn't clear the old data and it wouldn't change the current file).

Merry Christmas everyone!

I have gone back to the previous version of INI++

Marv

I don't disagree with that the writing should go to that new file that is chosen even if it didn't exist. That doesn't conflict with my suggestion at all actually, it would work very well together with it.

You can use the new filename even even if there's an expression that lets you select wether you should keep the old data, revert to default data, or clear everything in case the file you try to load doesn't exist.

Nivram: That is best. I have sent a fixed version to CT but it hasn't been released yet.

Niffas: I could make a new 'load' action which lets you choose its behaviour in a dialog? If so, what options should it have?

File is compressed? Yes/No
File is encrypted? Yes using current key/No
Change current file path? Yes/Only if success/No
-- If not 'no': Read only?
Clear all data first? Yes/No