Test versions of TextEditor.mcc 15.31 available

New test versions of TextEditor.mcc 15.31 for all supported systems are available on the TextEditor.mcc project page.

Please try to test this version as much as possible to find possibly remaining bugs due to the interal changes introduced with 15.29 and 15.30 and report any issues on that page. Crashlogs and detailed descriptions how to reproduce the issue are welcome.

Please do NOT provide crashlogs of old versions, as the known bugs have been fixed in the meantime. This testversion is meant to find any remaining bugs, although I hope to have found those already.

@tboeckel

This version seems to be working OK with NoWinED OS4. With 15.30 it reported errors.

clint

@all

In case someone has problems running yesterday's version on MOS please download this (te1531_mos_090929_2.lha) version. It fixes a MOS-only startup bug.

@tboeckel

Well, i even went back to 15.28 because i had a strange "repeating/looping input of the last typed char with complete lockup of YAM" bug.

Is this one fixed aswell?

@Raziel

Why do you think I am asking for help? I fixed all bugs which have been reported so far and which I was able to reproduce.

Honestly I really don't know if that specific bug is fixed already. Maybe yes, maybe no. Just try the 15.31 prerelease and tell me. I am certain you far better know how to recognize this misbehaviour than me.

If it is not fixed, then please add a comment to the bug tracker on the given page or open a new bug.

@tboeckel

What is the difference between the two versions ".mcc" and "-Test"? Is it just debugging output?
Is that why they are both 3x the old size?
What about the .mcp file?

@tonyw

The .mcp prefs module does not need to be updated to catch possible bugs in the .mcc class module.

The -Test program is a complete test program which does not rely on the external .mcc file, but has everything embedded (= statically linked). This makes it really easy to modify the class' behaviour without actually modifying a working .mcc class in the MUI: directory.

@all

I just uploaded new binaries here. These offer some house keeping variables for style and color handling as I suspect the bug to be hidden there.

Please test them again as much as possible.

Additionally you should have running tools like Sashimi or having a terminal connected to the serial port to capture the debug output.

Logs may either be attached to the project page if you are logged in at sf.net or sent to me privately.

@tboeckel

Crashlog attached to the Sourceforge comments.

Although it seems you already found the culprit.
I did try your advice and turned off UNDO and since then no more of those strange behaviour or crashes

@Raziel

I just uploaded binaries with Undo/Redo handling reimplemented from scratch.

Please try these again.

@tboeckel

You did it from scratch?

...testing, no crashes or misbehaviour so far...

@Raziel

Believe me, you would have done the same if you knew how the old implementation looked like and how confusing it was.

Allan had the bad idea to exploit all possible features of the C language without inserting any comments. When Jens got the source from Allan it could have won the first price at a "code obfuscation" contest.

@tboeckel

I thought so, yes, just after my post i read your comments on the TE CIA list

Great job, keep on

@tboeckel

Using this version mostly with YAM - so far everything's perfect. Thumbs up!

@hboeckel

Unfortunately i have to confirm SinOkur's bug report in YAM

Write two lines
Mark them
Delete them
UNDO

Text won't be displayed apart from the very last char, the rest of the text below the deleted part won't be scrolled down, thus overwriting some of it.
Resizing the window will make the text appear and the below text scroll, but closing the TE window will result in a Recoverable Alert which will nonetheless halt my system to a full stop :-/

Mmmh

@Raziel

This has been fixed in the meantime. Please download the binaries again.

@tboeckel

Confirmed fixed and running...

Thanks for the quick fix

@all

New binaries have just been uploaded. A bug in the Undo/Redo handling could have freed memory twice before.

@all

Once again there are new pre-release binaries available here. As usual, please test them as much as possible.

We are planning a new release around next weekend, so it would be really nice if you could help me to find any remaining bugs ahead of the release.