|
Zugg MASTER
Joined: 25 Sep 2000 Posts: 23379 Location: Colorado, USA
|
Posted: Mon Sep 24, 2007 9:10 pm
What more does CMUD 2.0x need for Public status? |
OK, here is what I'm planning for the rest of the week:
1) Fix some of the reported bugs in v2.03 (although there aren't very many critical issues that I've seen)
2) Update the documentation to reflect all of the new stuff (that will take most of the week)
3) Planning to add the zMUD Trigger Wizard system to CMUD.
So, what else is needed to make CMUD 2.0x a Public Version? I think it's getting pretty close. Do others agree?
The only other big thing that I can think of spending time on is getting everything to compile within Delphi 2007 on Vista. This involves updating a bunch of 3rd party components, which gives the advantage of fixing a few misc bugs in various components (mainly the toolbar system). It would also get me to spend more time in Vista to make it easier to deal with weird Vista issues. My guess is that it would probably take an additional week to get everything working within Delphi 2007. And I guess it's better to do this while in BETA testing, rather than trying to do this with a Public version. So maybe this is worth the effort at this time. |
|
|
|
Seb Wizard
Joined: 14 Aug 2004 Posts: 1269
|
Posted: Tue Sep 25, 2007 12:47 am |
4) More testing.
Judging by the posts in this forum, I guess that because there are some annoying bugs (even if not critical) in v2.03, many of us haven't tested it as much as we might have. And because it might be considered an early beta many people haven't tested it at all. In particular, these apparently threading related crashes and the scrolling speed slow down look like they might need quite a lot of testing to track down. (And people aren't going to be doing much beta testing if they find a really annoying (for them) bug!) |
|
|
|
Llwethen Novice
Joined: 08 Dec 2006 Posts: 37 Location: Lancaster,Oh
|
Posted: Tue Sep 25, 2007 12:06 pm |
I'd have to agree with Seb's statement about it not being fully tested.
The slower scrolling speed means I still use 1.34 when doing critical things. 2.03 also still hangs occasionally requiring the ESC key to be pressed. I wish Cmud would pop up and say "Hey I might be stuck as I haven't displayed a new line in 5 mins."
2.03 is close, oh so close but it just isn't production for me.
When are we going to be able to stop worry about corruption? I thought one of the great things about Cmud was going to be the use SqLite so we didn't have to worry about our work getting corrupted. Instead I find myself having to be far more paranoid with Cmud and backing up work. On a daily basis someone is told to reinstall and/or redo their packages. I find it really hard to recommend Cmud because of this problem. |
|
|
|
Thinjon100 Apprentice
Joined: 12 Jul 2004 Posts: 190 Location: Canada
|
Posted: Tue Sep 25, 2007 1:53 pm |
Tragically, I must agree with the above. I've restricted myself to using only one session while I've been running the beta 2.03 version, but still I get irregular behavior at times that I can't always track down, much less report. I had another complete lock-up the other day, but attributed that to having the Package Editor open (which, of course, shouldn't be crashing)... but even aside from that issue, if I leave CMud running for too long sometimes (overnight per-say) my main output window will be completely blank and unresponsive when I return... though apparently I'm still connected and receiving text, it's simply not displaying.
Don't get me wrong, I /LOVE/ the new features of CMud 2.0x... it's just stability issues for me ATM that keep it from being ready for public release. |
|
_________________ If you're ever around Aardwolf, I'm that invisible guy you can never see. Wizi ftw! :) |
|
|
|
Zugg MASTER
Joined: 25 Sep 2000 Posts: 23379 Location: Colorado, USA
|
Posted: Tue Sep 25, 2007 7:07 pm |
Well, this is pretty depressing for me. Because it's up to you guys to help with the testing. Just saying "needs more testing" isn't anything that I can do by myself. I am already testing is as much as I can. We need more beta testers, or we need existing beta testers to use it more seriously.
1) I am not aware of any reported problem with scrolling speed in 2.03. Where is this posted? Did I miss a post about this?
2) The issue with hanging also hasn't been given enough detail for me to figure out. Last I remember I was asking people if it only occurred when the settings editor was open, and I never heard a response to that question.
3) What data corruption are you talking about? I don't see any threads talking about data corruption. I see a thread talking about an issue where settings are not getting saved sometimes, but I'm not seeing anyone saying that they are getting completely corrupted settings or databases.
The reason we tell people to reinstall or redo their packages is because some people are using settings from old beta versions that had corruption bugs. But I'm not aware of any bugs in the later versions that are causing corruption. So once again, please don't propagate "rumors" like that. If there is still a data corruption problem, then please make a post about it.
4) I've run CMUD 2.03 overnight many times here with no problems. So if your main output window is blank and unresponsive, where is the post about this problem? Have you tried disabling triggers before leaving it overnight to see if that helps? Is the settings editor open when you leave it overnight?
I appreciate the help with beta testing, but I'm frustrated that I'm not seeing posts on these problems. Unless I have missed something when I was gone, it seems like it takes a topic like this to get people to post their problems. When you get a problem with the Beta version, don't just ignore it and hope that it will get fixed in another version...POST about it! Provide as much detail as you can. Send me your settings and tell me the steps that you used to get the problem (even if it means letting it run overnight). That's what beta testing is for.
Anyway, I'm stuck in a frustrating position right now. I'm not getting enough beta testers to help find these problems, and yet you guys are telling me that it needs more testing. At some point, I'm going to be forced to call it a public version just to get more people using it. But then CMUD will get an even worse reputation for stability like it did last year. Then again, some people are always going to think it's unstable just from some bad experience with a previous version no matter what improvements I make. And once this kind of reputation spreads, it's really hard to fix.
Maybe we need to get the word out to people that this is no longer an "early" beta version. Based upon the number of items I have on my bug list, it's getting close to a Public version. So try to use it more often and post about the problems you are seeing. |
|
|
|
Fang Xianfu GURU
Joined: 26 Jan 2004 Posts: 5155 Location: United Kingdom
|
Posted: Tue Sep 25, 2007 7:27 pm |
Zugg wrote: |
1) I am not aware of any reported problem with scrolling speed in 2.03. Where is this posted? Did I miss a post about this? |
It hasn't been posted about. I haven't seen any problem with scrolling myself.
Zugg wrote: |
3) What data corruption are you talking about? I don't see any threads talking about data corruption. I see a thread talking about an issue where settings are not getting saved sometimes, but I'm not seeing anyone saying that they are getting completely corrupted settings or databases. |
The trouble with data corruption is that it could've happened ages ago and the current version's just become more sensitive to it. Observing the forums, though, it's a bit of a standard response - if someone has a problem that nobody else can reproduce, corruption's one of the major suspects, and often using a new package fixes the problem. |
|
|
|
Zugg MASTER
Joined: 25 Sep 2000 Posts: 23379 Location: Colorado, USA
|
Posted: Tue Sep 25, 2007 10:24 pm |
We probably should have never used the word "corruption" in the first place.
In zMUD, your settings can get "corrupted". This means the *.MUD file is totally messed up and cannot be read. Trying to load the *.MUD settings will simply crash the program. This is caused because zMUD simply saves a chunk of memory to the *.MUD file, and if that memory was messed up because of an Access Violation error, then it's saving complete garbage to your *.MUD file and overwriting all of your settings.
In CMUD, you never get this kind of corruption. The *.PKG is an SQLite database. You are not going to have it completely overwritten with complete garbage like in the zMUD case. When we talk about "corruption" in CMUD, we usually mean that some script has some incorrect pointer (like it points to a module or class record that doesn't exist), or something like that. The *.PKG file will still load, but some script might not be working properly. This used to be a common problem because the drag/drop and copy/paste in the settings editor had bugs in older versions that would cause this. It shouldn't be happening any more, and in many cases the newer versions of CMUD will automatically fix problems like this.
So this isn't the same kind of horrible "corruption" that zMUD had. In zMUD you could lose everything. In CMUD you might just have problems with some scripts. And now that the XML Import/Export is working, it's easier to export your entire package and then re-import it. So, maybe we shouldn't be using the word "corruption"...it gives people the wrong idea.
In cases where creating a new package fixes the problem, then it would be nice if someone would send me the old package file that had the problem so that I could determine what was wrong with it. And we should all probably stop telling people to just reinstall, even though I am guilty of this myself. It would be better if we could actually narrow down the problem and get it fixed. |
|
|
|
Llwethen Novice
Joined: 08 Dec 2006 Posts: 37 Location: Lancaster,Oh
|
Posted: Wed Sep 26, 2007 12:04 am |
Zugg wrote: |
Well, this is pretty depressing for me. Because it's up to you guys to help with the testing. Just saying "needs more testing" isn't anything that I can do by myself. I am already testing is as much as I can. We need more beta testers, or we need existing beta testers to use it more seriously.
1) I am not aware of any reported problem with scrolling speed in 2.03. Where is this posted? Did I miss a post about this?
|
Using Alt-Q for testing
Version 1.34 - Triggers on 7.2/7.2/7.0
Triggers off 4.8/4.8/4.2
Version 2.03 - Triggers on 8.9/8.9/8.7
Triggers off 6.6/6.4/6.3
I thought the problem is the way that Cmud has to allocate memory as I use a 10k scrollback.
So doing Alt-Q until Cmud stops growing in task manager then running Alt-Q.
Version 1.34 - Triggers on 5.9/5.9/5.5
Triggers off 3.6/3.8/3.7
Version 2.03 - Triggers on 7.8/7.2/7.7
Triggers off 5.5/5.5/4.8
More importantly 1.34 just feels more reactive when in large session. I hadn't reported this as the initial problems were with stability and then the line gagging problem. Toril can sometimes just blast you with larges amounts of battle spam instantly easily several hundred lines per round. 2.03 I find myself reacting a couple of rounds behind others. 1.34 I'm right in sync with my group. This is something you only see at the top end of the game and not in the newbie areas.
I'm baffled as the triggers should be quicker with the new string caches. Gagging should be faster. So I use 1.34 in high level areas and 2.03 in the safer areas. |
|
|
|
Llwethen Novice
Joined: 08 Dec 2006 Posts: 37 Location: Lancaster,Oh
|
Posted: Wed Sep 26, 2007 12:18 am |
Zugg wrote: |
4) I've run CMUD 2.03 overnight many times here with no problems. So if your main output window is blank and unresponsive, where is the post about this problem? Have you tried disabling triggers before leaving it overnight to see if that helps? Is the settings editor open when you leave it overnight?
|
I'd love to say it happens when I do X but it is so random. Sometimes in combat, sometime just talking, sometimes not for hours and then it happens. Always hitting the Esc key and the screen output catches back up.
Perhaps it would help if we had a way of forcing a stack dump of Cmud if we think it's about to pop a gasket. If that could be done then when I think the screen is stuck I could dump the stack trace to a file, then press the Esc key. Assuming I caught a problem then I could submit what was going on at that exact incident.
I wonder if Dependency Walker in profile mode could help find any problems. |
|
|
|
Seb Wizard
Joined: 14 Aug 2004 Posts: 1269
|
Posted: Wed Sep 26, 2007 12:23 am |
Fang Xianfu wrote: |
Zugg wrote: |
1) I am not aware of any reported problem with scrolling speed in 2.03. Where is this posted? Did I miss a post about this? |
It hasn't been posted about. I haven't seen any problem with scrolling myself. |
It has been posted about - I don't know if turning off triggers fixes it or not, but if triggers are slowing down for some strange reason (and that is fixed by restarting CMUD), that obviously affects scrolling speed:
[2.03] Severe trigger slow-down when pattern matching significant spaces
Zugg wrote: |
We probably should have never used the word "corruption" in the first place.
...
In CMUD, you never get this kind of corruption. The *.PKG is an SQLite database. You are not going to have it completely overwritten with complete garbage like in the zMUD case. When we talk about "corruption" in CMUD, we usually mean that some script has some incorrect pointer (like it points to a module or class record that doesn't exist), or something like that. The *.PKG file will still load, but some script might not be working properly. This used to be a common problem because the drag/drop and copy/paste in the settings editor had bugs in older versions that would cause this. It shouldn't be happening any more, and in many cases the newer versions of CMUD will automatically fix problems like this. |
Well I had a, shall we call it, "messed up" package in 2.02. I'm sure all the data in it was fine. But it still gave me an error report every time I loaded it. And that can be pretty disconcerting and make one feel that there is something wrong with it! Now, as I said, I'm sure all the data was uncorrupted, and it's probably fixable via an export/import, but it did make me feel: "I'm glad I didn't spend hours on getting all my scripts to work properly in CMUD yet." Probably quite an unfounded feeling, but I might be scarred for life from some experiences with zMUD in the old days.
Zugg wrote: |
In cases where creating a new package fixes the problem, then it would be nice if someone would send me the old package file that had the problem so that I could determine what was wrong with it. And we should all probably stop telling people to just reinstall, even though I am guilty of this myself. It would be better if we could actually narrow down the problem and get it fixed. |
I probably still have the package, but it can be reproduced by the steps I've already described:
Seb wrote: |
100% reproducable steps in 2.02, but I haven't yet reproduced it in 2.03:
1. Open CMUD and Close the Sessions Window.
2. Enter this command into the command line:
#WINDOW Tells blah
3. Turn off the Status Bar and the Command Line (I don't know which of these causes the problem) if they are on.
4. Save Session as Test.
5. Close CMUD with the red X
6. Reopen CMUD
7. Open the Test session offline - the database window pops up for a second - I don't think it should be - separate bug?
8. Crash. |
Zugg wrote: |
Anyway, I'm stuck in a frustrating position right now. I'm not getting enough beta testers to help find these problems, and yet you guys are telling me that it needs more testing. At some point, I'm going to be forced to call it a public version just to get more people using it. But then CMUD will get an even worse reputation for stability like it did last year. Then again, some people are always going to think it's unstable just from some bad experience with a previous version no matter what improvements I make. And once this kind of reputation spreads, it's really hard to fix.
Maybe we need to get the word out to people that this is no longer an "early" beta version. Based upon the number of items I have on my bug list, it's getting close to a Public version. So try to use it more often and post about the problems you are seeing. |
I think you just need to try to fix the bugs that are annoying your beta testers, which will encourage them to test more. I'm not saying you're not, because you already are! But just keep plugging away at it for a few more releases. During that time, we may track down some of the more difficult to track down ones. Remember we're not getting paid for this - I know I spent about a day testing on each of the last two versions - so some encouragement for us too is helpful! (Beta testing is not that much fun!) But don't be discouraged either: you know the public release is close, and CMUD 2 will be a great improvement. By all means, after the next release I suggest perhaps, appeal for Beta testers in the CMUD General Forum...
BTW, another thing I think should be done well before the public release, if it actually causes any problems in scripts, is automatically changing #FUNC to #VARFUNC on imported MUD files and packages prior to v2.0. (Is the CMUD version number stored in a package? I think that would be a good idea - to save the version the package was last saved with, and first created with.) Or if this doesn't cause any problems, maybe make the compatibility report text a bit clearer. |
|
|
|
Llwethen Novice
Joined: 08 Dec 2006 Posts: 37 Location: Lancaster,Oh
|
Posted: Wed Sep 26, 2007 12:24 am |
Well Cmud's response to dependency walker was Guess that was busted idea.
|
|
|
|
Seb Wizard
Joined: 14 Aug 2004 Posts: 1269
|
Posted: Wed Sep 26, 2007 12:42 am |
Zugg wrote: |
And we should all probably stop telling people to just reinstall, even though I am guilty of this myself. It would be better if we could actually narrow down the problem and get it fixed. |
Oh, and as I described in another thread, I did have to uninstall and reinstall CMUD 2.03 as it wouldn't start due to a license problem. |
|
|
|
Arde Enchanter
Joined: 09 Sep 2007 Posts: 605
|
Posted: Wed Sep 26, 2007 3:35 am |
+ to all posted:
1) Fix Automapper "memLookup: Dataset is not in edit mode." error - the only crash-causing error for me for the last 2 weeks, happening every day.
2) As suggested by someone (sorry, I can't remember) add some tool for renumbering settings priorities with step 10 again. Sometimes after editing priority values goes with step 1 and no more room to insert more settings between them. You mention somewhere on drag-drop settings in tree view to change priorities automatically, but looks it not work for me. |
|
|
|
Thinjon100 Apprentice
Joined: 12 Jul 2004 Posts: 190 Location: Canada
|
Posted: Wed Sep 26, 2007 2:44 pm |
4) I've run CMUD 2.03 overnight many times here with no problems. So if your main output window is blank and unresponsive, where is the post about this problem? Have you tried disabling triggers before leaving it overnight to see if that helps? Is the settings editor open when you leave it overnight?
I had posted this before in the unconfirmed lockups post. I hadn't made mention of leaving it running overnight, but the issue of the screen being blank while still receiving code is there. |
|
_________________ If you're ever around Aardwolf, I'm that invisible guy you can never see. Wizi ftw! :) |
|
|
|
Seb Wizard
Joined: 14 Aug 2004 Posts: 1269
|
Posted: Wed Sep 26, 2007 2:51 pm |
Tip: You should put Quote="Zugg" tags around your first paragraph. It makes it less confusing.
|
|
|
|
Zugg MASTER
Joined: 25 Sep 2000 Posts: 23379 Location: Colorado, USA
|
Posted: Wed Sep 26, 2007 5:32 pm |
I'd like to get some issues from this thread separated into their own posts so that I can track them and fix them:
Llwethen Your results show a scrolling speed difference even when triggers are disabled. So, this is different than the post from Seb about spaces in triggers causing a slowdown. Could you create a new post with your scrolling speed differences in it so that we can discuss the scrolling speed problem properly.
Seb You stated in your post about the #WINDOW problem that:
Quote: |
100% reproducable steps in 2.02, but I haven't yet reproduced it in 2.03: |
You never answered whether or not this problem still existed in v2.03. From the above statement, I assumed you only saw this problem in 2.02. So you need to confirm whether this problem is still happening in v2.03 or not.
Quote: |
automatically changing #FUNC to #VARFUNC |
That is very difficult to do. That requires fully parsing the incoming settings, which is slow and subject to all sorts of problems if there are other compatibility problems. That's why the Compatibility Report exists...to flag issues that need to be fixed. But I don't want CMUD to automatically fix these issues because in many cases it could cause more harm than good, and I don't want anyone to lose their settings because of some problem in auto-conversion. The #FUNC/#VARFUNC issue is already flagged by the Compatibility Report, and it's easy to just click on the setting name and edit the script yourself.
Quote: |
But just keep plugging away at it for a few more releases. |
Yes, that is why the first item in the original list was:
1) Fix some of the reported bugs in v2.03
That is definitely a priority.
Quote: |
so some encouragement for us too is helpful! (Beta testing is not that much fun!) |
I completely agree and I'm sorry if my post came across badly. I *do* appreciate all of the help that the Beta Testers provide. I know that most testers are trying hard to help and that some of these bugs are really hard to reproduce (which is even more frustrating for all of us). I wish there was some good way to reward those Beta Testers who consistently produce high-quality bug reports. For example, when I get crash dumps that just say "it crashed" then that doesn't help at all. When I get crash dumps that say "1) First I did this, then 2) I did this, and then it crashed" they are much more useful and helpful in fixing bugs.
Arde
Quote: |
1) Fix Automapper "memLookup: |
Please create a new post about this so that we can talk about it. I'm not getting this crash myself, so I'll need a lot more details about what you are doing in your map.
Quote: |
add some tool for renumbering settings priorities with step 10 again |
This already exists. In the Settings Editor, select the Edit menu and then the "Reset priorities" command.
EveryoneIn general, I would like everyone to remember to create a new and separate thread for each problem they encounter. Just listing them in this thread makes it impossible for me to track the problems and just creates a long and confusing thread that everyone else will ignore.
In any case, I've decided to postpone the Delphi 2007 upgrade until more of these bugs are fixed. It doesn't make sense to update all of my 3rd party components while there are still this many "weird" bugs out there. Updating the components will just add other new bugs and will make it even harder to track down some of these issues. So I'll wait and do that upgrade when we are closer to a public release. |
|
|
|
Arde Enchanter
Joined: 09 Sep 2007 Posts: 605
|
Posted: Wed Sep 26, 2007 8:22 pm |
May I ask to add the option "Run CMUD in maximized window"?
|
|
|
|
Zugg MASTER
Joined: 25 Sep 2000 Posts: 23379 Location: Colorado, USA
|
Posted: Wed Sep 26, 2007 9:47 pm |
Quote: |
May I ask to add the option "Run CMUD in maximized window"? |
First, you should always make a new post if you want to suggest new features. Otherwise I don't have any way to track them in my database/wish-list.
But regarding a "maximized" window, you should be able to just maximize the main CMUD window and then exit. The window size/position is saved with your layout, and it should be remembering whether or not the window was maximized. |
|
|
|
Seb Wizard
Joined: 14 Aug 2004 Posts: 1269
|
Posted: Thu Sep 27, 2007 1:13 am |
Zugg wrote: |
Seb You stated in your post about the #WINDOW problem that:
Quote: |
100% reproducable steps in 2.02, but I haven't yet reproduced it in 2.03: |
You never answered whether or not this problem still existed in v2.03. From the above statement, I assumed you only saw this problem in 2.02. So you need to confirm whether this problem is still happening in v2.03 or not. |
Sorry, I assumed that you would know that if I had reproduced it in v2.03 I would have said so. So, no, I haven't reproduced it, but I am slightly concerned this bug may reoccur on another later version since there was nothing in the Changes log about fixing it. I noticed in v2.02 that it always opened the database window after I opened my session, even though I would shut it each time. I think maybe you had saved 2.02 with the database window open and that was causing that crash in 2.02. Anyway, that specific bug in v2.02 wasn't my point - my point was that even in very recent versions of CMUD, settings can get into a state where you get a crash immediately on opening your session. I'll clarify that bug report with what I've put here to make it easier to track. |
|
|
|
Arde Enchanter
Joined: 09 Sep 2007 Posts: 605
|
Posted: Thu Sep 27, 2007 3:49 am |
Zugg
It is too small feature to have its own topic.
It runs not maximized but with window having size of the whole screen. Occasionally I move it with mouse and only after that have notice that it was not actually maximized. |
|
_________________ My personal bug|wish list:
-Wrong Priority when copy-paste setting
-1 prompt trigger for Mapper, Session and General Options, not 3 different!
-#SECTION can terminate threads
-Buttons can't start threads |
|
|
|
eclpmb Novice
Joined: 29 Feb 2004 Posts: 36 Location: United Kingdom
|
Posted: Thu Sep 27, 2007 7:44 am |
I think reliably saving settings is fairly important; I am using 2.03 but i nearly gave up the fifth time i entered the same (simple) trigger and found it gone the next time i used CMUD
There is definately some sort of problem that CDATA seems associated with. |
|
|
|
Fang Xianfu GURU
Joined: 26 Jan 2004 Posts: 5155 Location: United Kingdom
|
Posted: Thu Sep 27, 2007 8:10 am |
Arde wrote: |
It is too small feature to have its own topic. |
No such thing. It doesn't matter how small a bug or idea is, it needs its own topic. Post, post like the wind! |
|
|
|
Seb Wizard
Joined: 14 Aug 2004 Posts: 1269
|
Posted: Mon Oct 01, 2007 1:58 am |
Zugg wrote: |
Quote: |
so some encouragement for us too is helpful! (Beta testing is not that much fun!) |
I completely agree and I'm sorry if my post came across badly. I *do* appreciate all of the help that the Beta Testers provide. I know that most testers are trying hard to help and that some of these bugs are really hard to reproduce (which is even more frustrating for all of us). I wish there was some good way to reward those Beta Testers who consistently produce high-quality bug reports. For example, when I get crash dumps that just say "it crashed" then that doesn't help at all. When I get crash dumps that say "1) First I did this, then 2) I did this, and then it crashed" they are much more useful and helpful in fixing bugs. |
What you said here (particularly the part about rewarding consistently good beta testers) made me think of this from 2006:
Zugg wrote: |
I'll be putting together a list of credits to show in the CMUD About dialog of people who have been really helpful in testing CMUD and in helping with the help files. I really appreciate any help you can offer, and I'm sure other CMUD users will appreciate it too. |
Are you still intending to do this? |
|
|
|
eclpmb Novice
Joined: 29 Feb 2004 Posts: 36 Location: United Kingdom
|
Posted: Mon Oct 01, 2007 7:34 am |
I have another problem in that when CMud does crash, the crash reporter dialogue where you enter your name and email address has the 'cancel' button only. There are two rey panels that look like buttons should go there but they're inactive.
|
|
|
|
Seb Wizard
Joined: 14 Aug 2004 Posts: 1269
|
Posted: Mon Oct 01, 2007 8:47 am |
eclpmb, that should go in a separate topic along with what version of CMUD and what operating system you are using. And please give details of what causes the crash.
|
|
|
|
|
|