YesNoOk
avatar

New build of the .95 series is up. (Read 69871 times)

Started by Nightwish, May 20, 2005, 00:46:52

0 Members and 1 Guest are viewing this.
Share this topic:
#51
Re: New build of the .95 series is up. |
May 21, 2005, 21:08:23
When I try to ctrl+f4 to close tabs, I get a crash with this error: 'MIRANDA32 caused an invalid page fault in
module RICHED20.DLL'

In the attached png, there's a lot of whitespace in the tabs.  Could the # of characters shown in the tab be changed dynamically to match the tabsize, but show no less than X characters?  Or is that a bad idea/too hard to implement?

Whatever was changed in avatar displaying has fixed the avatar mem leak I found on 9x!  Is that fix going to be added to SRMM or do you know?
#52
Re: New build of the .95 series is up. |
May 21, 2005, 21:16:45
OK... so there isn't any way to make new tabs appear on the second row as they do it on the first row? If there is not, then the free-space-between-tabs-on-second-row bug should be fixed (the one when there is 0px space) and maybe the labels should be centered, so there is not that much useless space on the right side of every tab. Just my opinion. It looks odd.

I don't wanna use the single-row thing because with about 6-8 tabs opened, the nicknames are "cut off" :-\ And i don't like that.

And what about the border color, is that one possible to add? And if you are already working on it, don't forget to separate colors for hovered border, pushed border, normal border and unread border :D (oh I'm just a big feature-requester :))
__
;D Sorry for my English ;D
#53
Re: New build of the .95 series is up. |
May 21, 2005, 21:21:46
OK... so there isn't any way to make new tabs appear on the second row as they do it on the first row? If there is not, then the free-space-between-tabs-on-second-row bug should be fixed (the one when there is 0px space) and maybe the labels should be centered, so there is not that much useless space on the right side of every tab. Just my opinion. It looks odd.

Please tell MS to fix the tab control :) I'am not going to do it.
__
Every program has at least one bug and can be shortened by at least one instruction -- from which, by induction, one can deduce that every program can be reduced to a single instruction that doesn't work.
My SMF-based forum fork
#54
Re: New build of the .95 series is up. |
May 21, 2005, 21:34:02
When I try to ctrl+f4 to close tabs, I get a crash with this error: 'MIRANDA32 caused an invalid page fault in
module RICHED20.DLL'

Thanks, fixed. It also happens on xp...

Quote
In the attached png, there's a lot of whitespace in the tabs.  Could the # of characters shown in the tab be changed dynamically to match the tabsize, but show no less than X characters?  Or is that a bad idea/too hard to implement?

Nope. The problem should be fixed by now. The tab control was using different fonts for calculating the available space and for finally drawing the labels. That's where the additional "padding" comes from (others had exactly the opposite - not enough space to show the label).

Quote
Whatever was changed in avatar displaying has fixed the avatar mem leak I found on 9x!  Is that fix going to be added to SRMM or do you know?

I changed the code for flicker free drawing when resizing the window. That code actually uses MORE gdi resources than the old code, because it needs to allocate an additional memory device context to perform "off-screen" drawing. And it involves one more BitBlt() operation than the old code.

Now, I really have NO idea why these changes should fix a leak which I've never found (I'am pretty sure, there isn't and never was one).

Well, sometimes, Windows acts really strange :)

As for SRMM - tabSRMM is open source, the source code is in Mirandas CVS. So they can just take the code and re-use it in srmm (or scriver, or anywhere else).
__
Every program has at least one bug and can be shortened by at least one instruction -- from which, by induction, one can deduce that every program can be reduced to a single instruction that doesn't work.
My SMF-based forum fork
#55
Re: New build of the .95 series is up. |
May 21, 2005, 21:42:16
Now, I really have NO idea why these changes should fix a leak which I've never found (I'am pretty sure, there isn't and never was one).

Well, sometimes, Windows acts really strange :)

As for SRMM - tabSRMM is open source, the source code is in Mirandas CVS. So they can just take the code and re-use it in srmm (or scriver, or anywhere else).

Yes, I was asking if you knew it would happen..  At any rate I'm SO HAPPY I can have avatars in the message window with no leaks.  It 'seems' to use less mem for me, after all I was losing 30kb every time it was redrawn.

Since it wasn't an expected fix.. can other 9x testers try it?
#56
Re: New build of the .95 series is up. |
May 21, 2005, 21:55:56
__
;D Sorry for my English ;D
#57
Re: New build of the .95 series is up. |
May 21, 2005, 21:57:57
Yes, I was asking if you knew it would happen..  At any rate I'm SO HAPPY I can have avatars in the message window with no leaks.  It 'seems' to use less mem for me, after all I was losing 30kb every time it was redrawn.

I think, it can be ported to srmm easily. No idea if SRMM is using 1:1 code from tabSRMM (I doubt, because tabSRMM has different sizing modes and all that stuff), but it should be similar.

Quote
Since it wasn't an expected fix.. can other 9x testers try it?

Nope, it wasn't expected. Like I said - the old code was ok, at least on 2k and later. I really had no idea where it could leak gdi handles or other resources. Not even bounds checker found any gdi leaks. So that must be a win9x specific thing. Also, the new code uses at least 2 gdi handles more (and a few k of mem for the off-screen drawing bitmap). The only thing I *could* imagine is that Win9x doesn't like using StretchBlt() on "on screen" device contexts. The new code doesn't any longer blit the bitmap directly on screen.

But ok.. if it works better. Good. It even looks better, because it doesn't flicker anymore.
__
Every program has at least one bug and can be shortened by at least one instruction -- from which, by induction, one can deduce that every program can be reduced to a single instruction that doesn't work.
My SMF-based forum fork
#58
Re: New build of the .95 series is up. |
May 21, 2005, 22:00:48
Please tell MS to fix the tab control :) I'am not going to do it.
So you're just gonna leave it like this? :'(
For now - yes. It has nothing to do with the new tab control. Multiline button tabs are always broken somewhat. That's why almost no application is using them (at least, they are very rarely used). The only thing which helps is to use a "fixed" tab width. Then it works with multiline buttons.
__
Every program has at least one bug and can be shortened by at least one instruction -- from which, by induction, one can deduce that every program can be reduced to a single instruction that doesn't work.
My SMF-based forum fork
#59
Re: New build of the .95 series is up. |
May 21, 2005, 22:07:04
Please tell MS to fix the tab control :) I'am not going to do it.
So you're just gonna leave it like this? :'(
For now - yes. It has nothing to do with the new tab control. Multiline button tabs are always broken somewhat. That's why almost no application is using them (at least, they are very rarely used). The only thing which helps is to use a "fixed" tab width. Then it works with multiline buttons.
Ha! You always have an answer... ok. Fixed width, I like that. Now I suppose this option is not in the pre5 (yet).
__
;D Sorry for my English ;D
#60
Re: New build of the .95 series is up. |
May 21, 2005, 22:55:04
One of the very rare applications that use multiline tabs. No wonder it's a M$ tool, too. ;)
Hey Alex, maybe you should ask them for an advice?! They love open-source developers! ;D



Koki, would you enlighten me about that utility of yours for making animated gifs?I might need a hand out here... ;)
#61
Re: New build of the .95 series is up. |
May 21, 2005, 23:05:36
One of the very rare applications that use multiline tabs. No wonder it's a M$ tool, too. ;)
Hey Alex, maybe you should ask them for an advice?! They love open-source developers! ;D
Koki, would you enlighten me about that utility of yours for making animated gifs?I might need a hand out here... ;)

Yes, but they are using "normal" tabs. Not the button style tabs. Normal tabs are ok, even when using variable width.
__
Every program has at least one bug and can be shortened by at least one instruction -- from which, by induction, one can deduce that every program can be reduced to a single instruction that doesn't work.
My SMF-based forum fork
#62
Re: New build of the .95 series is up. |
May 22, 2005, 00:23:58
I can't get seconds (%s) to show up in the message log, though it shows in my template preview.  Is there a 'show seconds' option I'm missing somewhere?
#63
Re: New build of the .95 series is up. |
May 22, 2005, 00:28:12
Seems double click to close tab no longer works? Clicking three times closes a tab, is this a bug?
#64
Re: New build of the .95 series is up. |
May 22, 2005, 00:30:13
I can't get seconds (%s) to show up in the message log, though it shows in my template preview.  Is there a 'show seconds' option I'm missing somewhere?

Yes. On the message log menu(s). Some of the old quick toggle settings are still there and working (they will just "skip" the variables).

Also, the nickname toggle still works with templates.
__
Every program has at least one bug and can be shortened by at least one instruction -- from which, by induction, one can deduce that every program can be reduced to a single instruction that doesn't work.
My SMF-based forum fork
#65
Re: New build of the .95 series is up. |
May 22, 2005, 00:31:15
Seems double click to close tab no longer works? Clicking three times closes a tab, is this a bug?

Nope. Double click doesn't work with button tabs. It may or may not work, but at the moment I don't care :)

__
Every program has at least one bug and can be shortened by at least one instruction -- from which, by induction, one can deduce that every program can be reduced to a single instruction that doesn't work.
My SMF-based forum fork
#66
Re: New build of the .95 series is up. |
May 22, 2005, 02:13:52
There are two small bugs in the new build...

1. Using shortcut key to close tab (Ctrl-W) causes Miranda to crash
2. When enabled the tabSRMM tray icon, the icon didn't kill cleanly when Miranda close (i.e. The icon is still there after MIM closed)

Thanks for the great work ^^
#67
Re: New build of the .95 series is up. |
May 22, 2005, 02:20:09
It'd be better if "double-click to close tab" would work. I'm also used to that feature from my beloved SlimBrowser.
And a feature I'd also love to have, both in SlimBrowser and tabSRMM (I guess I'll send a feature request to SB's team) is the one I came up with a little earlier, shown in the picture below:
A horizontal scrollbar that scrolls the tabs and notifies the direction of a new message tab by blinking the corresponding arrow(s).



(I had to install a buggy application just to be able to create this crappy animation, so don't blame me, blame the app's creators!)
#68
Re: New build of the .95 series is up. |
May 22, 2005, 02:36:11

2. When enabled the tabSRMM tray icon, the icon didn't kill cleanly when Miranda close (i.e. The icon is still there after MIM closed)

I've seen this. It is caused by a "silent" crash at shutdown (Miranda just disappears w/o giving an error message).

In that case, the shutdown sequence does not complete and tabSRMM cannot remove its tray icon.

So, the fix is to find the reason for the "crash at exit".
__
Every program has at least one bug and can be shortened by at least one instruction -- from which, by induction, one can deduce that every program can be reduced to a single instruction that doesn't work.
My SMF-based forum fork
#69
Re: New build of the .95 series is up. |
May 22, 2005, 03:21:48
Build updated to pre6.

See changelog. Downloadlink has been updated in the first post of this thread.
__
Every program has at least one bug and can be shortened by at least one instruction -- from which, by induction, one can deduce that every program can be reduced to a single instruction that doesn't work.
My SMF-based forum fork
#70
Re: New build of the .95 series is up. |
May 22, 2005, 03:23:50
It'd be better if "double-click to close tab" would work. I'm also used to that feature from my beloved SlimBrowser.

Double click will be fixed eventually.

As for the scroll bar - no way. I already explained why. It's against the main reason of using a tabbed interface.
__
Every program has at least one bug and can be shortened by at least one instruction -- from which, by induction, one can deduce that every program can be reduced to a single instruction that doesn't work.
My SMF-based forum fork
#71
Re: New build of the .95 series is up. |
May 22, 2005, 05:39:40
At midnight I'm seeing 0 instead of 12 for %a :(
#72
Re: New build of the .95 series is up. |
May 22, 2005, 09:13:08
Seems double click to close tab no longer works? Clicking three times closes a tab, is this a bug?
You can still use middle click to close tabs (opera way :) ). Try to use this until the double click gets fixed.
#73
Re: New build of the .95 series is up. |
May 22, 2005, 09:28:40
Build updated to pre6.

See changelog. Downloadlink has been updated in the first post of this thread.
Thank you, the flat tabs are now OK. But I can't make them "smaller" - they are still 100px wide even after I created a dword value in the DB and changed it to something else than 100. Does this option ("always use fixed width tabs") take it's setting from this value?
__
;D Sorry for my English ;D
#74
Re: New build of the .95 series is up. |
May 22, 2005, 14:22:10
First bug - tabs:
I use a new flat tabs (buttons) in multirow mode with labels consist of nick and status. And I don't use a fixed length of tabs.
1st situation: I create container with sigle row of tabs, row is full. Some of contacts change it's status so that label must be enlarged. So tabs is enlarged too but cannot be fit to one row and therefore second row is created. OK. But now if I select the last contact on the second row whole second row disapears (but message log was shown correctly). While I click on any remaining tab second row apears back.
2nd situation: I create container with two rows. Some of contacts change it's status so that tabs may fit in one row. Tabs are really fit in first row, but second row still present (without any tab).

Second small bug:
If I set template of Message In/Out to "%I%S %N, %D, %h:%m:%s: %M" and then I activate Show Icons only, the space between icon and contact name is missing - bug. If I have set Show Symbols only or in conjuction with Show Icons, the space between symbol and contact name apears - it is right. If I switch order of "%I" and "%S" the space disapear with Show Symbols only.

Third small "bug":
If I choose font "@Arial Unicode MS" for message log, after parethesis (left or right) is shown space. If I chose "Arial Unicode MS" font (without "@"), all is fine. OK I thing that this is not a bug of tabSRMM, but It would be nice to disallow select "@Arial Unicode MS" font.
BTW while I test PopUp 2.0.3.94 which add unicode support I found that "@Arial Unicode MS" doesn't show all UNICODE characters but "Arial Unicode MS" does.

What is difference between fonts with "@" preffix and without "@" preffix?
Last Edit: May 22, 2005, 14:24:29 by bOOT
#75
Re: New build of the .95 series is up. |
May 22, 2005, 14:47:32
I try send long message to another Miranda-IM client and tabSRMM allow me to type 7900 characters but the target client receive only this message: "** This message was blocked by the ICQ server ** The message was too long." Same result for 7899 characters. But message 7898 characters long was delivered successfully.
It is a bug? I'm sorry if not a bug.

edit
If I quote reveived message which has a maximum length (7898 characters) the input box contain more then 7898 characters (8412 for me). After I try send it I was immediatelly disconnected from ICQ network.
Last Edit: May 22, 2005, 14:53:01 by bOOT