Searching for specific values in logs and more.

Reverant · Post by **Reverant** » Fri Dec 22, 2006 7:42 am

Hi,

some ideas to ponder over Christmas:

1) I'm tweaking my ignition advance, so I'm looking for signs of knocking. I right click on the plots, select Knock -> Search for specific value. I enter -2 and it starts looking for that value from the start and finds the first occurence. I search again and it propably searches from the start AGAIN (but ignores the previous occurence) because it takes waaaay too long to find the next knock which is 2 seconds later. Can we have an option/checkbox to "Search from current cursor?". I know this is a PITA because you are propably using some generic getValue() window method, a permanent option in "Options" would do just fine! Or if you implement a new window class, a "find next" button as well, like in Microsoft Excel (or press F3 to search again).

2) Searching for manual marks (ie when hitting Spacebar in MegaTune) is tedious. Can you put a red vertical line in the blue progress bar on the bottom for each manual mark? That would make looking for manual marks a snap!

3) Last but not least, the great magical power of Undo. I have changed the wrong cell ooooh so many times...

Thanks for such a great app!
Jim

LT401Vette · Post by **LT401Vette** » Fri Dec 22, 2006 10:02 am

1) This is sort of interesting... It does search from the current cursor position, when it reaches the end it starts from the beginning, and stops when it gets back to the current cursor position. No generic method here. Everything to do with graphing was written from the ground up for MLV, starting with nothing bu blank Canvas Objects. The search function looks through the data set and then moves the cursor to the result if one is found.

What is interesting is that it is taking so long for you? With a half hour log it takes less than a second to find something at the end of the file for me. Some thing is happening different for you. I would like more detail on your version, JRE, OS, processor, memory.

2) That sounds feasible to make find next easier. Heck I could probably just leave the dialog open.

3) Undo, and interesting concept

, it shouldn't be too hard. If you are on the more recent versions changed cells are marked in red and the tool tip gives you the last saved value. But a quick CTRL-Z would be faster

First I have to fix a JRE 1.6 compatibility issue that was just brought to my attention. It looks like JRE 1.6 is adding a parameter to the msq when it is saved that MegaTune doesn't like

Reverant · Post by **Reverant** » Sat Dec 23, 2006 2:25 pm

LT401Vette wrote:What is interesting is that it is taking so long for you? With a half hour log it takes less than a second to find something at the end of the file for me. Some thing is happening different for you. I would like more detail on your version, JRE, OS, processor, memory.

MLV 2.60
JRE 1.4.2
SuSE Linux 10 (Kernel 2.6.13, KDE 3.5.5)
Intel Centrino 1.5GHz
1.5GB Ram
7200rpm, 8MB cache HDD

LT401Vette wrote:2) That sounds feasible to make find next easier. Heck I could probably just leave the dialog open.

My point #2 was for an easier way to find manual markers. Ie a red vertical line in the blue progress bar on the bottom.

LT401Vette wrote:3) Undo, and interesting concept , it shouldn't be too hard. If you are on the more recent versions changed cells are marked in red and the tool tip gives you the last saved value. But a quick CTRL-Z would be faster

Yes, I've been auto-updating, one of the very intuitive features, thanks!

Jim

LT401Vette · Post by **LT401Vette** » Sun Dec 24, 2006 5:37 am

I have a close match setup here. A 1.13 GHz (IBM Think Pad T23) with 1GBytes RAM and SuSE 10, only on JRE 1.5. Searches are still almost instant, so that can't be it.

What I am wondering, Is is possible it is the repaint that is not happening right away? How long are your searches taking, first one and next when it is near by?

Reverant · Post by **Reverant** » Sun Dec 24, 2006 11:27 am

LT401Vette wrote:What I am wondering, Is is possible it is the repaint that is not happening right away? How long are your searches taking, first one and next when it is near by?

Now that you mention it, yes, repainting generally sucks as far as windows are concerned. Ie I press close/ok on windows and I get a steady 2 or 3 seconds until I can click on any part of the main window (and it redraws the widgets/areas below the window closed). I'm going to run MLV from a shell and see if I get any messages (I assume the shell is the stderr?)

My X version is X.org V6.8.2, ATI Radeon 9100M with the Radeon (GPL) driver.

Jim

LT401Vette · Post by **LT401Vette** » Sun Dec 24, 2006 12:27 pm

Yeah, that doesn't sound right. I force a repaint .5 sec after the dialog closes just in case paint wasn't called when the window closes. But either way if you click on it before the paint, it should paint at least that section. So you may be right, there may be some type of error happening.

The output including errors are are actually all written to MegalogViewerLogFile.txt

Whether you start from a shell or not. If you add
debug=true
to MSGraph.properties, then is will use the shell as stdout.

LT401Vette · Post by **LT401Vette** » Wed Jan 03, 2007 11:55 am

I put your request number 2 in 2.61.

Reverant · Post by **Reverant** » Wed Jan 03, 2007 2:23 pm

Tested and works great, thanks Phil!

Jim

LT401Vette · Post by **LT401Vette** » Wed Jan 03, 2007 9:07 pm

CrystalBlue,
For the heck of it I put in F3 find next in 2.611

Now all that tabbing stuff may be a bit more work

Reverant · Post by **Reverant** » Wed Jan 03, 2007 11:31 pm

Phil,

it works great as well, and it is almost instant this time. Seems like the window drawing is the problem after all. I might try my luck with a newer JRE from Sun. Thanks,

Jim