Visit our Popular Forums

phiggins · 08-04-2018, 00:41

OK thanks Pavel. Will wait to see for the release of next version to test it.

did-g · 08-04-2018, 01:47

Hi,

As workaround did you try to disable opengl cache? On a fast computer it shouldn't impact speed too much.

If charts are often modified another assumption will break sooner or later and you'll get the wrong chart again. Caching mostly read only filesystem objects is OK, not so for read/write at least if you want to do it quickly.

chris95040 · 08-04-2018, 05:47

Quote:

Originally Posted by nohal

Paul...
Unless your real usage is even more abusive than my tests, it is now fixed in master branch.

Pavel

Why the caveat? Why not just fix it for good?

nohal · 08-04-2018, 05:49

Quote:

Originally Posted by chris95040

Why the caveat? Why not just fix it for good?

Chris...
You probably are good at this, can you please give me the exact testing scenario allowing me to say "fixed for good", please?

Or, better, if you are convinced the fix I did is not good enough, send your own PR which does "fix it for good".

Pavel

chris95040 · 08-04-2018, 06:59

Quote:

Originally Posted by nohal

Chris...
You probably are good at this, can you please give me the exact testing scenario allowing me to say "fixed for good", please?

Or, better, if you are convinced the fix I did is not good enough, send your own PR which does "fix it for good".

Pavel

You gave a caveat, caveats worded like that generally imply "i'm not sure i root caused the issue, nor am i certain the typing i did resolved it.". Sounds like you needed to do a little more thinking before typing. Look, if you're content to play bug whack-a-mole and stitch together a quilt of half-measures I guess thats up to you.

Me, I find the work much more satisfying when I think first, plan my work as a set of small, deterministic, understandable components communicating through predefined interfaces. When bugs come along I root cause them and fix them with confidence.

I can't for the life of me understand how work you choose to do in your free time (!!!) shows such a lack of passion or enthusiasm for getting it right. Why bother doing it then?

nohal · 08-04-2018, 07:05

Quote:

Originally Posted by chris95040

You gave a caveat, caveats worded like that generally imply "i'm not sure i root caused the issue, nor am i certain the typing i did resolved it.". Sounds like you needed to do a little more thinking before typing. Look, if you're content to play bug whack-a-mole and stitch together a quilt of half-measures I guess thats up to you.

Me, I find the work much more satisfying when I think first, plan my work as a set of small, deterministic, understandable components communicating through predefined interfaces. When bugs come along I root cause them and fix them with confidence.

I can't for the life of me understand how work you choose to do in your free time (!!!) shows such a lack of passion or enthusiasm for getting it right. Why bother doing it then?

This case is trivial, so please invest that moment of your time needed to cover it the correct way and show us how it has to be done. Not by these none-saying rants, by something actually usable, please.

Also explicitly explaining what commit 1d71b8f86c50019c523c2b4c59096976d752bbab does not do right and how it should have been done with more passion and enthusiasm would certainly help me understand what your problem really is.

And you are sure right in one thing, how I decide to spend my time, is my thing.

Thanks

Pavel

chris95040 · 08-04-2018, 07:10

You are the one who gave the caveat. That 'hey maybe this will work' attitude is why OpenCPN, and so much of the software we interact with today, is so flakey.

And like I said, its up to you how you spend your time. I personally can't imagine enjoying my time wallowing in spaghetti code, but its your Sunday!

nohal · 08-04-2018, 07:12

Quote:

Originally Posted by chris95040

You are the one who gave the caveat.

And like I said, its up to you how you spend your time.

Chris...

I am so very sorry for "giving the caveat", can we now please move forward with you educating us so that we are not so incredibly pathetic in the time coming?

Thank you

Pavel

chris95040 · 08-04-2018, 07:15

Quote:

Originally Posted by nohal

Chris...

I am so very sorry for "giving the caveat", can we now please move forward with you educating us so that we are not so incredibly pathetic in the time coming?

Thank you

Pavel

Draw pictures before you type. Thats the best thing I can tell you, but I suspect you will never quite understand that lesson. Enjoy your patchwork quilt.

nohal · 08-04-2018, 07:19

Quote:

Originally Posted by chris95040

Draw pictures before you type. Thats the best thing I can tell you, but I suspect you will never quite understand that lesson. Enjoy your patchwork quilt.

This is getting more and more interesting. Can you please show me a picture that covers this problem and it's fix? I spent the last 2 decades drawing UML diagrams pretty much daily and I think they also are kind of pictures with many boxes and stuff (do you agree?), but would probably not use one here...

Thank you so much in advance

Pavel

hoolie · 08-04-2018, 07:25

Quote:

Originally Posted by chris95040

Why the caveat? Why not just fix it for good?

Quote:

Originally Posted by chris95040

You gave a caveat, caveats worded like that generally imply "i'm not sure i root caused the issue, nor am i certain the typing i did resolved it.". Sounds like you needed to do a little more thinking before typing. Look, if you're content to play bug whack-a-mole and stitch together a quilt of half-measures I guess thats up to you.

Me, I find the work much more satisfying when I think first, plan my work as a set of small, deterministic, understandable components communicating through predefined interfaces. When bugs come along I root cause them and fix them with confidence.

I can't for the life of me understand how work you choose to do in your free time (!!!) shows such a lack of passion or enthusiasm for getting it right. Why bother doing it then?

Quote:

Originally Posted by chris95040

You are the one who gave the caveat. That 'hey maybe this will work' attitude is why OpenCPN, and so much of the software we interact with today, is so flakey.

And like I said, its up to you how you spend your time. I personally can't imagine enjoying my time wallowing in spaghetti code, but its your Sunday!

Quote:

Originally Posted by chris95040

Draw pictures before you type. Thats the best thing I can tell you, but I suspect you will never quite understand that lesson. Enjoy your patchwork quilt.

For goodness sake, you're rivalling Carcode for being a miserable old sod. Perhaps it hasn't occurred to you that OpenCPN development is necessarily a whole different ballpark to a commercial software team. The team do a very good job so either make a useful contribution or just have another beer and wind your neck in ... ...

chris95040 · 08-04-2018, 07:33

Huh. Surprising then that the work you choose to do in your free time shows such a lack of design effort or enthusiasm. UML is nice, whiteboards are even better.

I imagine the drawing would involve the interaction between the rendering engine, its cache, and its assets, but I can't retroactivity provide a neat drawing for a pile of spaghetti code. Thats the point, it should have been done first, before OpenCPN support for OpenGL even had a single line of code written. Then you'd have deterministic behavior and confidence in your fixes.

chris95040 · 08-04-2018, 07:47

Quote:

Originally Posted by hoolie

Perhaps it hasn't occurred to you that OpenCPN development is necessarily a whole different ballpark to a commercial software team.

The fact that its a distributed team of amateurs makes it even more important to have a proper architecture in place and to verify contributions follow that architecture.

Quote:

Originally Posted by hoolie

The team do a very good job

Perhaps you are right! Like I said, my data points are a 2-years-out-of-date observation that OpenCPN was incredibly flakey and a single nauseating caveat from an unenthusiastic coder fixing something that, conceptually, is quite simple.

nohal · 08-04-2018, 07:53

Quote:

Originally Posted by chris95040

Huh. Surprising then that the work you choose to do in your free time shows such a lack of design effort or enthusiasm. UML is nice, whiteboards are even better.

I imagine the drawing would involve the interaction between the rendering engine, its cache, and its assets, but I can't retroactivity provide a neat drawing for a pile of spaghetti code. Thats the point, it should have been done first, before OpenCPN support for OpenGL even had a single line of code written. Then you'd have deterministic behavior and confidence in your fixes.

I do have confidence in my fixes, of course. I do not have confidence in people, any people, correctly describing their use cases.
Can we now get back to our problem of you teaching us what we are supposed to do with what we have on the table, which, whether you or me like it or not, simply is a decent size codebase with a very long organic growth history?
Throw it all away, draw "pictures" and implement a new product? I think we will have to leave this amusement to you and enjoy the perfect result while not having to waste our time on OpenCPN anymore when you publish your result.

Pavel