Board index FlightGear Development

Is there a "standard reference system" for testing installs?

FlightGear is opensource, so you can be the developer. In the need for help on anything? We are here to help you.
Forum rules
Core development is discussed on the official FlightGear-Devel development mailing list.

Bugs can be reported in the bug tracker.

Re: Is there a "standard reference system" for testing insta

Postby jharris1993 » Wed Jan 04, 2017 7:46 pm

Thorsten,

I'm rolling over laughing imagining you trying to "herd cats" (smiley laughing so hard, tears are coming out of his eyes!)

All of your points are well taken. In any event, at the very least, I would like to try to produce bug reports that go - perhaps - just a tad beyond the "Doh! It don't work!" stage. (grinning) This is not to say that all, or even the majority of your bug reports are pieces of GAGH, but. . . . I have seen bugs on Ubuntu's bug list that should be bronzed. And if there were a "Bug Report Hall of Shame. . . ." 'Nuff said!

Obviously, as you mention, I cannot compete with VA class users who have their simms on for days at a time, but I would like to create bug reports that are reasonably well characterized. It has been my observation - even in commercial software testing - that the better characterized a bug is, the more clearly the report is written, and the easier the QA analyst makes it for the relevant dev to reproduce it, the more likely it is that YOUR bug will get seen and fixed. If a QA analyst does this repeatedly, he gets a "rep", and his name on a bug report gets attention. Conversely, ill written bugs, lousy characterization, and impossible reproducibility will also get you a "rep" - that you don't want!

The same topic - sideways, sort-of. . . .

I am putting together a test platform for FG, both Linux and Windows. My main Windows test platform will be Win7 64 bit. For Linux, I am planning on a 64 bit Mint install, and am also considering a different distribution in a different partition on the same HD. What would you recommend as a second Linux test distribution? Fedora? SUSE? CentOS? Raspbian? ( :D )

I am assuming here that the typical user installs (on Linux) for FG mirror the popularity of the distro, though I welcome any input you - or anyone else - can provide here. What I do NOT want to do is to have to re-re-re-recreate installs because I got the initial set wrong. I'd like to hit the popular distro's right off the bat.

Thanks for all your very useful commentary and advice.

(I'm still howling over those damn cats. . . )

Jim (JR)

Update:

A quick look over at Distrowatch shows that Might Mint is King of the Heap!
(Yes, I am an unabashed Mint fan-boy. Viz., an article of mine over at QA TechTips about Mint: http://www.qatechtips.com/2013/03/all-h ... linux.html)

The list of the top ten looks something like this:
1 Mint
2 Debian
3 Ubuntu
4 openSUSE
5 Manjaro
6 elementary
7 Fedora
8 Zorin
9 CentOS
10 deepin

So, maybe the question I should be asking is "which distribution classes should I be using?" Obviously Mint/Ubuntu/Debian are all part of one class, Fedora and (I think) CentOS are Red Hat re-spins, etc. I'm not to sure about openSUSE - the last time I tried it, it wobbled sideways and fell off the hard drive. . . .

I'm considering a Mint/Fedora install, though I'd appreciate any advice and/or suggestions you folks may wish to offer.

Jim (JR)
What say ye?

Jim (JR)

Some see things as they are, and ask "Why?"
I dream things that never were, and ask "Why Not".

Robert F. Kennedy

“Impossible” is only found in the dictionary of a fool.
Old Chinese Proverb
jharris1993
 
Posts: 139
Joined: Sun Nov 18, 2012 9:20 pm
Location: Worcester, MA. / Moscow, Russia
Callsign: $%^&ing Idiot!
Version: Whatever..
OS: Win-7 / Linux

Re: Is there a "standard reference system" for testing insta

Postby jharris1993 » Wed Jan 04, 2017 11:06 pm

Hooray wrote in Wed Jan 04, 2017 7:32 pm:Sorry, I only read a few lines of the previous response - so just responding briefly: FlightGear is cross-platform software, so unless you are facing something during install, startup/reset, it is rather unlikely that you will directly run into OS specific issues that are not obviously related to the OS/GPU at hand (e.g. OS specific file dialogs, error dialogs, driver glitches etc).

<snip>

All of this is to say that people are generally aware of such issues and that the bug tracker and the mailing list are the primary means to provide action-able bug reports.

<snip>

Something that a few of us have independently been working towards is to include more/better information in the binary itself (e.g. via the help/about dialog) and provide additional tools built right into the binary (such as the built-in profiler).

In my opinion, this is the most promising path forward - because it enables power-users to provide valuable information much more easily, while also allowing them to interface with the rest of the user-community to help review/triage bug reports and provide useful information to people able to actually debug/fix bugs.

To learn more, see:
http://wiki.flightgear.org/CrashRpt
http://wiki.flightgear.org/Towards_bett ... leshooting
http://wiki.flightgear.org/Resource_Tra ... FlightGear
http://wiki.flightgear.org/Draw_masks
http://wiki.flightgear.org/Built-in_Profiler


Ouch! OK, OK, you win! (Hooray has such a way of cutting through the BS and putting the facts out on the table where we can all see 'em.)

I took a look at these links - that is up until the "Resource_Tracking" link where I was overwhelmed in pointers, OOM crashes, etc. I promise to go look at the rest after my headache subsides. . . :? :shock: OK?

Tell me if I'm right or wrong: It's not so much a lack of well written bug reports as it is an inability to deal with the massive amounts of data you've already collected - water, water everywhere and not a drop to drink! - right?

Of the first three that I was able to get through, the "CrashRpt" one looks like something that I might have the skill to look at - though it looks like a full-time project all it's own - even if I never fire up FG again in my life! And I'm not even sure what I'd be getting involved in.

If there were some way to index bug report data with the corresponding "PDB" (symbol table?) data, it should be possible to create a correspondence between the two. How to re-attach the symbol data to the bug's data is something I'm not sure how to do - short of installing Visual Studio, which I cannot afford.

The Wiki makes it sound like it might be doable, (Viz., http://wiki.flightgear.org/CrashRpt => "Help Wanted"), but the scope of the problem - quite frankly - scares the &^%$#@!!! outta me.

Any ideas on how I might get my feet wet there without getting overwhelmed in data? Some sample data, PDB files, and a peek at the scripts to play with? Maybe a magic wand and a whole bucket full of spells too? :lol:

Jim (JR)
What say ye?

Jim (JR)

Some see things as they are, and ask "Why?"
I dream things that never were, and ask "Why Not".

Robert F. Kennedy

“Impossible” is only found in the dictionary of a fool.
Old Chinese Proverb
jharris1993
 
Posts: 139
Joined: Sun Nov 18, 2012 9:20 pm
Location: Worcester, MA. / Moscow, Russia
Callsign: $%^&ing Idiot!
Version: Whatever..
OS: Win-7 / Linux

Re: Is there a "standard reference system" for testing insta

Postby jharris1993 » Thu Jan 05, 2017 1:45 am

Oh boy, oh boy, oh boy, oh boy. . . .

I've been thinking about the rat's nest I'm sticking my head into, and decided that I'd better take a step backward or two. . . .

Let's turn the way-back machine back a month or so, and start this at the beginning.

1. I have a couple of granddaughters who are just fascinated with flight simulators.
When my wife and I discussed this with our son and I mentioned that I had some flight sim software. . . .
"You have flight simulator software?"
"Yep. Two or three by Microsoft and an Open Source version. Even have a flight-sim HOTAS joystick and throttle controller. . . "
"You ARE bringing them the next time you come, right?"
"Um. . . ."

2. Though normally any ideas I might have about messing with flight simulator software would be classified as "just playing games" by my wife, as soon as the granddaughters enter the equation it rapidly became a drop-dead priority and generous amounts of luggage space - and weight - were allotted for anything I might want to bring that I could even vaguely relate to "flight simulator" stuff for the granddaughters.

3. I rapidly whipped up a Win7-64 bit system to run FlightSim X, along with Flight Gear and crammed it into a suitcase with the Saitek HOTAS joystick I have.

4. Because I will be, (am now at), a location where the ADSL internet speed is horrid, I wanted to create an entirely local FG scenery store, obviating the need to access an external TerraSync scenery server. To do this, I made sure I bought the entire kit-and-caboodle on a thumb-drive before I left.

5. I unpacked and installed the entire FG scenery universe to my local hard drive and pointed my FG Scenery folder to that location, (which is how it is described in the various installer fora), creating and building the "Aircraft", "Textures", "Shared such-and-so", etc. directories as specified.

6. I attempted to run FG. . . . .
(a) Running the 2016.4.n releases on a large-screen monitor, (1900x1300 or so), FG would open a large window, and then reduce itself to a small view-port in the lower-left-corner of that window, as if it were hard-coded to some smaller resolution, (like maybe 1024x768?) The 2016.3.n releases do not do this.
(b) With the entire FG scenery universe installed and configured as noted above, launching FG results in horrendous amounts of time taken "loading scenery". I usually loose patience after about 20 or 30 minutes, clobber FG with the task manager, reboot, and run M$ Flight Sim X instead. I cannot prove, but suspect, that FG is attempting to load the entire bloody scenery universe, instead of just the chunk it needs.

Heroically, rather than just crybaby on the fora that FG is a piece of junk, I wanted to do more in-depth testing - using both Linux and Windows. (i.e. Are these just Windows issues, or are they common to both versions?) Of course, the requirements-based testing that I have done since, well, a long time ago. . . . lead me to ask if there was some kind of standard reference install, spec's, etc. which opened this particular can of worms.

This rapidly snowballed to the point where Hooray, (rightfully so), mentioned that there are better places for me to waste time than trying to define test reference standards. You know, things like the relatively simple task of completely re-indexing the entire bug-database data-blob, binding it to its respective symbol table libraries, re-linking, and creating commented source-code from the core-dump files FG has obtained over the years. . . . Simple, right? Piece-of-cake.

Holy Christmas!

Don't get me wrong here. I really do want to do whatever I can to help FG along as much as possible. However I don't want to get caught over committing myself and then falling flat on my face. Helping organize and create symbol-linked bug reports is a great idea. I just wonder how big a bite I am taking if I start a project like that? Maybe I should just punt, create a Windows and Linux install base, and play with it from there?

I don't want anyone to think that the clucking noise you hear in the background is me chickening out. . . . However, I fear that if I get over-involved in what could be a gigantic project, I would not be able to commit the kind of time it would take to see it through. And I do want to get some stick-time flying FG too. 8)

Ideas?

Jim (JR)
What say ye?

Jim (JR)

Some see things as they are, and ask "Why?"
I dream things that never were, and ask "Why Not".

Robert F. Kennedy

“Impossible” is only found in the dictionary of a fool.
Old Chinese Proverb
jharris1993
 
Posts: 139
Joined: Sun Nov 18, 2012 9:20 pm
Location: Worcester, MA. / Moscow, Russia
Callsign: $%^&ing Idiot!
Version: Whatever..
OS: Win-7 / Linux

Re: Is there a "standard reference system" for testing insta

Postby Thorsten » Thu Jan 05, 2017 6:42 am

(b) With the entire FG scenery universe installed and configured as noted above, launching FG results in horrendous amounts of time taken "loading scenery". I usually loose patience after about 20 or 30 minutes, clobber FG with the task manager, reboot, and run M$ Flight Sim X instead. I cannot prove, but suspect, that FG is attempting to load the entire bloody scenery universe, instead of just the chunk it needs.


That's frankly unlikely. Do you have *any* internet connection on then or not?

I remember I found one way of getting stuck (I think it was terrasync disabled but starting at an airport that's not installed) where it would load scenery forever.

The sad truth is that we all have our commandlines which 'just work' and don't try anything else - I've never installed FG from disc, I pull from the repositories for instance, I don't try real weather fetch, I just disable the option,... so not all options are routinely tested.

You can get to the bottom of this one by starting FG with the text console with log level to debug, watch the console output and just read up what it's doing when 'loading scenery' is on.

In general: If you want to help (and that's much appreciated) - maybe familiarize yourself with the set of tools - how to change log level, where to find the logs, how to monitor what's going on on the console, how to access the commandline that's being used, how to add things to the commandline.

Then take one thing (like the scenery loading forever), make a forum topic, try the tests suggested (like switching terrasync on and off). Then that can be classified and sent onward.

Move on to the next thing.

I can not emphasize enough the importance of having someone who is both able and willing to do systematic testing in response to queries on the other end of a bug report - I assure you this alone will make a big difference.
Thorsten
 
Posts: 9994
Joined: Mon Nov 02, 2009 8:33 am

Re: Is there a "standard reference system" for testing insta

Postby erik » Thu Jan 05, 2017 9:21 am

Thorsten wrote in Thu Jan 05, 2017 6:42 am:
(b) With the entire FG scenery universe installed and configured as noted above, launching FG results in horrendous amounts of time taken "loading scenery". I usually loose patience after about 20 or 30 minutes, clobber FG with the task manager, reboot, and run M$ Flight Sim X instead. I cannot prove, but suspect, that FG is attempting to load the entire bloody scenery universe, instead of just the chunk it needs.

That's frankly unlikely. Do you have *any* internet connection on then or not?

I've seen these too when developing the SIMD code. What seems to happen is that somewhere a parameter becomes NaN (not a number) or inf(inite) and then FlightGear assumes it should load the entire planet. If only I would know where to look in the code, then it would make sense to check for NaN and inf before starting scenery loading.

Erik
erik
 
Posts: 1504
Joined: Thu Nov 01, 2007 1:41 pm

Re: Is there a "standard reference system" for testing insta

Postby Richard » Thu Jan 05, 2017 11:27 am

I generally use virtual machines (VirtualBox or HyperV (windows)) to do this (in the day job); and it's really good for testing the installation process - especially the hard to test clean install case. For FG I suspect the runtime may be slow (or broken) as it's not that simple to virtualise the GPU and would rely on OpenGL software rendering.

I did once have a set of virtual machines for all of the major platforms (opensuse, ubuntu, OSX, windows XP -> 8.1) under VirtualBox. However (and strangely) it's easiest to do cross platform testing on a Mac because OSX doesn't play at all well (and questionably in licence terms) with most of the virtual machines.

Luckily Microsoft are currently making available virtual machines (that last 90 days) (for testing Microsoft edge) - so grabbing one of these might be useful. https://developer.microsoft.com/en-us/m ... tools/vms/
Richard
 
Posts: 618
Joined: Sun Nov 02, 2014 10:17 pm
Version: Git
OS: Win10

Re: Is there a "standard reference system" for testing insta

Postby jharris1993 » Sun Jan 08, 2017 9:07 pm

Thorsten wrote in Thu Jan 05, 2017 6:42 am:
<snip>

You can get to the bottom of this one by starting FG with the text console with log level to debug, watch the console output and just read up what it's doing when 'loading scenery' is on.

In general: If you want to help (and that's much appreciated) - maybe familiarize yourself with the set of tools - how to change log level, where to find the logs, how to monitor what's going on on the console, how to access the commandline that's being used, how to add things to the commandline.


OK. . . I'm in Windows. . . . . Other than the startup dialog where it offers me the Golden Opportunity to customize the startup parms, how do I "monitor what is going on, on the console"? Or does that translate to "open a command prompt and run FG manually instead of clicking the icon."

Or. . . . Do you mean run it under Linux using a terminal window?

And yes, I know how to do I/O redirection in both Windows and Linux, so (if needed) I can capture either/both standard output and standard error to logs.

One thing I CAN do - and I am dedicating HD space for it - is to run a dual-boot system with both Windows and Mint.

Re: Location of logs.
I am assuming that there are postings and/or Wiki articles that go into painful detail about all of this.

Re: My internet - in response to someone else's question.
Yes, I do have Internet - makes sense, right? - I don't think posting via carrier-pigeon is all that efficient, 'eh? :roll: :lol:

"Da bitch part" (as they said in Blazing Saddles is that this is an ADSL connection and it varies - randomly - from about 4mbit to 28.8 modem speed. (I use VyprVPN to either NYC or 'DC, which just about triples my speed. Even with that, it still stinks!) On the few occasions when I DO get decent line speeds, my wife wants to watch movies! Errrr. Of course this completely ignores the all to annoying times when my Internet connection just goes away into bye-bye land for whatever reason.

Needless to say, the less I have to depend on an Internet connection, the better I like it!

I've been pestering her for a fiber connection, so far no joy as over here you have to buy all the consumer-end-point equipment up-front - between $200 and $500 USD depending on what services you want, and how fast you want them.

Because of these factors, I worked to mitigate the problems this creates by:
1. Buying the whole enchalada on a thumb-drive before I left
2. Downloading the entire Win7 x86 and x64 Windows Update repository (WSUS Offline Updater is excellent), and brought it with me. (I have a multi-multi-meg cable internet connection back in the U.S. that I took full advantage of before I left.)

This way I can build-out Win7 systems and/or install FG along with a reasonable copy of the TerraSync universe, without depending on any kind of Internet connection.

Jim (JR)
What say ye?

Jim (JR)

Some see things as they are, and ask "Why?"
I dream things that never were, and ask "Why Not".

Robert F. Kennedy

“Impossible” is only found in the dictionary of a fool.
Old Chinese Proverb
jharris1993
 
Posts: 139
Joined: Sun Nov 18, 2012 9:20 pm
Location: Worcester, MA. / Moscow, Russia
Callsign: $%^&ing Idiot!
Version: Whatever..
OS: Win-7 / Linux

Re: Is there a "standard reference system" for testing insta

Postby wkitty42 » Mon Jan 09, 2017 12:49 am

FWIW: i have a (bash) script that you are welcome to... i use it when i start my sim from the command line... i don't do GUI startups for flightgear... it records everything from the normal flightgear log data that ends up in fgfs.log to the console output... what you would have to do to make it work on winwhatever i have no clue but you are welcome to it for your *nix endeavors... it should be easy enough to convert to winwhatever or even powershell if you are into that... just PM me and ask ;)
"You get more air close to the ground," said Angalo. "I read that in a book. You get lots of air low down, and not much when you go up."
"Why not?" said Gurder.
"Dunno. It's frightened of heights, I guess."
User avatar
wkitty42
 
Posts: 4755
Joined: Fri Feb 20, 2015 3:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 14.04.5

Re: Is there a "standard reference system" for testing insta

Postby jharris1993 » Mon Jan 09, 2017 4:49 pm

wkitty42 wrote in Mon Jan 09, 2017 12:49 am:FWIW: i have a (bash) script that you are welcome to... i use it when i start my sim from the command line... i don't do GUI startups for flightgear... it records everything from the normal flightgear log data that ends up in fgfs.log to the console output... what you would have to do to make it work on winwhatever i have no clue but you are welcome to it for your *nix endeavors... it should be easy enough to convert to winwhatever or even powershell if you are into that... just PM me and ask ;)


I'm asking!

I am doing this in the forum thread, instead of a PM as you asked, because the idea of being able to freely translate from 'nix to Windows is not as simple as it might appear from the 'nix standpoint. I began explaining this in my PM to you, and decided that this was worth sharing on the public forum. Ergo, here we are. You can PM me with whatever scripts and/or instructions you may wish to share. I can try them within Linux, and then see how well they convert to Windows. I will then report results.

BTW, (based on my limited experience with 'nix compared to Windoze), trying to translate startup configurations - even for programs that exist on both systems in essentially the same form - is a task of non-trivial complexity.

Part of the problem is the basic nature of the two beasties:
Linux:
Most 'nix systems, including Linux, is essentially a command line environment that just happens to have a slick GUI available to slap on top of it, on the off chance you like GUI's. The long and short of it is that anything you do via the GUI is (most of the time), ported to the "real" back-end commands that do the actual work. GParted is a slick tool, and is a classic example of a GUI front-end for the actual heavy lifting done by text commands in the background. Ubuntu/Mint's "Updater" - as well as Synaptic, are slick front-ends for apt-get. And so on. Noobs like me who actually LIKE GUI interfaces annoy those of you who really do know what you're doing, and consider a GUI "training wheels" for those of us who don't.

An example would be a bicycle with an outrigger motor attached. Using the bicycle with the motor is nice, but not necessary. Most of the advanced functionality of the bicycle - gear ratios, BMX'ing, (if you do that kind of thing), is (usually) done sans motor.

Windows:
On the other hand, Windows systems from about Windows ME onward - especially since XP - are entire monolithic systems that absolutely depend on their GUI. Sure, there are some things you can run from a command prompt, there are even more things you can do from an elevated, (administrative), command prompt. However - unlike Linux - you cannot do everything from the command line.

To communicate with Windows, command line tools have to use special "wedge" software, (rundll and svchost are great examples), to interface with the window manager, send messages into the Window's message queue, and extract replies. EVERYTING, (well, almost everything), is done via the Windows Messaging Subsystem. If you want to talk to Windows itself - or even a process/application running under Windows - there are certain very specific API's you must use.

An example would be a motorcycle - or even better, an automobile. Using the command line is like attaching specialized diagnostic equipment to the engine to monitor what is happening - however you cannot operate the vehicle just by using the diagnostic interface.

'nix has a GUI glued to the underlying CLI interface. Windows has a CLI glued to the underlying GUI interface. That is the fundamental difference.

Jim (JR)
What say ye?

Jim (JR)

Some see things as they are, and ask "Why?"
I dream things that never were, and ask "Why Not".

Robert F. Kennedy

“Impossible” is only found in the dictionary of a fool.
Old Chinese Proverb
jharris1993
 
Posts: 139
Joined: Sun Nov 18, 2012 9:20 pm
Location: Worcester, MA. / Moscow, Russia
Callsign: $%^&ing Idiot!
Version: Whatever..
OS: Win-7 / Linux

Re: Is there a "standard reference system" for testing insta

Postby wkitty42 » Tue Jan 10, 2017 2:56 am

i should start with the fact that i use the dnc script (aka download_and_compile.sh) to build my own flightgear... there is no GUI icon clicky thing unless one makes their own... in my case, i just use the scripts that dnc builds to be able to use flightgear and its tools...

the actual script i was talking about that i use mostly is the following... it runs flightgear in debug mode and logs everything to a log file of the same name as the script but with a ".log" extension...

runfgfsdebug:
Code: Select all
#!/bin/bash

# first set the logging and current directory variables
LOGFILE=$0-console.log
CWD=$PWD
RUNCMD="$HOME/flightgear-dev/next/run_fgfs_debug.sh --launcher $@"

function rotate () {
  # minimum file size to rotate in MBi:
  local MB="$1"
  # filename to rotate (full path)
  local F="$2"
  local msize="$((1024*1024*${MB}))"
  test -e "$F" || return 2

  local D="$(dirname "$F")"
  local E=${F##*.}
  local B="$(basename "$F" ."$E")"

  local s=

  printf "rotate msize=$msize file=$F -> $D | $B | $E\n"
  if [ "$(stat --printf %s "$F")" -ge $msize ] ; then
     for i in 8 7 6 5 4 3 2 1 0; do
       s="$D/$B-$i.$E"
       test -e "$s" && mv $s "$D/$B-$((i+1)).$E"
  # empty command is need to avoid exit iteration if test fails:
       :;
     done &&
     mv $F $D/$B-0.$E
  else
     printf "rotate skip: $F < $msize, skip\n"
  fi
  return $?
}

rotate 0 $LOGFILE


# Redirect stdout ( > ) into a named pipe ( >() ) running "tee"
exec > >(tee $LOGFILE)

# Without this, only stdout would be captured - i.e. your
# log file would not contain any error messages.
exec 2>&1



## do your script work commands here
### set your script's necessary variables
VAR1=some_value
VAR2=another_value

### run your commands here

printf "RUNCMD=$RUNCMD\n"
$RUNCMD

## end of script work commands

### unset your script's necessary variables
unset VAR1
unset VAR2

# return to the directory we were in when we started the script
cd $CWD

the key is the way it does the logging... you can see by the RUNCMD that it executes another script and passes the entire command line to that... that other script simply runs flightgear in the gdb debugger so i can grab backtraces on faults... this script i normally execute as "~/runfgfsdebug --log-level=debug" so that i can specify the logging level instead of hard coding it in this script or the other one...

if you look close, you can see that this script is really just a template script for the logging... i add or modify a few variables at the top of the script and then place all my commands within the "run your commands" block... if i add or modify the VAR1 variables, i adjust them at the end to unset them... the RUNCMD at the top is a new item i've been toying with and it seems to work pretty well... in the run block, it is printed so the full command is logged for reference and then it is executed directly... since it is pretty simple, it works well... it won't be so simple if you try to get too tricky with other quotes and escaped characters, though...

so anyway, that's the main thing... powershell should be able to do the same thing on winwhatever... the main part being the logging of STDOUT and STDERR which should grab all console output... it meshes pretty well with the ~/.fgfs/fgfs.log file, too, so you get two logs for the price and this one grabs the console output that the other simply doesn't/can't/won't...
"You get more air close to the ground," said Angalo. "I read that in a book. You get lots of air low down, and not much when you go up."
"Why not?" said Gurder.
"Dunno. It's frightened of heights, I guess."
User avatar
wkitty42
 
Posts: 4755
Joined: Fri Feb 20, 2015 3:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 14.04.5

Re: Is there a "standard reference system" for testing insta

Postby jharris1993 » Tue Jan 10, 2017 4:06 pm

wkitty42 wrote in Tue Jan 10, 2017 2:56 am:so anyway, that's the main thing... powershell should be able to do the same thing on winwhatever... the main part being the logging of STDOUT and STDERR which should grab all console output... it meshes pretty well with the ~/.fgfs/fgfs.log file, too, so you get two logs for the price and this one grabs the console output that the other simply doesn't/can't/won't...


Sounds like an enhancement request to me - an additional logging option that causes it to take messages that would normally go to STDERR and send them to the logs.

Jim (JR)
What say ye?

Jim (JR)

Some see things as they are, and ask "Why?"
I dream things that never were, and ask "Why Not".

Robert F. Kennedy

“Impossible” is only found in the dictionary of a fool.
Old Chinese Proverb
jharris1993
 
Posts: 139
Joined: Sun Nov 18, 2012 9:20 pm
Location: Worcester, MA. / Moscow, Russia
Callsign: $%^&ing Idiot!
Version: Whatever..
OS: Win-7 / Linux

Re: Is there a "standard reference system" for testing insta

Postby wkitty42 » Tue Jan 10, 2017 4:58 pm

yeah but then you can't pipe'em through grep to grab or filter out certain lines ;) actually, i log everything since i normally run in --log-level=debug and then filter for things i do want... eg: currently my script (posted above) has been grepping out a list of airports that the AI traffic manager can't find and whines about when it loads a craft headed or leaving one of them... i trimmed all that out so as to provide the basics of the script and how it logs for your and others to use...

another thing about writing things to the console is because that's where stuff is written that a sim-dev or craft-dev can see them during their creative and troubleshooting processes... that way they can see immediately what's happening as they trigger things... they don't always want to have to go digging in a log file... especially if it is several tens of megs or larger in size... then again, with my background in BBSes, spelunking logs is a pastime 30+ years long... plus i have the log for posting/sharing if one of the dev critters needs it ;)
"You get more air close to the ground," said Angalo. "I read that in a book. You get lots of air low down, and not much when you go up."
"Why not?" said Gurder.
"Dunno. It's frightened of heights, I guess."
User avatar
wkitty42
 
Posts: 4755
Joined: Fri Feb 20, 2015 3:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 14.04.5

Re: Is there a "standard reference system" for testing insta

Postby jharris1993 » Wed Jan 11, 2017 6:24 pm

Greetings!

I hope all of you had a wonderful Christmas and New-Years. (or whatever mid-winter holiday you celebrate!)

The difference between 'nix and Windows reminds me of the Linux utility "WINE" (Windows Is Not linEx <== :roll: Couldn't resist the bad pun!)

A lot of the techniques that are commonly used in a 'nix based system are totally inappropriate or difficult to use in a Windows environment without severely impacting the test environment. This is one of the main reasons that Windows Dev's often send bugs back - "Could Not Reproduce" - because the development environment includes things and makes assumptions that "clean" systems cannot do.

A lot of the techniques I am reading about involve compiling the source tree with various options enabled. I am not sure if I can even DO that in Windows without advanced - and expensive - development tools.

Are there things available that can be done using the "as built" version of FG, without having to resort to recompiling or jumping through other development hoops?

Jim (JR)
What say ye?

Jim (JR)

Some see things as they are, and ask "Why?"
I dream things that never were, and ask "Why Not".

Robert F. Kennedy

“Impossible” is only found in the dictionary of a fool.
Old Chinese Proverb
jharris1993
 
Posts: 139
Joined: Sun Nov 18, 2012 9:20 pm
Location: Worcester, MA. / Moscow, Russia
Callsign: $%^&ing Idiot!
Version: Whatever..
OS: Win-7 / Linux

Re: Is there a "standard reference system" for testing insta

Postby wkitty42 » Wed Jan 11, 2017 10:31 pm

good try at the pun! :)

WINE is actually a self-referencing acronym and stands for "WINE Is Not an Emulator"... WineHQ
"You get more air close to the ground," said Angalo. "I read that in a book. You get lots of air low down, and not much when you go up."
"Why not?" said Gurder.
"Dunno. It's frightened of heights, I guess."
User avatar
wkitty42
 
Posts: 4755
Joined: Fri Feb 20, 2015 3:46 pm
Location: central NC, USA
Callsign: wk42
Version: git next
OS: Kubuntu 14.04.5

Re: Is there a "standard reference system" for testing insta

Postby Richard » Thu Jan 12, 2017 4:58 am

jharris1993 wrote in Wed Jan 11, 2017 6:24 pm:A lot of the techniques I am reading about involve compiling the source tree with various options enabled. I am not sure if I can even DO that in Windows without advanced - and expensive - development tools.

Are there things available that can be done using the "as built" version of FG, without having to resort to recompiling or jumping through other development hoops?


Microsoft provides a free version of Visual Studio that builds FG just fine. Installing Visual Studio Community 2015 will also get you the tools for debugging; and if you download the built binaries from Jenkins you should be able to use the debugging tools with the pre built version.

It's a little tricky to setup a build environment under windows; but it is helpful really only when you're building.

The rest of the debugging / symbol tools can be equally well used with a prebuilt executable provided that you have the PDB files.
Richard
 
Posts: 618
Joined: Sun Nov 02, 2014 10:17 pm
Version: Git
OS: Win10

PreviousNext

Return to Development

Who is online

Users browsing this forum: No registered users and 2 guests