[headdie]

Working on the Revenge: Final Conflict mod tables I have hit a really annoying issue.  the mod is so error ridden that when in debug to generate a log file I end up resorting to rapidly hitting the N key for several seconds just to clear all the errors.

What I am requesting is an option to be selectable in the launcher which when selected suppresses these messages and dumps the usable information from these messages into a separate summery file in addition to the debug.log.  Something like:

Code: [Select]


Freespace 2 Debug Error Summery

Outer blast radius of weapon PTG is zero - EMP will not work. Add $Outer Radius to weapon table entry.
weapons.tbl(line 3270: Warning: Missing quotation marks in string list.
Unable to find WEAPON_LIST_TYPE string "Prometheus D" in ship: GTF Ulysses's primary dogfight banks.
Unable to find WEAPON_LIST_TYPE string "Hornet D" in ship: GTF Ulysses's secondary dogfight banks.
Unable to find WEAPON_LIST_TYPE string "Harpoon D" in ship: GTF Ulysses's secondary dogfight banks.
Unable to find WEAPON_LIST_TYPE string "Prometheus D" in ship: GTF Hercules's primary dogfight banks.


Obviously I am talking about this tool being most useful in situations like the one I am where a the modder is hitting dozens of errors at a time.

thoughts?

[chief1983]

You can suppress those popups via a command line option already I think.  Can't remember off the top off my head what it is.

[Sushi]

You can supress those popups via a command line option already I think.  Can't remember off the top off my head what it is.

But if you do, are the errors still written somewhere?

I can certainly see the use of what headdie is talking about: if there are multiple problems to fix, you wouldn't have to "fix one, restart freespace to discover the next, repeat until done."

[headdie]

-noparseerrors
"Disable Parsing Errors"
Disables some parsing warnings and makes other parse errors non-fatal. Do not rely on this when creating new tables. The errors are there for a reason.

Is this the one you are talking about and does it still post the error in the debug log?

[chief1983]

I thought they still got written to the log, you'd have to check.

[headdie]

I just tested it and I am still getting the error messages.

my flag settings are
-ambient_factor 620 -no_emissive_light -fov 1.0 -ogl_spec 120 -height -normal -spec_exp 15 -spec_point 2.7 -post_process -bloom_intensity 8 -debug_window -noparseerrors

[jr2]

  Umm.... it's spelled summary. 

[The E]

Personally, I am against any "suppress errors and warnings" flag. I can't see what good they would do, but I can see cases where people enable these things and then get sloppy in modding because debug doesn't throw any errors.

[headdie]

I can understand that.

how about in that case expanding the -noparseerrors flag so that it includes this summary or ensure it records all the errors in the debug log  that way the capability is still there but it's use needs to be looked up.

[Zacam]

Or, you use -no_warn instead. Or have it expand to operate on not just fred, which the code comments seem to suggest.

Also, there is a difference between not parsing warnings, and flat out errors. "noparseerrors" should actually be "noparsewarnings" because Errors are still errors.

And obviously, if the MOD cannot progress until they are resolved, maybe you should actually resolve them. Because sweeping them under the rug by any means (fair or foul) will not actually accomplish anything for you.

[karajorma]

I don't think he wants to sweep them under the rug so much as have them all outputted at once rather than having to run, fix bug, run, fix another bug, etc until they're all gone.

I can certainly see the value of that. Imagine if your C++ compiler only spit out one error at a time?

[Fury]

karajorma got it right. This would be extremely useful feature to have in both fs and fred. Going through one error at a time is really annoying when there's a lot of them. In the end you usually end up wasting a lot of time going through all sorts of errors when all you want to do is to get fso debug record all errors so you can get them fixed.

Not to mention that fred2_open is super annoying because its debug window can get bigger than current vertical resolution. I think I asked that to be fixed two years ago but it's still doing it.

[mjn.mixael]

Oh man.. I highly support this request..

[headdie]

Thanks for explaining it better Fury and karajorma that is what I am after

[MatthTheGeek]

Not to mention that fred2_open is super annoying because its debug window can get bigger than current vertical resolution. I think I asked that to be fixed two years ago but it's still doing it.

^ This.

Or how to make life on a crapbook even ****tier than it already is...

[Nuke]

Personally, I am against any "suppress errors and warnings" flag. I can't see what good they would do, but I can see cases where people enable these things and then get sloppy in modding because debug doesn't throw any errors.

id do it but make the flag cause its own popup (which would contain a warning message telling people why you shouldnt use the flag for anything but debugging), while suppressing other errors. this will assure that the flag doesn't get abused, but it gives you one popup instead of 50. point is you dont have to look at a bunch of popups while trying to fix a specific error (or a batch of errors).

[headdie]

I was thinking that having a message come up to state that errors had been detected and the user should check the log file and fix them

[Nuke]

some errors are reported as they happen. so just having a single popup to indicate that you are in a mode that should only be used for debugging. i dont see an issue of adding to that message that you should check the logs for actual errors. the real point of adding a single popup is so the feature does not get abused by people playing mods that have bugs in them.

[karajorma]

I can't imagine it would be too difficult to alter the warning function to work in the way that you have talked about. I like the idea of giving a single warning, not just to prevent abuse but more importantly because we do need to tell people that errors were actually encountered.