[Ohrrpgce] SVN: teeemcee/10727 Add debugerror, showbug, fatalbug functions and rearrange ErrorLevelEnum

[James Paige]

I really like the idea of making the error log viewable in the F9 menu. It
could also just show the section of g_debug relevant to the current (or
most recent) run of Test Game



On Sunday, September 30, 2018, Ralph Versteegen <teeemcee at gmail.com> wrote:

> (Part 1: just talking to myself)
>
> Our error reporting is quite a mess, and I certainly made it far messier
> when I added different error levels.
> This commit is just an exercise in hill-climbing: making a small
> improvement without thinking about the global optimum: whether the whole
> thing should actually be replaced.
> I'm pretty sick of telling people to look in g/c_debug.txt when they have
> a problem - it's not just for us; those error messages are often useful for
> people too.
> We usually just call debug to log an error. There are about 750 calls to
> debug which aren't commented out. Some are just debug noise, some are
> errors, some indicate corrupt data, some are engine bugs, some are fatal
> errors.
> I hope that by adding (in particular) showbug and debugerror that these
> will get used more instead of debug.
>
> I guess we don't want to pop up error messages unnecessarily while you're
> playing a game, just like we don't want to show script errors, if there's a
> possibility that the error can just be ignored.
> I also don't like how showerror prompts the user to quit the program (once
> 3 errors have occurred), because you don't want to do that while editing a
> game unless you're really stuck.
>
> (Part 2: some ideas)
>
> So maybe a certain severity of error (I guess that's debugerror; debug is
> used too much) should be logged if it happens in-game, but pops up a
> message if it happens while editing (or in Test Game)?
> But then there's a still a risk of it turning into spamming.
>
> So, instead, maybe what we should do is that when debugerror is called, we
> display a warning icon in the corner of the screen for a few seconds with
> some text like "F9 to see error", and in Custom's F9 menu is an option "See
> error log" (probably excluding debuginfo).
> Such an option in the F9 menu would be great either way.
> The warning icon could be shown in-game too, because a lot of errors
> indicate data that couldn't be loaded and you'd likely want to know that as
> a player. If it just displays for a few seconds it doesn't interrupt the
> game a fraction as badly as calling showerror.
>
> Also, I'd like to use CrashRpt to report whenever a bug is detected. Can
> send a minidump/stacktrace. We can do that on Linux/Mac too, just need
> extra code to upload to the server.
>
>
> On Mon, 1 Oct 2018 at 00:39, <subversion at hamsterrepublic.com> wrote:
>
>> teeemcee
>> 2018-09-30 04:39:04 -0700 (Sun, 30 Sep 2018)
>> 543
>> Add debugerror, showbug, fatalbug functions and rearrange ErrorLevelEnum
>> levels
>>
>> There are now aliases for most debugc error levels:
>>  debuginfo
>>  debug
>>  visible_debug
>>  debugerror
>>  showerror
>>  showbug
>>  fatalerror
>>  fatalbug
>> So I will mostly stop using debugc.
>>
>> Got rid of errShowError, renamed errShow to errShowError and errFatal to
>> errFatalError, added errShowInfo and errShowDebug (the latter is
>> visible_debug). Simplified visible_debug and dist_info.
>>
>> Also, forgot to update errorlevel.h and C/C++ code after renaming
>> ErrorLevelEnum
>> constants.
>> ---
>> U   wip/allmodex.bas
>> U   wip/array.c
>> U   wip/common.rbas
>> U   wip/common_base.bas
>> U   wip/common_base.bi
>> U   wip/const.bi
>> U   wip/distribmenu.bas
>> U   wip/errorlevel.h
>> U   wip/filelayer.cpp
>> U   wip/gfx_sdl2.bas
>> U   wip/misc.h
>> U   wip/surface.cpp
>> _______________________________________________
>> Ohrrpgce mailing list
>> ohrrpgce at lists.motherhamster.org
>> http://lists.motherhamster.org/listinfo.cgi/ohrrpgce-motherhamster.org
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.motherhamster.org/pipermail/ohrrpgce-motherhamster.org/attachments/20180930/17e6a157/attachment.html>