On 28/11/2019 12:08, Bo Berglund via lazarus wrote:
On Mon, 25 Nov 2019 10:31:51 +0100, Bo Berglund via lazarus
<lazarus@lists.lazarus-ide.org> wrote:
OK, This worked!
Now I have a Lazarus IDE running after starting via gdb.
Lets see how long it will take before the next crash...
So it has been running fine now for 72 hours while I have coded a lot.
No crashes that plagued me last week. :)
Two possibilities:
1) GDB itself makes the Lazarus application behave better?
2) Not using startlazarus when starting lazarus?
I did use the lazarus program for gdb and not startlazarus...
Yes, it is important to use "lazarus" not "startlazarus" in gdb.
gdb only debugs the program it started. By default it will not do
anything to sub processes. That also means restarting the IDE will stop
it from being debugged.
gdb should not have much effect. So there a slight changes in timing (eg
when libraries are loaded). This could change race-conditions.
It could be something different in the remote connection....
There could be a bug in startlazarus, and maybe it started an old/buggy
lazarus from some other place... (not sure, but....)
Or it could be, that when launching in GDB (and with that in a terminal)
there is a console.
That means, if someone put an unprotected "writeln" somewhere in the IDE
code, it could write to console. Otherwise stdout would be a closed handle.
You could try running lazarus, without gdb, but with output redirection
to a file.
Lazarus has some logging that checks for an open console. So there will
be content in that file, that does not confirm anything about a rogue
writeln. But if there were no crash, it could be a hint.
--
_______________________________________________
lazarus mailing list
lazarus@lists.lazarus-ide.org
https://lists.lazarus-ide.org/listinfo/lazarus
