SCID crashed today when trying to open a database from the File>Open menu. I decide to run from command line and got:
flatpak run io.github.benini.scid
scid: /run/build/scid/src/namebase.h:172: const char* NameBase::GetName(nameT, idNumberT) const:
Assertion `IsValidNameType(nt) && id < GetNumNames(nt)' failed.
SCID version 5.0.2.
OS: Ubuntu 24.04.1
DE: GNOME
I'm not a software developer but I was amazed at the combined assertion. Code smell?
After banging my head against the wall for a couple of hours, and discovering a bunch of things that troubled me, I found issue #140. A single execution of the cure did not work. I had to run
printf "0??????0??????1??????2??????3??????" >> recover.sn5
several times to recover the database -- surprising that this worked. I guess the id < GetNumNames(nt) was False because GetNumNames() returned a number that was too small; I peeked inside the file with xxd and there did not appear to be any player names according to the SCID5 format and several event names. When saving the database yesterday, I did not enter anything into the player names and only created event names. Looking at the recovered data, it seems that the event names for the last three 'games' were lost. I don't recall a system crash yesterday; I might have shutdown in a rush though.
Edit 1: Removed emotional rant after reading thread #140 from July 2023.
Edit 2: Having recovered and going back to work on my opening analysis, I'm now wary and keeping an eye on the .sn5 file and I've noticed that it seems to be updated only when I close SCID and not when I click (or use the hotkey) 'Game>Save: Replace game...' menu or complete the dialog on 'Game>Save: Add New Game...'. This doesn't seem like safe behaviour to me.
SCID crashed today when trying to open a database from the File>Open menu. I decide to run from command line and got:
SCID version 5.0.2.
OS: Ubuntu 24.04.1
DE: GNOME
I'm not a software developer but I was amazed at the combined assertion. Code smell?
After banging my head against the wall for a couple of hours, and discovering a bunch of things that troubled me, I found issue #140. A single execution of the cure did not work. I had to run
printf "0??????0??????1??????2??????3??????" >> recover.sn5several times to recover the database -- surprising that this worked. I guess the
id < GetNumNames(nt)was False becauseGetNumNames()returned a number that was too small; I peeked inside the file withxxdand there did not appear to be any player names according to the SCID5 format and several event names. When saving the database yesterday, I did not enter anything into the player names and only created event names. Looking at the recovered data, it seems that the event names for the last three 'games' were lost. I don't recall a system crash yesterday; I might have shutdown in a rush though.Edit 1: Removed emotional rant after reading thread #140 from July 2023.
Edit 2: Having recovered and going back to work on my opening analysis, I'm now wary and keeping an eye on the .sn5 file and I've noticed that it seems to be updated only when I close SCID and not when I click (or use the hotkey) 'Game>Save: Replace game...' menu or complete the dialog on 'Game>Save: Add New Game...'. This doesn't seem like safe behaviour to me.