Compiler eats my string literals ^^

[Lunasole]

Last night in 5.42 I encountered very strange and rare bug which I cannot reproduce (and don't know exact conditions) so just posting here.

It was like that:
1. I've opened IDE
2. Loaded project, did something, ran it several times, saved & closed
3. Opened another code file (not project), and here bug appeared

In following code string "pMin = " just MISSED from ReplaceString() return and only result of StrF was used, that happened both when running with debugger and in resulting exe.
I wasted 30min trying to find wtf is wrong with code, testing with all those purifiers and so on. And then I just restarted IDE and bug disappeared Does anyone exposed such case?
It looks to be some general compiler bug, not with ReplaceString() or StrF() or other library functions.

Code: Select all

PAGE_LINK_MAIN$ = ReplaceString(PAGE_LINK_MAIN$, "%S4", "pMin = " + StrF(n_pMin / 1024, 1), #PB_String_CaseSensitive, 1, 1)

[Fred]

If you use OS X or Linux 64-bit it could be the same issue as this: http://www.purebasic.fr/english/viewtop ... 24&t=62842

[helpy]

I once had the same problem. It was just an assignment (stringVar.s = "a string"). It was not executed after compilation (with and without debugger).

After restart of IDE IT worked.
I never could reproduce it.

This happend on Windows 7 64 Bit.

[Lunasole]

If you use OS X or Linux 64-bit it could be the same issue as this

No, that was on Win x86. No any crash or other visible errors, just few hardcoded strings (from a very many other hardcoded) were present in code and missing in executable. The StrF() return or other functions returns seems never missing, as I experimented with them when bug appeared.

I once had the same problem. It was just an assignment (stringVar.s = "a string"). It was not executed after compilation (with and without debugger).

After restart of IDE IT worked.
I never could reproduce it.

This happend on Windows 7 64 Bit.

Yes, that looks like my case

[doctorized]

Some rare times, with older 5.x versions I realized that some piece of code was not working properly. I tried to find out what was going on and when I was closing PB and ran it again, everything where back to normal. I was thinking that, maybe, I was loading something and I was not freeing it properly, but now I think it could be some sort of bug, as in some of these cases, the code was not doing something very "special" to have this behavior.

[Fred]

That could happens as we clean the compiler at every compile to not reload it from scratch and gain a few seconds. I would be interested to find a way to reproduce such issue, as it can be wierd to experience for sure.

[dobro]

@Fred:
as you know, with my editor (EPB)
I raise the compiler from scratch at each compilation
and frankly, we can not speak of "seconds"
I do not notice any particular downturn revive the compiler each time,
rather than leaving it in Ram ..

nowadays computers are anyway veloce, not to have this problem ...

******** in Froggy's ***************

comme tu le sais , avec mon Editeur (EPB)
je relance le compilateur a partir de zero a chaque compilation
et franchement, on ne peut pas parler de "secondes"
je ne note pas de ralentissement particulier a relancer le compilateur a chaque fois,
plutot que de le laisser en Ram ..

de nos jours les ordinateurs sont quand meme assez veloce, pour ne pas avoir ce probleme ...

[Lunasole]

I'll just up this topic, maybe there will be some more ppl encountered it?

For this time I already tried about 40 various scenario trying to reproduce that rare and really sneaky bug for interest, but all failed (5.42) ^^