[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Red Hat g++ packaging question

PS = Pete Stieber
PS>> Is there any way to see if Fedora 10 is planning a GCC 4.3.3
PS>> update?

JJ = Jakub Jelinek
JJ> The 4.3.2 in current F10 is closer to 4.3.3 than to 4.3.2.

Thanks for the reply Jakub. I see you are "owner" of the Fedora GCC 4.4 feature. I appreciate you taking the time.

I have code that suffered from the following GCC bug reports...

1. A problem with floating point calculations, that caused repeatable
crashing one of my apps:
I can work around it by adding this compiler option: -fno-if-conversion or by randomly changing other lines of code. 2. A problem with "code reordering" that messed up glibc which affected threading code found in wxWidgets wxThread class:
I worked around it by using joinable threads instead of detached threads.

Are the fixes found in 4.3.3 in Fedora's 4.3.2 on Fedora 10?

PS>> Does the Red Hat version of GCC recompile a source module
PS>> if it encounters an error on the first try?

JJ> Yes, if it encounters an internal compiler error,
JJ> the gcc/g++ compiler driver retries 3 times to
JJ> determine if a bug is reproduceable or not.  If not,
JJ> the most likely culprit is faulty hw, if yes, it
JJ> prepares preprocessed source together with options
JJ> used and tells you to file a bug with that data.

So I guess I should *not* file a bug report when the ICE is not repeatable?

Thanks again for taking the time to reply,

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]