GCC 9.4 Compiler Released With 190+ Bug Fixes

[phoronix]

Phoronix: GCC 9.4 Compiler Released With 190+ Bug Fixes

While GCC 11 is the latest stable compiler series for the GNU Compiler Collection, for those still making use of GCC 9 that initially debuted in 2019 there is a new point release this week...

https://www.phoronix.com/scan.php?page=news_item&px=GCC-9.4-Released

[skeevy420]

With all those bug fixes it's clearly inferior software


The joke is I'm making fun of people who make fun of KDE bug fixes

[xnor]

The joke is I'm making fun of people who make fun of KDE bug fixes

But if you fix all KDE bugs, what is left of KDE?

[skeevy420]

Not GNOME

[pipe13]

With all those bug fixes it's clearly inferior software


The joke is I'm making fun of people who make fun of KDE bug fixes

Really. 'Cuz I ran into a gcc-9 bug just last night on CentOS-7 devtoolset-9. (That machine is currently playing at Windows, so I can't check the exact gcc-9 minor version.) The code in question compiles just fine with gcc-8.5 (CentOS-8 Stream), 10.x (Whatever Fedora 32 used) and gcc-11 (Fedora-34), The code looks like a very simple template, pretty sure it's a gcc-9 bug and hope it's among those fixed.

[RahulSundaram]

Really. 'Cuz I ran into a gcc-9 bug just last night on CentOS-7 devtoolset-9. (That machine is currently playing at Windows, so I can't check the exact gcc-9 minor version.) The code in question compiles just fine with gcc-8.5 (CentOS-8 Stream), 10.x (Whatever Fedora 32 used) and gcc-11 (Fedora-34), The code looks like a very simple template, pretty sure it's a gcc-9 bug and hope it's among those fixed.

Did you file a bug report? Hoping isn't a very effective strategy.

[pipe13]

Did you file a bug report? Hoping isn't a very effective strategy.

Notabug. Won't fix.

I'm an unpaid volunteer maintainer. I cheerfully file bug reports against Fedora and CentOS-8 Stream. But against CentOS-7 I draw the line.

I have no personal stake in CentOS-7. My only interest in CentOS-7 are my product's users. On CentOS-7 I specify scl devtoolset-8, which works Just Fine.

At best scl devtoolset-9 just confuses the issue. At worst fixing any problems there might give someone somewhere an excuse to remain on that obsolete platform.

The open-source product I help maintain is Ball Aerospace's Opticks GIS. It does multi-spectral stuff unavailable on any other open-source GIS. Opticks is Really Neat.

We've recently picked up some Windows developers, who have introduced version creep of various components. CMake 2.8 no longer works for us. Fortunately CentOS-7 has a cmake3 (3.17.5) which does. Someone has also introduced an OpenCV dependency somewhat newer than the archaic version 2.4.5 CentOS-7. For that I must specify an OpenCV build from source. I don't need that.

I'll do what I can for my CentOS-7 users, but I'd rather work with Windows10-Linux compatibility from a more modern platform.

As for those who need gcc-9, CentOS-8 Stream's gcc-toolset-9 has gcc version 9.2.1-2, which also compiles Opticks Just Fine.

The code compilation problem I see on CentOS-7 devtoolset-9 (gcc 9.1.1-2) is a bit more complicated than I'd originally supposed in my earlier comment. Distilling it down for Bugzilla would require time I simply do not have, for a problem that doesn't really exist.


FWIW, tonight I checked that a bugzilla had been filed against gdal on CentOS-8 Stream (it won't install), and filed a new one against scl (it won't itself enable gcc-toolset-9).


But thanks for asking. Always appreciate your -- and any vendor's -- participation in these threads.

[RahulSundaram]

Notabug. Won't fix.

I'm an unpaid volunteer maintainer. I cheerfully file bug reports against Fedora and CentOS-8 Stream. But against CentOS-7 I draw the line.

I have no personal stake in CentOS-7. My only interest in CentOS-7 are my product's users. On CentOS-7 I specify scl devtoolset-8, which works Just Fine.

But thanks for asking. Always appreciate your -- and any vendor's -- participation in these threads.

To be clear, I am not a vendor. If you are hoping for bugs to be fixed, you have a vested interest in seeing them get fixed and bug reports would be a more effective approach than forum posts but whatever floats your boat.