On Tue, Feb 18, 2020 at 8:37 PM Bernd Edlinger <bernd.edlin...@hotmail.de> wrote: > > > It has been almost a year since GCC 8.3 has been released and GCC 8.4 > > release should have been released already, so we should concentrate on > > getting it out soon. Unfortunately we have two P1s, one of them is > > waiting for reporter's input, so we might as well just ignore it unless > > the input is provided, but the other, C++ FE one, looks something that > > should be fixed. If we get rid of the P1s, I'd like to create > > 8.4-rc1 on Wednesday, Feb 26th and release 8.4 the week afterwards. > > If you have any queued backports, please commit them to 8 branch > > (and 9 branch too, we'd like to release 9.3 soon too). > > Hi Jakub, > > it just occurred to me that my patch here is a kind of > security relevant one: > > https://gcc.gnu.org/ml/gcc-patches/2020-02/msg01060.html > > since every time the collect2 process is interrupted via a signal > it can delete random files from the hard drive, since the > signal handler may be using the path name, and passes it to the unlink > function before it is initialized. > > The patch doe not apply cleanly to gcc-8, but just fixing > that issue, without tackling the cleanup at the same time should > be feasible, if you like that for this version?
Sure, these kind of fixes are always welcome, but please post them for review. Richard. > > Thanks > Bernd.
