| Summary: | net-misc/gogoc-1.2-r3 crashes with glibc-2.22-r1 | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Simone Scanzoni <nonno.cicala> |
| Component: | [OLD] Unspecified | Assignee: | Bernard Cafarelli <voyageur> |
| Status: | RESOLVED WONTFIX | ||
| Severity: | major | ||
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | AMD64 | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: |
emerge --info (after reverting to glibc-2.21-r1)
client_dns_server_address_ipv6.patch gogoc-1.2-setenv.patch |
||
Created attachment 430196 [details, diff]
client_dns_server_address_ipv6.patch
The program is trying to assign a NULL pointer to an environment variable.
I don't know if that didn't happen for some reason with glibc-2.21, or why it didn't crash if it happened, but I wrote a patch to set the variable only if it isn't a NULL.
This solved the problem for me.
Created attachment 436808 [details, diff] gogoc-1.2-setenv.patch I use this patch from https://bugzilla.redhat.com/show_bug.cgi?id=1269322 and it also solve problem (In reply to hamper.nhs from comment #3) > Created attachment 436808 [details, diff] [details, diff] > gogoc-1.2-setenv.patch > > I use this patch from https://bugzilla.redhat.com/show_bug.cgi?id=1269322 > and it also solve problem This sets the variable as an empty string, instead of not assigning it as I did (in both cases if it was a NULL pointer). (In reply to Simone Scanzoni from comment #4) > (In reply to hamper.nhs from comment #3) > > Created attachment 436808 [details, diff] [details, diff] [details, diff] > > gogoc-1.2-setenv.patch > > > > I use this patch from https://bugzilla.redhat.com/show_bug.cgi?id=1269322 > > and it also solve problem > > This sets the variable as an empty string, instead of not assigning it as I > did (in both cases if it was a NULL pointer). BTW, I've been using my patch almost daily since then and it works fine. Thanks for the initial bugreport and patch(es), but it looks like you will have to use another tunnel broker now: service was closed, package last rites in progress in bug #589760 (In reply to Bernard Cafarelli from comment #6) > Thanks for the initial bugreport and patch(es), but it looks like you will > have to use another tunnel broker now: service was closed, package last > rites in progress in bug #589760 The software isn't bound to the service! I've been already using gogoc with a different broker than gogoNET for months! |
Created attachment 421934 [details] emerge --info (after reverting to glibc-2.21-r1) I tried glibc-2.22-r1 and gogoc began to crash giving errors like this: gogoc[3588]: segfault at 8 ip 00007fa99f3ca49a sp 00007ffcaaa84798 error 4 in libc-2.22.so[7fa99f349000+190000] I need it for IPv6, so I had to revert to glibc-2.21-r1 (btw, if you are so crazy to downgrade from 2.22 to 2.21, you'll need a /usr/lib64/locale/locale-archive from 2.21, I took one from the stage3 tarball and copied it with a live disc, because as soon as your system has the binaries from 2.21 with a newer local-archive bash stops working and the system becomes unusable, you won't have the opportunity to regenerate the locales).