Summary: | net-dns/maradns-3.5.0022 fails to compile (lto): DwRecurse.c:54:16: error: type of [-Werror=lto-type-mismatch] | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Agostino Sarubbo <ago> |
Component: | Current packages | Assignee: | No maintainer - Look at https://wiki.gentoo.org/wiki/Project:Proxy_Maintainers if you want to take care of it <maintainer-needed> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | dev-portage, maradns, sam |
Priority: | Normal | ||
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | |||
Bug Blocks: | 618550 | ||
Attachments: | build.log |
Description
Agostino Sarubbo
2022-07-26 18:31:20 UTC
Created attachment 794654 [details]
build.log
build log and emerge --info
lto_tinderbox has reproduced this issue with version 3.5.0022 - Updating summary. Dear portage team, may I ask why this ebuild fails in src_compile but does not die there and instead dies in src_install? I can reproduce with the latest, but at that point there was the ebuild: https://github.com/gentoo-mirror/gentoo/blob/992c7c565d30ff9dc92dc09eeca2d054cd24ad46/net-dns/maradns/maradns-3.5.0020.ebuild It's to do with the unconventional Makefile. Probably something like running all those commands together only uses the last one's exit code for the rule? Not sure, I haven't played with it. I suspect if you run 'make' manually with those *FLAGS set, it'll exit with exit code 0 (i.e. no failure). The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=02a35ae88a3b84bf2381ffb4244102bdd86f1418 commit 02a35ae88a3b84bf2381ffb4244102bdd86f1418 Author: Eli Schwartz <eschwartz93@gmail.com> AuthorDate: 2024-03-03 05:03:44 +0000 Commit: Sam James <sam@gentoo.org> CommitDate: 2024-03-03 05:51:12 +0000 net-dns/maradns: mark as LTO-unsafe Closes: https://bugs.gentoo.org/861293 Signed-off-by: Eli Schwartz <eschwartz93@gmail.com> Signed-off-by: Sam James <sam@gentoo.org> net-dns/maradns/maradns-3.5.0022.ebuild | 7 ++++++- net-dns/maradns/maradns-3.5.0036.ebuild | 7 ++++++- 2 files changed, 12 insertions(+), 2 deletions(-) Upstream here. The issue has been fixed upstream in https://github.com/samboy/MaraDNS/commit/d828cdf60f50272fa826ca2e5244e6d5a9e7f737 I have been wanting to have compile time flags to make sure my int_fast32_t variables were consistent everywhere; older compilers would not find those glitches (since the two types are the same right now) so I considered it a non-issue until GCC had some obscure compile time flags to catch those issues. (In reply to Sam Trenholme from comment #6) > Upstream here. The issue has been fixed upstream in > https://github.com/samboy/MaraDNS/commit/ > d828cdf60f50272fa826ca2e5244e6d5a9e7f737 > > I have been wanting to have compile time flags to make sure my int_fast32_t > variables were consistent everywhere; older compilers would not find those > glitches (since the two types are the same right now) so I considered it a > non-issue until GCC had some obscure compile time flags to catch those > issues. Thank you! Yeah, LTO has turned out to be a goldmine for this sort of thing, because of the better diagnostics it can produce. |