|
Description
Mihai Moldovan
2018-03-14 12:12:53 UTC
Created attachment 523868 [details, diff]
speech-tools-2.5-all-gcc42.patch
Created attachment 523870 [details, diff]
speech-tools-2.5-all-GentooLinux.patch
Created attachment 523872 [details, diff]
speech-tools-2.5-all-sharedlib.patch
Created attachment 523874 [details, diff]
speech-tools-2.5-all-gcc43-include.patch
Created attachment 523876 [details, diff]
speech-tools-2.5-all-remove-shared-refs.patch
Created attachment 523878 [details, diff]
speech-tools-2.5-all-base_class.patch
Created attachment 523880 [details, diff]
speech-tools-2.5-all-etcpath.patch
Created attachment 523882 [details, diff]
speech-tools-2.5-all-gentoo-config.patch
Created attachment 523884 [details, diff]
speech-tools-2.5-all-ldflags-fix.patch
Created attachment 523886 [details, diff]
speech-tools-2.5-all-mixed-cxxflag-cflag-fix.patch
Created attachment 523888 [details, diff]
speech-tools-2.5-all-ncurses-tinfo.patch
Created attachment 576406 [details, diff]
updated speech-tools-2.1-gcc7.patch
It looks to me that the patch speech-tools-2.1-gcc7.patch contains obvious error (a null-terminated string is created at uninitialized address).
There is a semantic difference between C++ and C: in C character literals are always int, in C++ they are either char or int, context dependent. In the case in question code was written in C-style and was correct for C but is broken in C++.
Please consider the updated variant.
(In reply to Alexander Bezrukov from comment #12) > Please consider the updated variant. Sorry, please disregard my comment. The original patch is broken but it is not needed at all and is not applied. And the original code is correct even in C++. Thanks. It works locally, but needs additional patching (both from the latest Gentoo patchset and some Debian patches for known segfaults and other problems). Thanks for taking that up! The ebuild should probably also see an EAPI bump from 5 to 7. I haven't done that back than because my PM lacked EAPI 7 support. Pulling the patches from Debian also sounds like a good idea (they backported most of them anyway from upstream). Probably just a bit time consuming to figure out what's needed and we already have, including a bit of rebasing. Has the Gentoo patch set been updated? The patches I dropped weren't necessary as far as I could tell, but maybe something happened during the last year. Some patches are Debian-specific like wave-format change: it is more compact, but incompatible with upstream, so I don't want to use it, at least now. Gentoo patchset was updated as well: support for ncurses with separate terminfo (tinfo) was added. ping ping Hi! I finally have a vacation and time for Gentoo, though my resources are limited to an old netbook, inet connection quality is not suitable for ssh to use other hosts . So it will take some time and I'm going to fix this stuff within few days. The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=063df9aae04bc82fef0ec6d51144052251cb4b3b commit 063df9aae04bc82fef0ec6d51144052251cb4b3b Author: Peter Levine <plevine457@gmail.com> AuthorDate: 2020-05-30 07:14:29 +0000 Commit: Miroslav Šulc <fordfrog@gentoo.org> CommitDate: 2020-08-23 15:33:19 +0000 app-accessibility/speech-tools: Bump speech-tools-2.5.0 Closes: https://bugs.gentoo.org/650474 Closes: https://github.com/gentoo/gentoo/pull/16011 Bug: https://bugs.gentoo.org/685810 Package-Manager: Portage-2.3.100, Repoman-2.3.22 Signed-off-by: Peter Levine <plevine457@gmail.com> Signed-off-by: Miroslav Šulc <fordfrog@gentoo.org> app-accessibility/speech-tools/Manifest | 1 + .../files/speech-tools-2.5.0-fno-common.patch | 22 ++++ .../speech-tools/speech-tools-2.5.0.ebuild | 141 +++++++++++++++++++++ 3 files changed, 164 insertions(+) |
