Re: [gentoo-dev] [PATCH] elisp.eclass: Call makeinfo with --no-split option
Ulrich Müller writes:
> Signed-off-by: Ulrich Müller
(Why?)
> ---
> eclass/elisp.eclass | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/eclass/elisp.eclass b/eclass/elisp.eclass
> index 1fa8aab29753..e2f629edf132 100644
> --- a/eclass/elisp.eclass
> +++ b/eclass/elisp.eclass
> @@ -139,7 +139,7 @@ elisp_src_configure() { :; }
> elisp_src_compile() {
> elisp-compile *.el
> if [[ -n ${ELISP_TEXINFO} ]]; then
> - makeinfo ${ELISP_TEXINFO} || die
> + makeinfo --no-split ${ELISP_TEXINFO} || die
> fi
> }
[gentoo-dev] Last rites: net-irc/ngircd
[resending as I forgot the other list.] Hi, First, I apologise for letting the package rot a bit. It's wonderful software that's pretty neat, but I've lost all motivation for caring for it because upstream insist on using ancient automake for its ansi2knr tool, removed in later automake releases. See https://github.com/ngircd/ngircd/issues/261#issuecomment-1369266180. While we could just patch the generated configure, it's an ongoing problem if we have to make future changes, and given I've been championing the Modern C work, I just can't really bring myself to do it here. I'm last-riting net-irc/ngircd accordingly. Feel free to adopt it and unmask it if you're willing to do that handrolling of the configure patch for https://bugs.gentoo.org/946998 or somehow persuade upstream. (It might be possible to maintain a tiny meson port for it. Maybe upstream would even accept such a thing if it lived alongside autotools in the repo.) # Sam James (2024-12-27) # Unfixable modern C issues because of reliance on ancient automake. # Out of date. See bug #946998. Removal on 2025-01-27. net-irc/ngircd thanks, sam
Re: [gentoo-dev] [PATCH] elisp.eclass: Call makeinfo with --no-split option
> On Fri, 27 Dec 2024, Sam James wrote:
> (Why?)
By default, makeinfo will split large output files into smaller
subfiles. Although this doesn't currently happen for any package,
it might happen in future, and it would break some assumptions.
The threshold for splitting is 300 kB, and we have at least three
packages with larger info files. The reason they don't break is that
they use the upstream Makefile rather than the eclass function.
>> --- a/eclass/elisp.eclass
>> +++ b/eclass/elisp.eclass
>> @@ -139,7 +139,7 @@ elisp_src_configure() { :; }
>> elisp_src_compile() {
>> elisp-compile *.el
>> if [[ -n ${ELISP_TEXINFO} ]]; then
>> -makeinfo ${ELISP_TEXINFO} || die
>> +makeinfo --no-split ${ELISP_TEXINFO} || die
>> fi
>> }
signature.asc
Description: PGP signature
[gentoo-dev] Last rites: net-irc/ngircd
Hi, First, I apologise for letting the package rot a bit. It's wonderful software that's pretty neat, but I've lost all motivation for caring for it because upstream insist on using ancient automake for its ansi2knr tool, removed in later automake releases. See https://github.com/ngircd/ngircd/issues/261#issuecomment-1369266180. While we could just patch the generated configure, it's an ongoing problem if we have to make future changes, and given I've been championing the Modern C work, I just can't really bring myself to do it here. I'm last-riting net-irc/ngircd accordingly. Feel free to adopt it and unmask it if you're willing to do that handrolling of the configure patch for https://bugs.gentoo.org/946998 or somehow persuade upstream. (It might be possible to maintain a tiny meson port for it. Maybe upstream would even accept such a thing if it lived alongside autotools in the repo.) # Sam James (2024-12-27) # Unfixable modern C issues because of reliance on ancient automake. # Out of date. See bug #946998. Removal on 2025-01-27. net-irc/ngircd thanks, sam signature.asc Description: PGP signature
