Hasn't built with Clang for the last 4+ years, no new releases for 7 years...
Can probably go ahead and presume this will never build with Clang again, which
makes it something we don't want to carry.
60 "known" downloads in analytics.
We're using an unstable release, which has now been unstable for over a decade.
Has moved from Google Code to GitHub but attracted no notable interest/forks,
and no newer commits since the last one prior to moving in 2009. Also has
0 "known" installations on analytics.
Basically, don't think it's worth presuming this is ever going to be stable
again.
It doesn't build under the vendor experiment, which is mandatory from Go 1.7.
If you can fix this, please file that fix upstream & we'll happily accept a PR
restoring the formula, but it's not used enough to merit holding up Go 1.7.
The error:
```
==> godep go build -o /usr/local/Cellar/fugu/1.1.1/bin/fugu main.go usage.go version.go
package main
imports github.com/mattes/fugu
imports github.com/docker/docker/registry
imports github.com/docker/docker/engine
imports github.com/docker/docker/utils
imports github.com/docker/docker/pkg/archive
imports github.com/docker/docker/vendor/src/code.google.com/p/go/src/pkg/archive/tar: must be imported as src/code.google.com/p/go/src/pkg/archive/tar
godep: go exit status 1
```
Upstream hasn't cut a release in over 18 months. In the mean time this has gone
through 10 formula revisions & has a whole vendored LibreSSL because upstream
isn't keeping up with LibreSSL development, sadly, which is rapid & regularly
breaking.
I'm uncomfortable keeping an unsupported LibreSSL buried inside this formula
when it looks increasingly like that's going to become a permanent fixture rather
than the temporary hack it was intended to be when inserted initially.
We wouldn't accept this formula today due to notability issues in our audit and
our usage data shows only 18 "known" installations since analytics were
switched on a few months ago.
Let's remove this from Homebrew until the situation improves in terms of
mandating LibreSSL but not matching its release cadence.
- removals from boneyard
- note that lmutil had previously been migrated from binary to boneyard
Closes#3501.
Signed-off-by: ilovezfs <ilovezfs@icloud.com>
Newer versions (8.0+) of icmake are broken on non-GNU systems in new ways.
Icmake isn't used by anything in Homebrew core; it's just there to support
yodl in the future, for zsh development. Abandoning it instead of getting
the newer versions to build.
Closes#3487.
Signed-off-by: Andrew Janke <andrew@apjanke.net>
I'm only intending for this to be temporary, but we're stuck between a rock and
a hard place on this formula & it's holding up work that's really more significant.
Essentially, for the new release Debian moved to requiring `libmd`, which is a
cryptographic library that is portable on BSD and Linux, and was included in OS X
until 10.8 when it was removed.
Various implementations are available to us:
We could use the "original" `libmd` that today needs heavy patching to build on OS X.
We could try to revive the old Apple `libmd` implementation, which doesn't build as-is.
We could attempt to use the one Debian uses, which although despite being slightly
modernised by Debian still doesn't build on OS X.
We could also ask Debian to go back to using a completely portable mechanism and
drop the `libmd` dependency.
nant currently blocks https://github.com/Homebrew/homebrew-core/pull/2011, and looks like it will for the foreseeable.
Upstream commits which fix this issue include:
4d2e97f2cf85a688731609a4a2baa1
But we're not going to adopt an untested 4k+ diff when there's no obvious upstream
release coming; the last nant release was mid 2012, we're already carrying a patch
from 2013 & another build fix from 2015, and neither are in any release yet.
We're more than happy to take this formula back when there's a new release covering
the above, but till then the amount of hacking needed to bring it back to life from
the last release is onerous & I'd recommend people install it directly from git
if they want to use it in the meantime.
Hard dependency on grass, also boneyarded, which is blocking GCC
6.1.0. See PR #676 for more discussion.
Closes#1756.
Signed-off-by: Martin Afanasjew <martin@afanasjew.de>
Requires the now removed Judy, and literally two people installed this formula
from bottle over the last 365 days.
Closes#1628.
Signed-off-by: Dominyk Tiller <dominyktiller@gmail.com>
The formula does not currently build useful installations of
GASNet. The stub/include makefiles the GASNet installer writes contain
references to `$(homebrew --prefix)/Library/ENV/4.3/ld` etc. which
don't work outside the homebrew installation environment. These need to
be patched to be useful, so that they resolve to the compiler used to
build/install GASNet.
Closes#1368.
Signed-off-by: Martin Afanasjew <martin@afanasjew.de>
Moves mit-scheme from the x11 tap to core with modifications
for disabling the X11 dependency.
Closes#1030.
Signed-off-by: Alex Dunn <dunn.alex@gmail.com>
Upstream has ignored the 64-bit patches from the original Homebrew pull
request, as well as subsequent 64-bit reports:
https://sourceforge.net/p/libodbcxx/bugs/25/
Trunk hasn't been updated since 2010 and the last tag was in 2009.
ClosesHomebrew/homebrew#49980.
Signed-off-by: Dominyk Tiller <dominyktiller@gmail.com>
MLton is whole-program, optimizing compiler for Standard ML.
A previous mlton formula simply installed the upstream binary release
and was moved to the boneyard as a binary-only formula (see
Homebrew/homebrew#21780).
This new mlton formula builds from source, using the upstream binary
release to bootstrap.
ClosesHomebrew/homebrew#48694.
Signed-off-by: Andrew Janke <andrew@apjanke.net>
