summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
* scripts/bootstrap-prefix: work around bash + gcc14 bootstrap problemhsk172024-12-181-1/+17
| | | | | | | Fetch patch from upstream to address /usr/include/termcap.h requirement. Closes: https://bugs.gentoo.org/946257 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: add support for Darwin Clang bootstrap pathAlexey Sokolov2024-10-181-25/+127
| | | | | | | | | This is a slightly modified version of https://github.com/gentoo/prefix/pull/34/commits/6319cbe1d15c3831bd4a2e5b52c4367abc68747d Bug: https://bugs.gentoo.org/758167 Signed-off-by: Alexey Sokolov <alexey+gentoo@asokolov.org> Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: use newer cmake too bootstrapFabian Groffen2024-10-181-4/+5
| | | | | Signed-off-by: Alexey Sokolov <alexey+gentoo@asokolov.org> Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: bump snapshot (for macOS Sequoia)Fabian Groffen2024-10-011-3/+3
| | | | | | | | | - add timeout to wget (because defaults are unbearable, and gnu is down for the moment) - update snapshot including gcc-14.2.0 fix for macOS - add sed-4.9 to the list of seds to try Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: switch compiler to a properly bootstrapped oneFabian Groffen2024-09-301-6/+21
| | | | | | | | | | We need include-fixed headers and the machinery around it, besides that it is highly discouraged to skip bootstrapping on "foreign" envs as the one we're in at stage2. Rename USE=bootstrap to USE=system-bootstrap Update coreutils bootstrap to 9.5 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: try harder to find matching SDK to OSFabian Groffen2024-09-281-3/+5
| | | | | | | When xcrun returns the latest SDK, no symlink is present, so don't use that to decide whether or not to find a matching SDK on macOS. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: drop obsolete darwin rpath workaroundFabian Groffen2024-09-281-22/+0
| | | | | | | This has been fixed for some time, and for sure in GCC-14 versions we use. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: bump snapshot dateFabian Groffen2024-09-231-1/+1
| | | | | | | This brings in a profile for Darwin24 (Sequoia) but not a working bootstrap for that platform yet. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: use libffi-3.4.5 for Darwin24Fabian Groffen2024-09-181-0/+1
| | | | Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: remove automatic CPU_FLAGS_* settingFabian Groffen2024-09-051-41/+0
| | | | | | | | It doesn't do much, and turns the Prefix from generic to specific to the running machine, which is an undesirable side-effect. Closes: https://bugs.gentoo.org/939038 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: use DISABLE_USE for circular depsFabian Groffen2024-08-211-18/+20
| | | | | | | | | Use set of USE-flags that are necessary to break circular deps, and reduce more to just weed unnecessary dependencies. Use DISABLE_USE for the circular ones so they can be properly reused. Bug: https://bugs.gentoo.org/936629 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: try and fix cyclic deps for nghttp2Fabian Groffen2024-08-211-89/+59
| | | | | | | | bring back the old code which used a minimal set of USE-flags to mask which is necessary to break cycles such as one involving nghttp2 and curl. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: first complete @system before using USE-flagsFabian Groffen2024-08-201-1/+11
| | | | | | | | | Complete @system before trying to set default USE-flags. This may cause re-compilations but is necessary to avoid circular deps e.g. with curl, so take the extra step, much like how in the past we would emerge -e @world after @system. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: drop bits that assume we don't have a C++ compilerFabian Groffen2024-08-201-17/+0
| | | | | | | The situation where we bootstrap with a non-C++ compiler is long gone. Closes: https://bugs.gentoo.org/803551 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: ensure etc/portage/profile exists before writingFabian Groffen2024-08-201-2/+3
| | | | Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: avoid USE=cet with binutils during bootstrapFabian Groffen2024-08-191-0/+4
| | | | | | | | | | the host compiler used may not support CET, causing binutils to fail, so wait for that later when we got our own compiler Thanks hsk17 for the analysis and proposed fix. Bug: https://bugs.gentoo.org/936629 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: workaround binutils being dep-cleanedFabian Groffen2024-08-191-0/+5
| | | | | | | Thanks hsk17 for noticing and providing a patch. Bug: https://bugs.gentoo.org/936629 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: try and unify USE-flag disablingFabian Groffen2024-08-191-14/+26
| | | | | | | | | | As pointed out by hsk17, we have two places in which we disable different sets of USE-flags, try to unify them. In addition add http3, quic and curl_quic_openssl during bootstrap. Bug: https://bugs.gentoo.org/936629 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: set CPU_FLAGS_* when we canFabian Groffen2024-08-111-0/+41
| | | | Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: bump snapshot for Solaris fixesFabian Groffen2024-07-301-1/+1
| | | | Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: use isainfo to pick up current archFabian Groffen2024-07-251-1/+7
| | | | | | | uname -p on Solaris will return historical values, so use isainfo to get the currently targetted architecture Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: disable USE=pie for Solaris bootstrapFabian Groffen2024-07-251-0/+7
| | | | | | | | Host compiler doesn't like -pie in the workload it has to do during bootstrap, so go through lengths to disable it. (Why can't we just override masks/forces from base?) Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: make macOS Ventura and up skip GCC5Fabian Groffen2024-07-241-0/+1
| | | | | | | The host toolchain cannot compile GCC-5, but it *can* compile GCC-14, which is exactly what we need, so skip GCC-5. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: emulate readlink -f behaviour for older macOSFabian Groffen2024-07-231-2/+9
| | | | | | | Avoid some ugly warning and fallback to (likely) wrong SDK with failures as result due to not expanding the MacOSX.sdk symlink properly. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: be selective on where to use GCC-5 in stage1Fabian Groffen2024-07-221-9/+21
| | | | | | | GCC-5 doesn't compile on ARM64, newer versions of macOS might also prove to be capable of compiling GCC-14. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: use GCC-5 on all Darwin bootstrapsFabian Groffen2024-07-211-50/+122
| | | | | | | | | | | | | | | | | With recent macOS we need GCC-14 in order to make sense of system headers. However, GCC-14 gets confused by (older) Clang installed on the host system. So we need GCC-5 to get us a C11/C++11 compiler that can compile GCC-14 to do the real work. Unfortunately for GCC-5 to work properly on systems that do not have /usr/{include,lib} we need wrappers such as we use in Gentoo Prefix proper, thus part of that needs to be pulled and put in place too in order to get correctly compiled and running binaries. This change succeeds bootstrap on Darwin 17, 19 and 20, 21 and 22 pending. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: improve emake error reportingFabian Groffen2024-07-211-1/+4
| | | | | | | If make fails, run with forced -j1 to get a clearer error for debugging purposes. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: define PORTAGE_INST_{UID,GID}Fabian Groffen2024-07-141-0/+9
| | | | | | | | | | Based on the patch by Etienne Buira <etienne.buira@free.fr> in bug #933100. Mainline Portage doesn't set the Portage UID/GID, so define them in make.conf during bootstrap for RAP targets. Closes: https://bugs.gentoo.org/933100 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* pts/rsync-generation/update-rsync-master: fix PYTHONPATHFabian Groffen2024-07-101-1/+1
| | | | Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: bump snapshot for xz-utils fixFabian Groffen2024-06-231-1/+1
| | | | | | | | We need 3f4b5440ef1e9b77b465dcaf0a0e104238ff6ed8 (from bug #934515) to make the bootstrap succeed on macOS/Darwin. Closes: https://bugs.gentoo.org/934757 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: bump snapshot for zstd compile-fixFabian Groffen2024-06-171-1/+1
| | | | Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/auto-bootstraps/analyse_result: fix syntax warningFabian Groffen2024-06-162-3/+3
| | | | | | this probably changed inbetween a Python release or two Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: bump snapshot to an existing oneFabian Groffen2024-06-141-1/+1
| | | | | Closes: https://bugs.gentoo.org/934308 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/rsync-generation/mksnapshot: don't download a crap 404 messageFabian Groffen2024-06-141-1/+1
| | | | | | | Fail instead of downloading a 404/403/500 whatever message. Bug: https://bugs.gentoo.org/934308 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: bump bootstrap snapshotFabian Groffen2024-05-301-1/+1
| | | | | | | This brings in gcc-13.3.0 which helps on macOS Sanoma, but we still don't get up to success. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: attempt to fix b736fab6 do_emerge_pkgsFabian Groffen2024-05-281-7/+11
| | | | | | | | Use ROOT, which should be the same between stages, such that we don't reference tmp/tmp. Thanks huxnu for pointing out. Bug: https://bugs.gentoo.org/933079 Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: unbreak bootstraps due to silly meson depFabian Groffen2024-05-281-4/+32
| | | | | | | | gentoo-functions requires meson nowadays, which cannot be bootstrapped this early because it requires a Python that the python eclass finds so work around this by providing a dummy gentoo/functions.sh Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* bootstrap-prefix: use full path for `do_emerge_pkgs`Matt Jolly2024-05-281-2/+8
| | | | | | | | | | | | | In the stage3 bootstrap we always prefer to use tools that have been built for stage3; to accomplish this we ensure that it is the first thing evaluated in PATH. Unfortunately, Portage, Python, and Python-exec are often pulled into the depgraph at some point before we're fully boostrapped. To ensure that we don't try and execute ${EPREFIX}/usr/bin/emerge before we're ready, always provide the full path to the bootstrap Python interpreter and emerge script. Signed-off-by: Matt Jolly <kangie@gentoo.org> Closes: https://github.com/gentoo/prefix/pull/37 Signed-off-by: Sam James <sam@gentoo.org>
* scripts/bootstrap-prefix: bump snapshot for Darwin Python fixFabian Groffen2024-04-101-1/+1
| | | | Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: bump bootstrap snapshotFabian Groffen2024-04-081-1/+1
| | | | | | fix upstream distfile change for darwin-miscutils Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: snapshot bump to fix Solaris bootstrapFabian Groffen2024-04-071-1/+1
| | | | | | | bump for binutils-config' -rpath handling, which affects probably more than just Solaris Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/bootstrap-prefix: bump snapshot for Darwin fixesFabian Groffen2024-04-061-1/+1
| | | | | | | | - binutils-config ld duplicate paths elimination - gcc math.h fix for darwin - openssl using --push-state --as-needed --pop-state LDFLAGS unconditional Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/auto-bootstraps/dobootstrap: fix shellcheck fixFabian Groffen2024-04-051-1/+1
| | | | | | | previous fix included some quotes in the output, making no sense and failing portage down the line Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/auto-bootstraps/dobootstrap: unbreak after shellcheckFabian Groffen2024-04-051-7/+7
| | | | | | | | the quoting-eagerness of shellcheck isn't always resulting in the desired output, in this case quoted empty strings broke where they previously would be ignored as whitespace separators Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/auto-bootstraps/process_uploads: allow processing of temp files tooFabian Groffen2024-04-021-1/+3
| | | | Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/rsync-generation/mksnapshot: disable all but bz2 snapshotsFabian Groffen2024-03-311-2/+8
| | | | | | | | | | | who am I kidding, the bootstrap and the mirror retaining is all hardwired to bz2, and other compression formats don't really make a big enough dent to warrant lots of changes to allow picking a different compresion format disable all but bz2, that's the only thing we use afterall Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/rsync-generation/mksnapshot: xz can read lzip archives, so drop xzFabian Groffen2024-03-311-1/+0
| | | | | | | lzip (lzma) compresses better than xz (lzma2) but xz can read/decompress lzip archives, so drop the xz archive in favour of lzip Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/rsync-generation/mksnapshot: drop gzip snapshotFabian Groffen2024-03-311-1/+0
| | | | | | | | | | gzip compresses to around 75MB which is much higher than bzip2's 57MB The gzip snapshot was historically never used ever by the bootstrap script, it always used and hardcoded bz2. So don't bother creating a gzip variant. Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/rsync-generation/mksnapshot: increase zstd compression levelFabian Groffen2024-03-311-1/+1
| | | | | | | | | Using -9 we end up at 58MB, which is more than gzip2's 57MB. So up the level to -19 (the max, basically like how we configure the other compressors too) to get 52MB, which is still much more than xz (48MB) and lzip (47MB). Signed-off-by: Fabian Groffen <grobian@gentoo.org>
* scripts/rsync-generation/mksnapshot: fix zstd file extensionFabian Groffen2024-03-311-1/+1
| | | | Signed-off-by: Fabian Groffen <grobian@gentoo.org>