Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Cannot manually compile with 'bmake install' due to missing signatures #399

Closed
jacobvosmaer opened this issue Aug 21, 2016 · 11 comments
Closed

Comments

@jacobvosmaer
Copy link

Hi, I am trying to manually compile lang/clang on OS X (because no binary package is provided in Darwin/trunk). This fails with the following error:

=> Installing llvm-3.8.0{,nb*} from /Users/jacobvosmaer/pkgsrc-packages/All;http://pkgsrc.joyent.com/packages/Darwin/trunk/x86_64//All
pkg_add: No valid signature found, rejected
pkg_add: 1 package addition failed
=> No binary package found for llvm-3.8.0{,nb*}; installing from source.
=> Bootstrap dependency digest>=20010302: found digest-20160304
===> Checking for vulnerabilities in llvm-3.8.0nb2
===> Installing binary package of llvm-3.8.0nb2
pkg_add: No valid signature found, rejected
pkg_add: 1 package addition failed
*** Error code 1

I tried following the instructions in https://pkgsrc.joyent.com/docs/building/ when configuring the pkgsrc tree.

I think it makes sense what happens here; bmake install creates a temporary package that it feeds to pkg_add and pkg_add uses VERIFIED_INSTALLATION=always via /opt/pkg/etc/pkg_install.conf. What I don't understand is how this is supposed to not go wrong. If I were able to locally produce packages that are trusted by /opt/pkg/etc/gnupg/pkgsrc.gpg that would surprise me. :)

Is there maybe some step missing in the 'Building packages' documentation?

@jperkin
Copy link
Collaborator

jperkin commented Aug 21, 2016

Yeh I need to document this better. The problem is that the pkg_install.conf that I ship by default requires signed packages (VERIFIED_INSTALLATION=always), but creating signed packages requires some additional setup and to configure pkg_add to permit the key that you setup (as obviously you can't sign as me!) I'll try to get to this at some point soon, but in the meantime you can remove the VERIFIED_INSTALLATION line and allow your unsigned packages.

@jperkin
Copy link
Collaborator

jperkin commented Aug 21, 2016

I'll also fix the issue causing clang to not be in the repository for the next build. It's due to the build not fitting in the default ramdisk (http://us-east.manta.joyent.com/pkgsrc/public/reports/Darwin/trunk/x86_64/20160812.1847/llvm-3.8.1/build.log) so I'll move it to local disk.

@jacobvosmaer
Copy link
Author

@jperkin awesome!

I will poke around a bit to see if I can figure out the 'sign my own packages' thing.

@jacobvosmaer
Copy link
Author

Some things I find hard re configuring signing packages created by bmake install, in case this helps.

  • gpg --import as described in https://gist.github.com/drscream/c45419950d8af648e2c6 does not work because/when I try to use an already existing key pair: it says the pubkey is already in the (main) keyring
  • getting gpg-agent to work across sudo is finnicky
  • having to do anything with gpg in the first place is... challenging for me. I don't use it every day and every time I do it is no fun.

So I see why this would be hard to document. :( Sorry I can't be of more help.

@jacobvosmaer
Copy link
Author

This works but it is not ideal:

VERIFIED_INSTALLATION=trusted

In pkg_install.conf. That throws up a prompt that lets me ignore the missing signature during bmake install.

@drscream
Copy link

@jacobvosmaer could you share your exact error by using the gpg --import?

Because I run everything as root on the build zone I need to import the key which I have created on one of my own machines. If you use your key which should be already imported in the keyring you don't need todo that again.

@jperkin
Copy link
Collaborator

jperkin commented Aug 22, 2016

Yeh, it can be tricky to get right and there are some GPG corner cases where e.g. it's impossible to correctly choose a signing key when you have multiple imported (especially annoying when you want to use the same machines for different builds).

@jacobvosmaer
Copy link
Author

@drscream hi I just noticed that is your gist I link to :)

I use a primitive OS without zones (OS X) where my build user is my normal user. That gives me this:

$ /opt/pkg/bin/gpg2 --primary-keyring /opt/pkg/etc/gnupg/pkgsrc.gpg --import cjv.pub 
gpg: key A5386FBD: "Jacob Vosmaer <contact@jacobvosmaer.nl>" not changed
gpg: Total number processed: 1
gpg:              unchanged: 1

So the issue here is that I'm reusing that key. Giving root it's own GPG keyring is not a crazy idea, I might try that. But to be honest I am OK with VERIFIED_INSTALLATION=trusted for my use case (personal laptop) so I am not sure if I want to spend much more time fighting GPG. :)

I am not quite sure how to resolve this GitHub issue here I opened. For me, the problem is solved, but the documentation on https://pkgsrc.joyent.com/docs/building/ is still incomplete.

Feel free to close this @jperkin

jperkin pushed a commit that referenced this issue Aug 24, 2016
Changes from 3.2.2 to 3.2.3
Improvements

    It is now possible to use HDF5 with the new shared library naming scheme (>= 1.8.10, hdf5.dll instead of hdf5dll.dll) on Windows (gh-540). Thanks to Tadeu Manoel.
    Now :program: ptdump sorts output by node name and does not print a backtrace if file cannot be opened. Thanks to Zbigniew Jędrzejewski-Szmek.

Bugs fixed

    Only run tables.tests.test_basics.UnicodeFilename if the filesystem encoding is utf-8. Closes gh-485.
    Add lib64 to posix search path. (closes gh-507) Thanks to Mehdi Sadeghi.
    Ensure cache entries are removed if fewer than 10 (closes gh-529). Thanks to Graham Jones.
    Fix segmentation fault in a number of test cases that use index.Index (closes gh-532 and gh-533). Thanks to Diane Trout.
    Fixed the evaluation of transcendental functions when numexpr is compiled with VML support (closes gh-534, PR #536). Thanks to Tom Kooij.
    Make sure that index classes use buffersizes that are a multiple of chunkshape[0] (closes gh-538, PR #538). Thanks to Tom Kooij.
    Ensure benchmark paths exist before benchmarks are executed (PR #544). Thanks to rohitjamuar.

Other changes

    Minimum Cython version is now v0.21

Changes from 3.2.1.1 to 3.2.2
Bug fixed

    Fix AssertionError in Row.__init_loop. See gh-477.
    Fix issues with Cython 0.23. See gh-481.
    Only run tables.tests.test_basics.UnicodeFilename if the filesystem encoding is utf-8. Closes gh-485.
    Fix missing missing PyErr_Clear. See gh-#486.
    Fix the C type of some numpy attributes. See gh-494.
    Cast selection indices to integer. See gh-496.
    Fix indexesextension._keysort_string. Closes gh-497 and gh-498.

Changes from 3.2.1 to 3.2.1.1

    Fix permission on distributed source distribution

Other changes

    Minimum Cython version is now v0.21

Changes from 3.2.0 to 3.2.1
Bug fixed

    Fix indexesextension._keysort. Fixes gh-455. Thanks to Andrew Lin.

Changes from 3.1.1 to 3.2.0
Improvements

    The nrowsinbuf is better computed now for EArray/CArray having a small chunkshape in the main dimension. Fixes #285.

    PyTables should be installable very friendly via pip, including NumPy being installed automatically in the unlikely case it is not yet installed in the system. Thanks to Andrea Bedini.

    setup.py has been largely simplified and now it requires setuptools. Although we think this is a good step, please keep us informed this is breaking some installation in a very bad manner.

    setup.py now is able to used pkg-config, if available, to locate required libraries (hdf5, bzip2, etc.). The use of pkg-config can be controlled via setup.py command line flags or via environment variables. Please refer to the installation guide (in the User Manual) for details. Closes gh-442.

    It is now possible to create a new node whose parent is a softlink to another group (see gh-422). Thanks to Alistair Muldal.

    link.SoftLink objects no longer need to be explicitly dereferenced. Methods and attributes of the linked object are now automatically accessed when the user acts on a soft-link (see gh-399). Thanks to Alistair Muldal.

    Now ptrepack recognizes hardlinks and replicates them in the output (repacked) file. This saves disk space and makes repacked files more conformal to the original one. Closes gh-380.

    New pttree script for printing HDF5 file contents as a pretty ASCII tree (closes gh-400). Thanks to Alistair Muldal.

    The internal Blosc library has been downgraded to version 1.4.4. This is in order to still allow using multiple threads inside Blosc, even on multithreaded applications (see gh-411, gh-412, gh-437 and gh-448).

    The print_versions() function now also reports the version of compression libraries used by Blosc.

    Now the setup.py tries to use the ‘-march=native’ C flag by default. In falls back on ‘-msse2’ if ‘-march=native’ is not supported by the compiler. Closes gh-379.

    Fixed a spurious unicode comparison warning (closes gh-372 and gh-373).

    Improved handling of empty string attributes. In previous versions of PyTables empty string were stored as scalar HDF5 attributes having size 1 and value ‘0’ (an empty null terminated string). Now empty string are stored as HDF5 attributes having zero size

    Added a new cookbook recipe and a couple of examples for simple threading with PyTables.

    The redundant utilsextension.get_indices() function has been eliminated (replaced by slice.indices()). Closes gh-195.

    Allow negative indices in point selection (closes gh-360)

    Index wasn’t being used if it claimed there were no results. Closes gh-351 (see also gh-353)

    Atoms and Col types are no longer generated dynamically so now it is easier for IDEs and static analysis tool to handle them (closes gh-345)

    The keysort functions in idx-opt.c have been cythonised using fused types. The perfomance is mostly unchanged, but the code is much more simpler now. Thanks to Andrea Bedini.

    Small unit tests re-factoring:

        print_versions() and tests.common.print_heavy() functions

            moved to the tests.common module

        always use print_versions() when test modules are called as scripts

        use the unittest2 package in Python 2.6.x

        removed internal machinery used to replicate unittest2 features

        always use tests.common.PyTablesTestCase as base class for all test cases

        code of the old tasts.common.cleanup() function has been moved to tests.common.PyTablesTestCase.tearDown() method

        new implementation of tests.common.PyTablesTestCase.assertWarns() compatible with the one provided by the standard unittest module in Python >= 3.2

        use tests.common.PyTablesTestCase.assertWarns() as context manager when appropriate

        use the unittest.skipIf() decorator when appropriate

        new :class:tests.comon.TestFileMixin: class

Bugs fixed

    Fixed compatibility problems with numpy 1.9 and 1.10-dev (closes gh-362 and gh-366)
    Fixed compatibility with Cython >= 0.20 (closes gh-386 and gh-387)
    Fixed support for unicode node names in LRU cache (only Python 2 was affected). Closes gh-367 and gh-369.
    Fixed support for unicode node titles (only Python 2 was affected). Closes gh-370 and gh-374.
    Fixed a bug that caused the silent truncation of unicode attributes containing the ‘0’ character. Closes gh-371.
    Fixed descr_from_dtype() to work as expected with complex types. Closes gh-381.
    Fixed the tests.test_basics.ThreadingTestCase test case. Closes gh-359.
    Fix incomplete results when performing the same query twice and exhausting the second iterator before the first. The first one writes incomplete results to seqcache (gh-353)
    Fix false results potentially going to seqcache if tableextension.Row.update() is used during iteration (see gh-353)
    Fix Column.create_csindex() when there’s NaNs
    Fixed handling of unicode file names on windows (closes gh-389)
    No longer not modify sys.argv at import time (closes gh-405)
    Fixed a performance issue on NFS (closes gh-402)
    Fixed a nasty problem affecting results of indexed queries. Closes gh-319 and probably gh-419 too.
    Fixed another problem affecting results of indexed queries too. Closes gh-441.
    Replaced “len(xrange(start, stop, step))” -> “len(xrange(0, stop - start, step))” to fix issues with large row counts with Python 2.x. Fixes #447.

Other changes

    Cython is not a hard dependency anymore (although developers will need it so as to generated the C extension code).

    The number of threads used by default for numexpr and Blosc operation that was set to the number of available cores have been reduced to 2. This is a much more reasonable setting for not creating too much overhead.
jperkin pushed a commit that referenced this issue Jan 16, 2017
Version 0.34
------------

Released on 2016-12-21.

Bug fixes:

* `#398 <https://github.com/Kozea/WeasyPrint/issues/398>`_:
  Honor the presentational_hints option for PDFs.
* `#399 <https://github.com/Kozea/WeasyPrint/pull/399>`_:
  Avoid CairoSVG-2.0.0rc* on Python 2.
* `#396 <https://github.com/Kozea/WeasyPrint/issues/396>`_:
  Correctly close files open by mkstemp.
* `#403 <https://github.com/Kozea/WeasyPrint/issues/403>`_:
  Cast the number of columns into int.
* Fix multi-page multi-columns and add related tests.


Version 0.33
------------

Released on 2016-11-28.

New features:

* `#393 <https://github.com/Kozea/WeasyPrint/issues/393>`_:
  Add tests on MacOS.
* `#370 <https://github.com/Kozea/WeasyPrint/issues/370>`_:
  Enable @font-face on MacOS.

Bug fixes:

* `#389 <https://github.com/Kozea/WeasyPrint/issues/389>`_:
  Always update resume_at when splitting lines.
* `#394 <https://github.com/Kozea/WeasyPrint/issues/394>`_:
  Don't build universal wheels.
* `#388 <https://github.com/Kozea/WeasyPrint/issues/388>`_:
  Fix logic when finishing block formatting context.


Version 0.32
------------

Released on 2016-11-17.

New features:

* `#28 <https://github.com/Kozea/WeasyPrint/issues/28>`_:
  Support @font-face on Linux.
* Support CSS fonts level 3 almost entirely, including OpenType features.
* `#253 <https://github.com/Kozea/WeasyPrint/issues/253>`_:
  Support presentational hints (optional).
* Support break-after, break-before and break-inside for pages and columns.
* `#384 <https://github.com/Kozea/WeasyPrint/issues/384>`_:
  Major performance boost.

Bux fixes:

* `#368 <https://github.com/Kozea/WeasyPrint/issues/368>`_:
  Respect white-space for shrink-to-fit.
* `#382 <https://github.com/Kozea/WeasyPrint/issues/382>`_:
  Fix the preferred width for column groups.
* Handle relative boxes in column-layout boxes.

Documentation:

* Add more and more documentation about Windows installation.
* `#355 <https://github.com/Kozea/WeasyPrint/issues/355>`_:
  Add fonts requirements for tests.


Version 0.31
------------

Released on 2016-08-28.

New features:

* `#124 <https://github.com/Kozea/WeasyPrint/issues/124>`_:
  Add MIME sniffing for images.
* `#60 <https://github.com/Kozea/WeasyPrint/issues/60>`_:
  CSS Multi-column Layout.
* `#197 <https://github.com/Kozea/WeasyPrint/pull/197>`_:
  Add hyphens at line breaks activated by a soft hyphen.

Bux fixes:

* `#132 <https://github.com/Kozea/WeasyPrint/pull/132>`_:
  Fix Python 3 compatibility on Windows.

Documentation:

* `#329 <https://github.com/Kozea/WeasyPrint/issues/329>`_:
  Add documentation about installation on Windows.


Version 0.30
------------

Released on 2016-07-18.

WeasyPrint now depends on html5lib-0.999999999.

Bux fixes:

* Fix Acid2
* `#325 <https://github.com/Kozea/WeasyPrint/issues/325>`_:
  Cutting lines is broken in page margin boxes.
* `#334 <https://github.com/Kozea/WeasyPrint/issues/334>`_:
  Newest html5lib 0.999999999 breaks rendering.


Version 0.29
------------

Released on 2016-06-17.

Bug fixes:

* `#263 <https://github.com/Kozea/WeasyPrint/pull/263>`_:
  Don't crash with floats with percents in positions.
* `#323 <https://github.com/Kozea/WeasyPrint/pull/323>`_:
  Fix CairoSVG 2.0 pre-release dependency in Python 2.x.


Version 0.28
------------

Released on 2016-05-16.

Bug fixes:

* `#189 <https://github.com/Kozea/WeasyPrint/issues/189>`_:
  ``white-space: nowrap`` still wraps on hyphens
* `#305 <https://github.com/Kozea/WeasyPrint/issues/305>`_:
  Fix crashes on some tables
* Don't crash when transform matrix isn't invertible
* Don't crash when rendering ratio-only SVG images
* Fix margins and borders on some tables


Version 0.27
------------

Released on 2016-04-08.

New features:

* `#295 <https://github.com/Kozea/WeasyPrint/pull/295>`_:
  Support the 'rem' unit.
* `#299 <https://github.com/Kozea/WeasyPrint/pull/299>`_:
  Enhance the support of SVG images.

Bug fixes:

* `#307 <https://github.com/Kozea/WeasyPrint/issues/307>`_:
  Fix the layout of cells larger than their tables.

Documentation:

* The website is now on GitHub Pages, the documentation is on Read the Docs.
* `#297 <https://github.com/Kozea/WeasyPrint/issues/297>`_:
  Rewrite the CSS chapter of the documentation.
jperkin pushed a commit that referenced this issue Mar 20, 2017
# Change Log

## [3.5.5] - 2017-02-24

[3.5.5]: hashie/hashie@v3.5.4...v3.5.5

### Added

* [#326](hashie/hashie#326): Added `Hashie::Extensions::Mash::KeepOriginalKeys` to give Mashes the ability to keep the original structure given to it - [@michaelherold](https://github.com/michaelherold).

### Fixed

* [#415](hashie/hashie#415): Fixed Mash logging keys multiple times which lead to a bad user experience or, in some cases, errors - [@michaelherold](https://github.com/michaelherold).

## [3.5.4] - 2017-02-22

[3.5.4]: hashie/hashie@v3.5.3...v3.5.4

### Added

* [#412](hashie/hashie#412): Added a Hashie::Extensions::Mash::SymbolizeKeys extension that overrides the default stringification behavior for keys - [@michaelherold](https://github.com/michaelherold).

### Fixed

* [#409](hashie/hashie#409): Fixed Railtie detection for projects where Rails is defined but Railties are not availble - [@CallumD](https://github.com/callumd).
* [#411](hashie/hashie#411): Fixed a performance regression from 3.4.3 that caused a 10x slowdown in OmniAuth - [@michaelherold](https://github.com/michaelherold).

## [3.5.3] - 2017-02-11

[3.5.3]: hashie/hashie@v3.5.2...v3.5.3

### Fixed

* [#402](hashie/hashie#402): Use a Railtie to set Hashie.logger on rails boot - [@matthewrudy](https://github.com/matthewrudy).
* [#406](hashie/hashie#406): Ensure that subclasses that disable warnings propagate that setting to grandchild classes - [@michaelherold](https://github.com/michaelherold).
* Your contribution here.

## [3.5.2] - 2017-02-10

[3.5.2]: hashie/hashie@v3.5.1...v3.5.2

### Added

* [#395](hashie/hashie#395): Add the ability to disable warnings in Mash subclasses - [@michaelherold](https://github.com/michaelherold).
* [#400](hashie/hashie#400): Fix Hashie.logger load and set the Hashie logger to the Rails logger in a Rails environment - [@michaelherold](https://github.com/michaelherold).

### Fixed

* [#396](hashie/hashie#396): Fix for specs in #381: Incorrect use of shared context meant example was not being run - [@biinari](https://github.com/biinari).
* [#399](hashie/hashie#399): Fix passing Pathname object to Hashie::Mesh.load() - [@albb0920](https://github.com/albb0920).

### Miscellanous

* [#397](hashie/hashie#397): Add the integration specs harness into the main test tasks - [@michaelherold](https://github.com/michaelherold).

## [3.5.1] - 2017-01-31

* [#392](hashie/hashie#392): Fix for #391: Require all dependencies of Hashie::Mash - [@dblock](https://github.com/dblock).

[3.5.1]: hashie/hashie@v3.5.0...v3.5.1

## [3.5.0] - 2017-01-31

* [#386](hashie/hashie#386): Fix for #385: Make `deep_merge` always `deep_dup` nested hashes before merging them in so that there are no shared references between the two hashes being merged. - [@mltsy](https://github.com/mltsy).
* [#389](hashie/hashie#389): Support Ruby 2.4.0 - [@camelmasa](https://github.com/camelmasa).

[3.5.0]: hashie/hashie@v3.4.6...v3.5.0

### Added

* [#381](hashie/hashie#381): Add a logging layer that lets us report potential issues to our users. As the first logged issue, report when a `Hashie::Mash` is attempting to overwrite a built-in method, since that is one of our number one questions - [@michaelherold](https://github.com/michaelherold).

### Changed

* [#384](hashie/hashie#384): Updated to CodeClimate 1.x - [@boffbowsh](https://github.com/boffbowsh).

### Fixed

* [#369](hashie/hashie#369): If a translation for a property exists when using IndifferentAccess and IgnoreUndeclared, use the translation to find the property - [@whitethunder](https://github.com/whitethunder).
* [#376](hashie/hashie#376): Leave string index unchanged if it can't be converted to integer for Array#dig - [@sazor](https://github.com/sazor).
* [#377](hashie/hashie#377): Dont use Rubygems to check ruby version - [@sazor](https://github.com/sazor).
* [#378](hashie/hashie#378): Deep find all searches inside all nested hashes - [@sazor](https://github.com/sazor).
* [#380](hashie/hashie#380): Evaluate procs default values of Dash in object initialization - [@sazor](https://github.com/sazor).

### Miscellanous

* [#387](hashie/hashie#387): Fix builds failing due to Rake 11 having a breaking change - [@michaelherold](https://github.com/michaelherold).
jperkin pushed a commit that referenced this issue Mar 20, 2017
kramdown 1.13.2 released				2017/01/07

This release fixes some minor issues - updating is recommended.

Changes

3 bug fixes:

* Fix footnote link spacing to use non-breaking space (pull request #399 by
  Martyn Chamberlin)

* Show warning for unreferenced footnote definitions (fixes #400 reported by
  Kyle Barbour)

* Fix test cases with respect to Ruby 2.4 (fixes #401 reported by Connor Shea)


kramdown 1.13.1 released				2016/11/25

This release fixes the GFM header ID generation for more cases, updating is
very recommended.

Changes

1 bug fix:

* Fix GFM header ID generation when code spans, math elements, entities,
  typographic symbols or smart quotes are used (fixes #391 reported by Nick
  Fagerlund)


kramdown 1.13.0 released				2016/11/20

The biggest change in this release is the introduction of a converter for man
pages. Although there already exist two solutions (ronn and kramdown-man),
both are not completely satisfactory:

* Ronn doesn't use standard Markdown syntax for all elements.
* kramdown-man only converts a subset of the available element types.

The new man page converter uses standard kramdown syntax and supports nearly
all element types, including tables.

This release also brings some enhancements for the GFM parser. One thing to
note is that the header ID generation is now more compatible to GFM which also
means that some IDs will be different - so check the documents on which you
use the GFM parser, especially when you are using Jekyll or Github Pages.

Organizational-wise, issues and pull requests on Github that pertain to
feature requests have been closed and are now tracked through a dedicated
kramdown project on Github.

Changes

4 minor changes:

* Add new converter for man pages
* Header ID generation for the GFM parser is now more compatible to GFM (fixes
  #267, requested by chadpowers)
* Update to the MathJax math engine to allow formatting the preview as code /
  pre > code (pull request #372 by Florian Klampfer)
* Allow tabs in table separator lines (pull request #370 by Shuanglei Tao)

2 bug fixes:

* Compactly nested lists are now handled correctly after fixing a bug in
  indentation detection (fixes #368 reported by Christopher Brown)
* GFM parser: Allow indenting the delimiting lines of fenced code blocks for
  better GFM compatibility (pull request #369 by Shuanglei Tao)

2 other fixes and enhancements:

* Added information on how to run tests to README.md (fixes #377 reported by
  Aron Griffis)
* Added information about how to use KaTeX with the MathJax math engine (fixes
  #292 reported by Adrian Sieber, information by Dato Sim�«Ñ)
@jperkin jperkin closed this as completed in 600ce87 May 4, 2017
@rmustacc
Copy link

rmustacc commented May 4, 2017

github autoclose, re-opening

@rmustacc rmustacc reopened this May 4, 2017
wiedi pushed a commit to wiedi/pkgsrc-legacy that referenced this issue Jan 28, 2018
pkgsrc changes:
- switch to using qt5 (to follow upstream)
- handle ${PREFIX} in more places
- fix package COMMENT (remove "for IBus")
- GCC_REQD 4.7 for c++11
- add patch comments

See PR pkg/52689 for more details.

Upstream changes:
(from https://github.com/google/mozc/blob/master/docs/release_history.md)

Summary of changes between 2.19.2644.102 and 2.20.2673.102 as follows.

    Third party libraries:
        None.
    Build related changes:
        --qtver GYP build option was removed (280e38f).
        Mozc for macOS now uses macOS 10.11 SDK by default (b2a74bb).
    Major changes:
        src/data/installer/credits_ja.html was removed (2ec6c8f).
        Mozc for macOS now generates 64-bit executables. 32-bit machine is no longer supported on macOS.
        Mozc for Android now has more translations (d914458).
    Fixed issues:
        TritonDataCenter#187: build_mozc.py always generates 32 bit binaries on 64 bit OSX
        TritonDataCenter#327: Switch to Qt5 from Qt4
        TritonDataCenter#348: DirectWrite may fail to render text in certain enviromnents
        TritonDataCenter#391: ImportError: gen_zip_code_seed.py
        TritonDataCenter#399: OK/Cancel buttons on Mozc key binding editor dialog cannot be clicked on Windows
        TritonDataCenter#400: Close icon on GUI dialogs do not work on Windows
    Total commits:
        30 commits.

Summary of changes between 2.18.2613.102 and 2.19.2643.102 as follows.

    Third party libraries:
        protobuf: e8ae137 -> c44ca26
        Dropped dependency on fonttools
    Build related changes:
        --qtver=5 GYP build option is implicitly assumed on macOS and Linux builds (f76c304). On Windows, --qtver=4 is still the default.
    Major changes:
        Mozc for macOS now supports 10.12 as a runtime environment.
        Mozc for Android now uses on-device font to render keytop icons (f5dcad).
    Fixed issues:
        TritonDataCenter#263: Incorrect position in voiced sound marks on the key pad in Android
        TritonDataCenter#384: HUAWEI P9 lite does not show MozcView.
        TritonDataCenter#388: Having multiple abbreviation user dictionary entries with the same reading should be supported
        TritonDataCenter#389: Emoticon user dictionary entry should not be treated a content word
    Total commits:
        39 commits.

Summary of changes between 2.17.2532.102 and 2.18.2612.102 as follows.

    Third party libraries:
        protobuf: d5fb408 -> e8ae137
        GYP: e2e928b -> 4ec6c4e
        breakpad: d2904bb -> 85b27e4
        Dropped dependency on zlib
    Build related changes:
        Renamed src/mozc_version_template.txt to src/data/version/mozc_version_template.bzl
        Reference build environment now uses Ubuntu 14.04.5 (a7cbf72)
        Reference build environment now uses Ninja 1.7.1 (d2bc62b)
        Removed --android_compiler GYP option (5ce7fa6)
        Android build requires Android NDK r12b (5ce7fa6)
    Major changes:
        Improved Store Apps compatibility on Windows (0488082)
    Fixed issues:
        NPE in UserDictionaryToolActivity.onPostResume on Android (09b47c3)
        TritonDataCenter#273: Compilation errors in Android arm64 and mips64 build
        TritonDataCenter#373: Unexpected size bloat of the APK
        TritonDataCenter#374: Duplicate candidates after Undo
        TritonDataCenter#375: 90- is suggested from 090-
        TritonDataCenter#376: Suggestion-only user dictionary entry may not work
        TritonDataCenter#377: Abbreviation user dictionary entry may not work
        TritonDataCenter#378: Suppression word may not work
        TritonDataCenter#379: Single character noun user dictionary entry may not work
        TritonDataCenter#380: Dependency on dictionary/pos_matcher.h from session/session_server.cc is missing in GYP rules
        TritonDataCenter#382: Fix typo
    Total commits:
        84 commits.

Summary of changes between 2.17.2405.102 and 2.17.2531.102 as follows.

    Third party libraries:
        protobuf: 1a59a71 -> d5fb408
    Build related changes:
        Building Mozc for Windows requires Visual Studio 2015 update 3.
        --qtdir option is no longer supported in Linux desktop build (d003076).
    Major changes:
        Updated system dictionary.
        Removed several Shift-JIS-based normalizations on Windows (26241b0).
        Mozc for Windows requires SSE2 even on 32-bit environment.
        Mozc for Windows supports Windows 7 SP1 and later only.
        Mozc for macOS supports macOS 10.9 and later only.
        Mozc for desktop platforms (Windows, macOS, desktop Linux) supports Qt5 behind --qtver=5 GYP option.
    Fixed issues:
        Fix OOM when importing too large dictionary file on Android (5c859ae)
        TritonDataCenter#298: Fix NPE on Samsung devices on showing toast
        TritonDataCenter#315: Switch to Visual C++ 2015
        TritonDataCenter#372: Discontinue the support of Windows Vista
    Total commits:
        128 commits.

Summary of changes between 2.17.2355.102 and 2.17.2404.102 as follows.

    Third party libraries:
        protobuf: 172019c -> 1a59a71
    Build related changes:
        Building Mozc requires protobuf 3.0 or later.
        --android_stl GYP option is removed in Android build. You cannot use GNU STL to build Mozc for Android anymore.
    Major changes:
        None.
    Fixed issues:
        TritonDataCenter#369: Unexpected software keyboard layout can be chosen
        TritonDataCenter#370: Mozc keeps crashing on Android N Developer Preview 5
        TritonDataCenter#371: Shortcut word in personal dictionary should not be used for multi segment conversion
    Total commits:
        50 commits.

Summary of changes between 2.17.2323.102 and 2.17.2354.102 as follows.

    Third party libraries:
        None.
    Build related changes:
        None.
    Major changes:
        2.17.2323.102 and later commits in OSS repository preserve the original CL commit date in Google internal repository.
            Consider to specify --topo-order option to git log to see commits in the actual commit order.
        Multiple performance improvements in Android.
    Fixed issues:
        None.
    Total commits:
        32 commits.

Summary of changes between 2.17.2314.102 and 2.17.2322.102 as follows.

    Third party libraries:
        googletest: 1d53731 -> 82b11b8
        WTL: 9.0.4140 -> 9.1.5321
    Build related changes:
        Building macOS binaries now requires Ninja instead of xcodebuild.
    Major changes:
        None.
    Fixed issues:
        TritonDataCenter#247: Use ninja to build Mac binaries
        TritonDataCenter#355: Native resource leak due to the missing pthread_detach call in mozc::Thread::Detach
        TritonDataCenter#361: ImmSetCandidateWindow() with CFS_EXCLUDE isn't supported on Win Vista and Win7
    Total commits:
        15 commits.
@CodingMarkus
Copy link

Why not making

VERIFIED_INSTALLATION=trusted

default, then close this issue, and everyone is happy, right?

Every browser allows users to override cert failures as well and its always up to the user to put their system at risk. And in the current state, it's impossible for users to build on packets which makes using pkgsrc more or less pointless as only a small subset of packages is pre-built and users will just use macports, fink, or brew as these provide zero hassle.

@jperkin
Copy link
Collaborator

jperkin commented Aug 23, 2021

I have a few thoughts on this, but the most important thing to say is that my binary repositories will always attempt to build everything in pkgsrc, so if a package is not available, then by definition it's currently broken, and there will be a corresponding failure log listed in the bulk build report posted to pkgsrc-bulk. This will mean that any attempt to build from source will also fail, and the user will enter the wonderful world of pkgsrc development and having to resolve all of the issues in the third-party software. Changing the value of VERIFIED_INSTALLATION will be only one of the many steps necessary to get going, and if users are serious about getting involved in pkgsrc work then they will almost certainly be better off building their own bootstrap and packages so that they are not tied in to having to reproduce my exact environment.

I have a number of issues with setting VERIFIED_INSTALLATION=trusted. First and foremost my binary package repositories are designed to be as simple and secure as possible. Giving the user the option of potentially installing compromised software isn't one I want to delegate to them. At no point should I be shipping unsigned packages, and so always is the best option so that there is a clear failure mode when attempting to install an unsigned or invalid package. It's also unclear how the pkg_add prompt would work from the pkgin environment (which certainly isn't designed for interaction), especially when running in unattended modes (e.g. pkgin -y install or even when being called from ansible/salt/chef/etc).

Since this issue was opened I've added a bunch more docs on the wiki, one of which is how to configure signed packages, so I think we can just close this now. As always if there are parts of the documentation that aren't clear then please let me know.

@jperkin jperkin closed this as completed Aug 23, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

5 participants