Packaging for ubuntu bionic #2
Open
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…where one band is larger than 2 GB (2.3 regression)
…entification of raw binary formats such as ENVI (fixes OSGeo#1961)
…d falsely return true if the point was aligned with a segment, but not between the nodes. Impact correct reconstruction of holes in shapefile driver
…a items whose value is a list on several lines
…lue to NoData on Float32 rasters (fixes OSGeo#1987)
…hreaded reads (fixes OSGeo#1244) Kudos to @tbonfort for the easy reproducer.
…hread which created it (fixes OSGeo#1989). This workarounds a PROJ bug also fixed per OSGeo/PROJ#1726
…SGeo#2011) When writing several GDAL datasets in parallel from several threads, a thread can force a dirty block of another thread/dataset to be written. By doing so, it will remove it from the list of cached blocks, so if the thread that explictly manipulates this dataset reacquire the block, it will see an old version of the block. The effect of this can be particularly seen with GeoTIFF files with multi band pixel leaving organization, where acquiring a block of a band causes corresponding blocks of other bands to be cached. The main fix done here is in GDALRasterBlock::Internalize() to avoid evicting a dirty block of another dataset when there are other candidates. This considerably reduces the likelihood of the bug. Another fix is done in GDALDataset::EnterReadWrite() to wait for the completion of the flushing of pending dirty blocks of the dataset before reading new blocks. I believe there could still be problems if the block cache size is too small regarding the number of threads.
…ce that has masks and overviews
… whose content is not a temperature, but a derived quantity, such as spread
…) = 'foo' to ES query language
…CTER(4096)) = 'foo'
…6b1c. Does not affect released versions. Fixes https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=19057
…ter count capability
…eRPCTransformer() Cherry-picked from c95b091
craigds
force-pushed
the
kx-2.4-bionic
branch
3 times, most recently
from
4db9730
to
c49ab23
Compare
rcoup
approved these changes
Mar 10, 2020
|
|
||
| # skip known failures | ||
| rm gcore/rfc30.py | ||
| TRAVIS=YES TRAVIS_BRANCH=trusty pytest -v -k 'not (test_ogr_fgdb_13 or test_vsisync or test_vsis3_init or test_vsis3_1 or test_vsis3_cleanup or test_eedai_GOOGLE_APPLICATION_CREDENTIALS)' | ||
| TRAVIS=YES TRAVIS_BRANCH=ubuntu_1804 pytest -v -k 'not (test_ogr_fgdb_13 or test_vsisync or test_vsis3_init or test_vsis3_1 or test_vsis3_cleanup or test_eedai_GOOGLE_APPLICATION_CREDENTIALS)' |
There was a problem hiding this comment.
can we review the not list here based on what actually fails under bionic?
gdal/debian/rules
Outdated
| --with-java=/usr/lib/jvm/default-java \ | ||
| --with-perl \ | ||
| --with-python \ | ||
| --with-kea=/usr/bin/kea-config \ | ||
| --with-jpeg12 \ |
There was a problem hiding this comment.
12-bit jpeg support (jpeg12) is default now?
…00 < side < 400px
…pped to the input data-type. Includes a test Cherry-picked from ddc8ac0
* Update build config to 2.3.0 * Add extra VSI flags to link GDAL statically. As recommended by https://trac.osgeo.org/gdal/wiki/VSIPreload * Use aptly-uploader to uplaod to aptly * Make python[3]-gdal and gdal-bin require exact version of libgdal20 * Remove vsipreload: In libgdal-dev, it conflicts with the old libgdal1h package, which we still need (postgis depends on it). We also don't use it any more. * CI via buildkite * Don't strip symbol tables from our debian builds
This fixes an issue where a keep-alive'd connection via a proxy is closed by the server, and subsequent range requests fail because the proxy's headers are treated as the response headers. In Curl>=7.54 there's an option to suppress passing proxy response headers through to the application, but for older libcurl versions we add a reasonably conservative workaround.
Easier to fake buildkite builds locally
Not the newest, but I got errors building against 8.0.1 and I want a closer-to-trusty version to make this simpler
Merged most of gdal_2.4.2+dfsg-1~bionic0.debian.tar.xz from https://launchpad.net/~ubuntugis/+archive/ubuntu/ppa/+packages
We (trusty) used to have it implicitly because libpq-dev depended on it, but it no longer does in bionic.
Bionic packaging removes the rpath from the libraries, but we need it for our kakadu and filegdb libs.
craigds
force-pushed
the
kx-2.4-bionic
branch
from
56972ea
to
194e54c
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
(no need to actually merge this; it can live as a separate branch for now so we can easily build either branch in CI).
This contains packaging/build changes for ubuntu bionic, for the kx-2.4 branch.
todo