pakcs merge requestshttps://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests2023-04-18T07:40:49Zhttps://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/34Add hie.yaml for the Haskell langauge server2023-04-18T07:40:49ZFredrik WieczerkowskiAdd hie.yaml for the Haskell langauge serverSimilar to curry-frontend!62, configuring the language server explicitly makes it easy to work on the frontend, in this case from the parent `pakcs` repository. Since the language server only inspects the `hie.yaml` at the root of the wo...Similar to curry-frontend!62, configuring the language server explicitly makes it easy to work on the frontend, in this case from the parent `pakcs` repository. Since the language server only inspects the `hie.yaml` at the root of the workspace, we need one here too.
The language server's implicit configuration unfortunately is not sufficient for this: In the `pakcs` repository it will default to using the system GHC rather than Stack, presumably since it cannot find a `stack.yaml` in the root of the repository. This causes errors if the user's system GHC configuration differs in any way from the declared Stack configuration.https://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/33Add .vscode and .DS_Store to gitignore2023-01-30T14:12:23ZFredrik WieczerkowskiAdd .vscode and .DS_Store to gitignoreThis prevents these automatically generated files from being accidentally committed.This prevents these automatically generated files from being accidentally committed.https://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/32Update frontend to most recent master (using GHC 9.2)2022-12-16T10:51:48ZFredrik WieczerkowskiUpdate frontend to most recent master (using GHC 9.2)Since GHC 9.2 ships a native ARM macOS code generator, this makes it a lot easier to build PAKCS on ARM macOS.
Marked as a draft, since Stack cannot seem to find a suitable GHC 9.2.5 distribution in CI, perhaps the Stack version in the ...Since GHC 9.2 ships a native ARM macOS code generator, this makes it a lot easier to build PAKCS on ARM macOS.
Marked as a draft, since Stack cannot seem to find a suitable GHC 9.2.5 distribution in CI, perhaps the Stack version in the Docker image is too old?https://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/31Read home directory from USERPROFILE on Windows2022-10-19T15:17:37ZFredrik WieczerkowskiRead home directory from USERPROFILE on WindowsA small change that falls back to `USERPROFILE` (the default environment variable for the user's home directory on Windows) if `HOME` is unset.A small change that falls back to `USERPROFILE` (the default environment variable for the user's home directory on Windows) if `HOME` is unset.https://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/29Move generatePrologBasics into Makefile2022-10-19T15:14:56ZFredrik WieczerkowskiMove generatePrologBasics into MakefileThe motivation for this change is twofold:
- First, we get a cleaner separation of concerns, since `sicstusbasics.pl` no longer contains shell commands that describe how to copy itself around, instead the Makefile deals with this.
- Sec...The motivation for this change is twofold:
- First, we get a cleaner separation of concerns, since `sicstusbasics.pl` no longer contains shell commands that describe how to copy itself around, instead the Makefile deals with this.
- Secondly, under MinGW on Windows `make` (including recursive invocations) has access to Unix tools like `sed` or `cat`, but subprocess calls from `sicstus` do not (since `sicstus` uses the native Microsoft VC runtime rather than a Unix emulation). This resulted in errors when trying to call these tools from `generatePrologBasics`, which we sidestep by moving the logic into the Makefile directly.https://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/28Quote variables where possible in scripts2022-10-19T15:12:26ZFredrik WieczerkowskiQuote variables where possible in scriptsCurrently there are few subtle issues where the scripts won't deal with paths containing spaces correctly. This is especially problematic on Windows where paths like `C:\Program Files\...` are common and e.g. often already contained by t...Currently there are few subtle issues where the scripts won't deal with paths containing spaces correctly. This is especially problematic on Windows where paths like `C:\Program Files\...` are common and e.g. often already contained by the path to the user's Prolog interpreter. This PR attempts to fix this by quoting as may path-ish variables as possible (which is good practice anyway since using unquoted expansions in combination with commands like `rm` can e.g. lead to unexpected deletions).https://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/27Replace 'which' calls with 'command -v'2022-10-19T15:10:39ZFredrik WieczerkowskiReplace 'which' calls with 'command -v'In an effort to make PAKCS compatible with more operating systems, e.g. Windows, this PR replaces `which` with `command -v`. This both avoids spawning a process to check for the path of executables and subtle inconsistencies in how failu...In an effort to make PAKCS compatible with more operating systems, e.g. Windows, this PR replaces `which` with `command -v`. This both avoids spawning a process to check for the path of executables and subtle inconsistencies in how failures are treated (see [this Stackoverflow post for further rationale](https://stackoverflow.com/questions/592620/how-can-i-check-if-a-program-exists-from-a-bash-script)).
A specific example of where the previous/current implementation caused trouble: While most `which` implementations (e.g. on Linux/macOS) will avoid output when the executable could not be found and no tty is attached, `which` on MinGW/Windows won't and thus assign the error text as if it were the path to various variables:
```
which: no <executable> in (<user PATH>)
```https://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/26Update frontend (GHC 8.10)2021-11-12T16:06:14ZFredrik WieczerkowskiUpdate frontend (GHC 8.10)See curry-frontend!49.See curry-frontend!49.https://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/25fix a CI todo and some cleanup2021-07-16T13:10:17ZBennet Bleßmannstu201758@mail.uni-kiel.defix a CI todo and some cleanup* main point:
[Issue 273034](https://gitlab.com/gitlab-org/gitlab/-/issues/273034) closed since I last had a look
this allows to generate nighly and test versions under the same package as releases,
but with the version contain...* main point:
[Issue 273034](https://gitlab.com/gitlab-org/gitlab/-/issues/273034) closed since I last had a look
this allows to generate nighly and test versions under the same package as releases,
but with the version containing a corresponding label and build metadata
* cleanup changes that should have been part of 5d67c555
* fixed new release using the wrong date format (yyyy-mm-dd instead of dd/mm/yyyy)
* also move the link generation for nightly from `update_version.sh` to `release_helper.sh` (as 5d67c555 did for release)https://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/24Fix failing "Test Download" scheduled Pipeline job2021-06-14T13:04:39ZBennet Bleßmannstu201758@mail.uni-kiel.deFix failing "Test Download" scheduled Pipeline jobAfter my last CI fix, one of the daily scheduled Test Download jobs started to fail as it implicitly relied on the old cabal behaviour which was changed in the new version.
This tries to fix that by switching these jobs back to the old i...After my last CI fix, one of the daily scheduled Test Download jobs started to fail as it implicitly relied on the old cabal behaviour which was changed in the new version.
This tries to fix that by switching these jobs back to the old image.
This also explicitly sets the main image rather than using latest, this should make changing the image in the future easier, as one can create and push the new image without breaking the existing setup and then make a commit to change the selected image, taking advantage of CI to test image change explicitly.https://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/23upgrade ghc in ci image2021-06-08T13:33:23ZBennet Bleßmannstu201758@mail.uni-kiel.deupgrade ghc in ci imagehttps://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/22mention `fix_asset_link.sh` in CI Readme2021-03-17T17:21:04ZBennet Bleßmannstu201758@mail.uni-kiel.demention `fix_asset_link.sh` in CI Readmehttps://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/21generate link, where we have all the required information2021-02-16T20:34:23ZBennet Bleßmannstu201758@mail.uni-kiel.degenerate link, where we have all the required informationthis way `update_version.sh` and `gitlab_release.sh` can't easily diverge form another and `release_helper.sh`this way `update_version.sh` and `gitlab_release.sh` can't easily diverge form another and `release_helper.sh`https://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/20fix generated release links2021-02-15T23:39:22ZBennet Bleßmannstu201758@mail.uni-kiel.defix generated release linksadd script to fix release perma linksadd script to fix release perma linkshttps://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/19Fix incorrect path for creating pakcs vX.Y.Z.version files in the curry-lang....2021-02-08T00:07:02ZBennet Bleßmannstu201758@mail.uni-kiel.deFix incorrect path for creating pakcs vX.Y.Z.version files in the curry-lang.org repohttps://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/18fix error for tag release pipeline2021-02-05T23:41:36ZBennet Bleßmannstu201758@mail.uni-kiel.defix error for tag release pipelinehttps://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/17update ci readme2021-02-02T23:41:38ZBennet Bleßmannstu201758@mail.uni-kiel.deupdate ci readmehttps://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/16Split run_download_make CI job and adjust custom section title style to match...2021-01-30T20:52:54ZBennet Bleßmannstu201758@mail.uni-kiel.deSplit run_download_make CI job and adjust custom section title style to match build in sectionshttps://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/15Add custom collapsible section to ci log2021-01-28T23:20:36ZBennet Bleßmannstu201758@mail.uni-kiel.deAdd custom collapsible section to ci loghttps://git.ps.informatik.uni-kiel.de/curry/pakcs/-/merge_requests/14Master setup ci2020-12-21T22:00:05ZBennet Bleßmannstu201758@mail.uni-kiel.deMaster setup ciCI Einrichtung für den Master.
Einige Änderungen an den Makefiles damit manche Fehler in der CI zum Abbruch führen statt ignoriert zu werden.
Aktuell wird noch ein Docker Image unter meinen Docker-Hub Account genutzt, diese sollte wahrs...CI Einrichtung für den Master.
Einige Änderungen an den Makefiles damit manche Fehler in der CI zum Abbruch führen statt ignoriert zu werden.
Aktuell wird noch ein Docker Image unter meinen Docker-Hub Account genutzt, diese sollte wahrscheinlich irgendwann nochmal umgestellt werden.