- User Since
- Jun 1 2018, 2:58 AM (16 w, 5 d)
Sun, Sep 23
@mpickering Thank you for the hint about check_main. The patch is now much simpler. I also tested with the -fobject-code flag and switching between GHC and GHCi and I didn't get any error messages.
- Merge branch 'master' into T11647
- Improve patch for Trac #11647
Thu, Sep 20
Thanks for this code review!
I'll rework this patch.
Wed, Sep 19
Sun, Sep 16
@osa1 I fixed the failing tests. Please note the changing version numbers for packages containers and ghc.
- Merge branch 'master' into T13862
- T13862 Fix failing tests due to new message
Thu, Sep 13
@osa1 Thanks for your comments.
- T13862 Back out new DynFlag
- T13862 Change text of error message.
Sat, Sep 8
Why not simply change the error message to:
Use -v (or :set -v` in ghci) to see a list of the files searched for. `
Mon, Sep 3
Sun, Sep 2
Sat, Sep 1
I now changed to the error message, that's more consistent with the error message issued on a LANGUAGE pragma error.
- Error msg more consistent with LANGUAGE pragma error (Trac #15053)
Thu, Aug 30
Ah so you wanted to update languagePragParseError too? Why?
@osa1 I started with modifying the languagePragParseError message. But as an non native English speaker I obviously was not very successful... I'll wait 3 or 4 days for additional comments and then decide.
@Phyx Can you live with the message text proposed by Ömer?
Wed, Aug 29
@Phyx Thanks for your nice text. I changed to your version. I added a colon (:) after Input was. PLease inform me, if you dont like it.
@tdammers Thanks for your comment. In the new error message text, there is no more something like non-parasable.
- Use nicer text for error message (Trac #15053)
Aug 25 2018
- Improve wording of error messsage (Trac #15053)
Aug 23 2018
Agreed: white-space separated is much better! My version of the error message is just a first proposal.
Aug 18 2018
For the changes requested by nomeata see: https://phabricator.haskell.org/D5076
Aug 13 2018
@nomeata Thanks for your comment!
Aug 12 2018
@thomie Thanks for the hint. I have added a line to the 8.8.1 release notes.
- Add little note to release guide 8.8.1 (Trac #10869)
Aug 9 2018
Aug 8 2018
I guess specifying -E together with -keep-hscpp-files will write both files?
@thomie The patch is now much shorter. Thanks for your kind help.
- Improve patch after code review (Trac #10869).
- Rename flag form keep-hspp-files to kepp-hscpp-files.
- Do not copy the file but reuse code from the keep-hc/s/llvm-files flags.
Aug 7 2018
@thomie: Many thanks for your code review and your kind hints to improve the patch.
Aug 3 2018
@dfeuer: I added 2 additonal tests. The first as you suggested with the same recursive function twice with 2 different names. The second one with 2 different functions with the same subfunction in the where clause. I also renamed the existing test.
Please inform me, if you see other possibilites to improve the tests for Trac #9441
- Add two additional tests for Trac #9441
Aug 2 2018
Aug 1 2018
Jul 29 2018
Many thanks for your kind help. I have now fixed all your issues. Please inform me, if you see other necessary changes.
- Fix remaining issues from code review (Trac #12674)
Jul 28 2018
Thanks for the code review and your comments. Please look at my replies and feel free to take your decisions.
- Improve tests (Trac #12674)
Jul 26 2018
Jul 21 2018
Uiuii! This is an intermittend problem! I tested under Windows: With backslash it worked, with forward slash it didin't, so I assumed the testcase was wrong. But obviously I didn't test enough!!
Thanks for your kind remarks and sorry for bothering you!
Jul 20 2018
Please inform me, if these 2 tests can be written with less redundandency. The only difference is in the Makefile: Run the program ./T9405 (Linux) vs .\T9405 (Windows).
Reset filemodes back to 644
Jul 6 2018
Oops, the failure of testcase cabal01is severe! To fix it, we would need to change Cabal: Cabal should be aware of the new location of the *.o file. However this gives a dependency between the versions of Cabal and GHC. The user of a GHC version with the patch for Trac Trac #14025 must also use a version of Cabal with a patch for Trac #14025 (and vice versa).
Considering, that Trac #14025 solves a very tiny problem, this is a too heavy burden on the user. Therefore I abandon this solution and I'll develop a new solution, just fixing the case with the absolute path.
Jul 3 2018
Jun 25 2018
Jun 20 2018
I'll do this. But this week I'm away from my desktop computer. So maybe
it will take till mid next week.
Thanks and best wishes. Roland
Jun 17 2018
@bgamari : You wrote:
I added the requested changes by bgamari and by the linter. Please feel free to request additional changes!
This is my first contribution to GHC, where I wrote real Haskell code. Therefore this is just a draft version and it needs some additional improvments.