- User Since
- Nov 17 2014, 8:23 PM (252 w, 3 d)
Oct 29 2017
What is (State# s -> r)? It looks like the State# s token is being consumed by the callback, which seems a bit odd.
May 21 2017
Improve docs + make inheritance actually work
May 20 2017
May 19 2017
May 15 2017
May 12 2017
Remove guards in favor of patterns
May 8 2017
I tweaked the testlib a little bit to allow changing the whitespace-normaliser.
Fix fixWhitespace and also the tests
Hm, it seems testing is not as straightforward – I think the stderr normaliser turns tabs into 8 spaces? So it can't really tell the difference between the two…
Apr 12 2017
Apr 11 2017
Apr 9 2017
Mar 18 2017
Unhide 'bold' from Outputable
Mar 12 2017
Mar 11 2017
Mar 10 2017
I made some tweaks to hsc2hs to work around this problem. This means GHC will build fine even with an older GHC, albeit with warnings about missing pragmas.
Disambiguate Read type as Integer
Heh, I was actually expecting more … critique here, considering this is the first time I have ever messed with a lexer :P
Remove unrelated comment
Jan 26 2017
Jan 25 2017
Commit to /bin/bash and re-enable stderr validation
Jan 16 2017
I wrote this patch to preempt concerns regarding the performance impact of D2718. Given that the performance of GHC is trending concern, I felt I would doing a disfavor in having submitted D2718 that has the potential to worsen the status quo.
Jan 10 2017
Update test stderrs
Jan 9 2017
Unfilter qualified imports of Pretty + try to improve docs of printForC
Jan 6 2017
Dec 19 2016
Thanks everyone for reviewing this! :)
Dec 17 2016
Split caching into a separate differential
Cache files in buffers to improve efficiency + Add margin and ellipsis
Migrate to SysTools.Terminal
Improve FastString docs
Dec 16 2016
ignore_stderr in topHandler03
I tested this on an Ubuntu box and got the same result as yours.
Dec 13 2016
Dec 10 2016
Had to coerce the arc linter scripts into playing nicely with the binary files.
Dec 9 2016
Ensure width is at least 1
Add tests and tweak style slightly
Dec 8 2016
I went with the first approach. The code now should be able to detect MinTTY with little fear of false positives (assuming no-one would go out of their way to name their non-TTY pipes \msys-fake-pty0-to-master).
Forgot to divide by sizeof(wchar_t)
Use GetFileInformationByHandleEx to detect Cygwin/MSYS terminals
Ignore stderr for API annotations
Oh wait I found another one: https://fossies.org/linux/vim/src/iscygpty.c
After I finally got this to build on Windows, I discovered that moving the “is a TTY?” test before the environment variable tests means that on non-Windows terminals (such as MinTTY) the test will never succeed. This is because MinTTY does not implement a true Windows console, but a fake one using pipes. Therefore, Windows API has no idea that the pipe does indeed lead to a terminal.
Other broken tests that aren't fixed in this patch:
Dec 7 2016
Remove accidental infinite loop