I've been trying to get the tests running on Windows; the latest roadblock was test11. test11 works correctly using vim.exe from the officially built binary distribution (7.3.48), but not on vim.exe built from the latest source (7.3.496). I finally got to play with 'hg bisect', very fun.
This patch does three things, all of them related and necessary:
1. Revert the only not-100%-straightforward change that was in patch 124.
Patch 124 was mainly about moving the "did-i-have-an-eol-when-i-loaded"
setting from global to per-buffer, but it also changed the point at which
the setting was reset. This patch undoes that last part. There was a
comment about the line number becoming invalid due to edits, which I didn't
follow up on.
2. Correctly sets has("filterpipe") as "false" on the Windows console,
leaving it enabled in the GUI as intended by patch 240.
Almost all the code in patch 240 was guarded with FEAT_GUI_W32; for some
reason the change to add has("filterpipe") wasn't. I took a look at
generalizing the pipe code to cover the console, but it's very GUI-specific
as written. For the time being the console doesn't do pipes.
3. In test11, set shelltemp before filtering through gzip if
has("filterpipe") is false.
These three things were interacting in nefarious ways; I'll spare you the
agony, but suffice to say the automated tests act very strangely when
filters, piping, gzip, and self-modifying tests interact...
I suspect this patch fixes a number of potential filtering issues on the
Windows console.
----
David Pope
--
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php
fix-windows-console-filterpipe-and-eol-save.patch
Description: Binary data
