On 09/25/2016 12:37 AM, Matt Mackall wrote:
On Sat, 2016-09-24 at 00:23 +0200, Pierre-Yves David wrote:
On 09/22/2016 08:21 PM, Matt Mackall wrote:
# HG changeset patch
# User Matt Mackall <m...@selenic.com>
# Date 1474294560 18000
# Mon Sep 19 09:16:00 2016 -0500
# Node ID 6db1b2e7d19ed317404c1275db46780d40ececb8
# Parent 133b35066bef5d9c31d13c4f0b2e4a50d1ceae87
extdata: add basic template support
This lets us expose the per-entry "freeform data".
I guess the same feedback apply here. Maybe we should go for a single
generic template name (with argument) and let user create alias if they
really needs to. What do you think?
Again, the registrar framework isn't a good fit
I think the registrar framework is mostly intended for extensions. Where
the module is loaded for sure and adding new template/revset requires to
access with the internal data structure.
diff -r 133b35066bef -r 6db1b2e7d19e mercurial/templater.py
--- a/mercurial/templater.py Mon Sep 19 09:13:11 2016 -0500
+++ b/mercurial/templater.py Mon Sep 19 09:16:00 2016 -0500
@@ -19,6 +19,7 @@
parser,
registrar,
revset as revsetmod,
+ scmutil,
templatefilters,
templatekw,
util,
@@ -352,6 +353,13 @@
try:
v = context.process(key, safemapping)
except TemplateNotFound:
+ # check for extdata last
+ if 'ctx' in mapping:
+ ctx = mapping['ctx']
Could we use 'get' here:
ctx = mapping.get('ctx')
if ctx is not None:
I can't imagine why that would be preferable.
Using the key only once make the code more maintainable. I find it a bit
more pythonic.
That was a small nits and its not very important. I'm fine with the
current form if you prefer it.
Cheers,
--
Pierre-Yves David
_______________________________________________
Mercurial-devel mailing list
Mercurial-devel@mercurial-scm.org
https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel