12:58 a.m.
Repo: git://git.gitano.org.uk/gitano.git
Ref: dsilvers/fix-272
SHA1: ec642db685d98a9fa264511a5c5ae0786ac0339d
When I consolidated the behaviour of multiple-value Lace values, I failed to
take into account that some of the values are both set-type and also list-type.
This results in a huge nasty traceback when the owner of a repo pushes an
adminref change containing a hook when using the default ruleset. This pair
of commits first exposes the problem by tweaking one of the treedelta tests
and then corrects it.
Daniel Silverstone (2):
TESTING: Use prefix instead of is
LACE: Correct a bug in iterating table form values
lib/gitano/lace.lua | 5 ++++-
testing/admin-patches/must-affect-FOO.patch | 2 +-
2 files changed, 5 insertions(+), 2 deletions(-)
--
2.11.0
12:58 a.m.
New subject: [PATCH 1/2] TESTING: Use prefix instead of is
In order to increase coverage of the 'prefix' match type in
list context, this change alters one of the many treedelta
related admin patches.
---
testing/admin-patches/must-affect-FOO.patch | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/testing/admin-patches/must-affect-FOO.patch
b/testing/admin-patches/must-affect-FOO.patch
index 07c570e..81cce3c 100644
--- a/testing/admin-patches/must-affect-FOO.patch
+++ b/testing/admin-patches/must-affect-FOO.patch
@@ -6,7 +6,7 @@ index aafa17c..388766a 100644
# Okay, if we're altering the admin ref, in we go
include global:adminchecks is_admin_ref
-+deny "Needs a FOO" op_is_normal ![treediff/targets is FOO]
++deny "Needs a FOO" op_is_normal ![treediff/targets prefix FOO]
+
# Now we're into branch operations. Owners can do any normal operation
# Normal ops are create/delete/fastforward on refs
--
2.11.0
12:58 a.m.
New subject: [PATCH 2/2] LACE: Correct a bug in iterating table form values
When a tabular value is given to Lace, we assume that it is in 'set'
form (i.e. the keys are the important thing to match against) and
as such, we need to ensure we skip non-string keys since tables might
be dual form (set, and list).
---
lib/gitano/lace.lua | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/lib/gitano/lace.lua b/lib/gitano/lace.lua
index 685d625..3bea565 100644
--- a/lib/gitano/lace.lua
+++ b/lib/gitano/lace.lua
@@ -137,7 +137,10 @@ local function _do_simple_match(ctx, key, matchtype, value)
else
local ret = false
for k in pairs(kk) do
- ret = ret or check(value, k)
+ -- Only bother if the key is a string (set-form)
+ if type(k) == "string" then
+ ret = ret or check(value, k)
+ end
end
if matchtype:sub(1,1) == "!" then
ret = not ret
--
2.11.0
4:57 a.m.
Looks good and makes sense to me.
It changes the space indentation to tabs in lib/gitano/lace.lua, but
that's not a big concern to me.
On 18 March 2018 at 15:58, Daniel Silverstone
<dsilvers(a)digital-scurf.org> wrote:
Repo: git://git.gitano.org.uk/gitano.git
Ref: dsilvers/fix-272
SHA1: ec642db685d98a9fa264511a5c5ae0786ac0339d
When I consolidated the behaviour of multiple-value Lace values, I failed to
take into account that some of the values are both set-type and also list-type.
This results in a huge nasty traceback when the owner of a repo pushes an
adminref change containing a hook when using the default ruleset. This pair
of commits first exposes the problem by tweaking one of the treedelta tests
and then corrects it.
Daniel Silverstone (2):
TESTING: Use prefix instead of is
LACE: Correct a bug in iterating table form values
lib/gitano/lace.lua | 5 ++++-
testing/admin-patches/must-affect-FOO.patch | 2 +-
2 files changed, 5 insertions(+), 2 deletions(-)
--
2.11.0
_______________________________________________
gitano-dev mailing list
gitano-dev(a)gitano.org.uk
http://listmaster.pepperfish.net/cgi-bin/mailman/listinfo/gitano-dev-gita...
10:26 p.m.
On Wed, May 16, 2018 at 20:57:21 +0100, Richard wrote:
Looks good and makes sense to me.
Excellent.
It changes the space indentation to tabs in lib/gitano/lace.lua, but
that's not a big concern to me.
Oops. I've fixed that and merged the change, thank you for the review.
D.
--
Daniel Silverstone http://www.digital-scurf.org/
PGP mail accepted and encouraged. Key Id: 3CCE BABE 206C 3B69
1935
days inactive
2015
days old
4 comments
2 participants
participants (2)
-
Daniel Silverstone -
Richard