Re: [PATCH] t0050: remove the unused $test_case variable
On Wed, Feb 07 2018, Johannes Sixt jotted:
> Am 07.02.2018 um 09:07 schrieb Ævar Arnfjörð Bjarmason:
>>
>> On Wed, Feb 07 2018, Johannes Sixt jotted:
>>
>>> Am 07.02.2018 um 00:13 schrieb Ævar Arnfjörð Bjarmason:
The $test_case variable hasn't been used since
decd3c0c28 ("t0050-*.sh: mark the rename (case change) test as
passing", 2014-11-28) when its last user went away.
Let's remove the "say" as well, since it's obvious from subsequent
output that we're testing on a case sensitive filesystem.
>>>
>>> Am I misunderstanding the message? I think it reports properties of
>>> the test environment. And the tests do run on case-insensitive
>>> filesystems. IMO, the message should be kept.
>>
>> It's obvious from subsequent output whether the FS is case sensitive or
>> not, so I thought it was redundant to keep this report at the top since
>> we didn't have the variable setting anymore.
>
> There are test cases that do different things depending on whether the
> CASE_INSENSITIVE_FS prerequisite is set. I think it was the intent to
> report whether it is set and not whether one or the other value of the
> (now unused) variable is used somewhere.
>
> BTW, the message texts do not show which variant is taken (these are
> without your patch):
>
> On Windows:
>
> t>sh t0050-filesystem.sh
> will test on a case insensitive filesystem
> will test on a filesystem lacking symbolic links
> ok 1 - detection of case insensitive filesystem during repo init
> ok 2 - detection of filesystem w/o symlink support during repo init
> ok 3 - setup case tests
> ok 4 - rename (case change)
> ok 5 - merge (case change)
> not ok 6 - add (with different case) # TODO known breakage
> ok 7 - setup unicode normalization tests
> ok 8 - rename (silent unicode normalization)
> ok 9 - merge (silent unicode normalization)
> # still have 1 known breakage(s)
> # passed all remaining 8 test(s)
> 1..9
>
> On Linux:
>
> t@master:1002> ./t0050-filesystem.sh
> ok 1 - detection of case insensitive filesystem during repo init
> ok 2 - detection of filesystem w/o symlink support during repo init
> ok 3 - setup case tests
> ok 4 - rename (case change)
> ok 5 - merge (case change)
> ok 6 # skip add (with different case) (missing CASE_INSENSITIVE_FS)
> ok 7 - setup unicode normalization tests
> ok 8 - rename (silent unicode normalization)
> ok 9 - merge (silent unicode normalization)
> # passed all 9 test(s)
> 1..9
>
> I'd even argue that there should be messages on Linux, too.
Thanks. Let's just drop this patch. I thought it would still print out
something similar to that "missing CASE_INSENSITIVE_FS" at a quick
glance last night, but was obviously wrong.
Re: [PATCH] t0050: remove the unused $test_case variable
Am 07.02.2018 um 09:07 schrieb Ævar Arnfjörð Bjarmason:
On Wed, Feb 07 2018, Johannes Sixt jotted:
Am 07.02.2018 um 00:13 schrieb Ævar Arnfjörð Bjarmason:
The $test_case variable hasn't been used since
decd3c0c28 ("t0050-*.sh: mark the rename (case change) test as
passing", 2014-11-28) when its last user went away.
Let's remove the "say" as well, since it's obvious from subsequent
output that we're testing on a case sensitive filesystem.
Am I misunderstanding the message? I think it reports properties of
the test environment. And the tests do run on case-insensitive
filesystems. IMO, the message should be kept.
It's obvious from subsequent output whether the FS is case sensitive or
not, so I thought it was redundant to keep this report at the top since
we didn't have the variable setting anymore.
There are test cases that do different things depending on whether the
CASE_INSENSITIVE_FS prerequisite is set. I think it was the intent to
report whether it is set and not whether one or the other value of the
(now unused) variable is used somewhere.
BTW, the message texts do not show which variant is taken (these are
without your patch):
On Windows:
t>sh t0050-filesystem.sh
will test on a case insensitive filesystem
will test on a filesystem lacking symbolic links
ok 1 - detection of case insensitive filesystem during repo init
ok 2 - detection of filesystem w/o symlink support during repo init
ok 3 - setup case tests
ok 4 - rename (case change)
ok 5 - merge (case change)
not ok 6 - add (with different case) # TODO known breakage
ok 7 - setup unicode normalization tests
ok 8 - rename (silent unicode normalization)
ok 9 - merge (silent unicode normalization)
# still have 1 known breakage(s)
# passed all remaining 8 test(s)
1..9
On Linux:
t@master:1002> ./t0050-filesystem.sh
ok 1 - detection of case insensitive filesystem during repo init
ok 2 - detection of filesystem w/o symlink support during repo init
ok 3 - setup case tests
ok 4 - rename (case change)
ok 5 - merge (case change)
ok 6 # skip add (with different case) (missing CASE_INSENSITIVE_FS)
ok 7 - setup unicode normalization tests
ok 8 - rename (silent unicode normalization)
ok 9 - merge (silent unicode normalization)
# passed all 9 test(s)
1..9
I'd even argue that there should be messages on Linux, too.
-- Hannes
Re: [PATCH] t0050: remove the unused $test_case variable
On Wed, Feb 07 2018, Johannes Sixt jotted:
> Am 07.02.2018 um 00:13 schrieb Ævar Arnfjörð Bjarmason:
>> The $test_case variable hasn't been used since
>> decd3c0c28 ("t0050-*.sh: mark the rename (case change) test as
>> passing", 2014-11-28) when its last user went away.
>>
>> Let's remove the "say" as well, since it's obvious from subsequent
>> output that we're testing on a case sensitive filesystem.
>
> Am I misunderstanding the message? I think it reports properties of
> the test environment. And the tests do run on case-insensitive
> filesystems. IMO, the message should be kept.
It's obvious from subsequent output whether the FS is case sensitive or
not, so I thought it was redundant to keep this report at the top since
we didn't have the variable setting anymore.
Re: [PATCH] t0050: remove the unused $test_case variable
Am 07.02.2018 um 00:13 schrieb Ævar Arnfjörð Bjarmason:
The $test_case variable hasn't been used since
decd3c0c28 ("t0050-*.sh: mark the rename (case change) test as
passing", 2014-11-28) when its last user went away.
Let's remove the "say" as well, since it's obvious from subsequent
output that we're testing on a case sensitive filesystem.
Am I misunderstanding the message? I think it reports properties of the
test environment. And the tests do run on case-insensitive filesystems.
IMO, the message should be kept.
Signed-off-by: Ævar Arnfjörð Bjarmason
---
t/t0050-filesystem.sh | 8
1 file changed, 8 deletions(-)
diff --git a/t/t0050-filesystem.sh b/t/t0050-filesystem.sh
index b29d749bb7..606ffddd92 100755
--- a/t/t0050-filesystem.sh
+++ b/t/t0050-filesystem.sh
@@ -7,14 +7,6 @@ test_description='Various filesystem issues'
auml=$(printf '\303\244')
aumlcdiar=$(printf '\141\314\210')
-if test_have_prereq CASE_INSENSITIVE_FS
-then
- say "will test on a case insensitive filesystem"
- test_case=test_expect_failure
-else
- test_case=test_expect_success
-fi
-
if test_have_prereq UTF8_NFD_TO_NFC
then
say "will test on a unicode corrupting filesystem"
