[pacman-dev] [PATCH] pacman: better warning message when skipping duplicate targets

[Michael Straube]

Am 11.12.18 um 20:51 schrieb Andrew Gregory:
> On 12/11/18 at 06:14pm, Michael Straube wrote:
>> Am 09.12.18 um 19:47 schrieb Andrew Gregory:
>>> On 12/09/18 at 06:31pm, Michael Straube wrote:
>>>> Change the warning message to reflect the reason when skipping duplicate
>>>> targets. skipping target -> skipping duplicate target
>>>>
>>>> FS#49377
>>>>
>>>> Signed-off-by: Michael Straube <michael.straube at posteo.de>
>>>> ---
>>>>    src/pacman/remove.c | 2 +-
>>>>    src/pacman/sync.c   | 2 +-
>>>>    2 files changed, 2 insertions(+), 2 deletions(-)
>>>
>>> Should we just remove the error altogether and move the message to
>>> DEBUG?  The user doesn't need to do anything in response to it and
>>> I can't think of any reason a front-end would want to actually die
>>> from it.  It seems to just be useless line noise that requires
>>> front-ends to check for it specifically just to ignore it.
>>>
>>
>> Sounds reasonable. On the other hand I can imagine that some people
>> would complain that too much from what's going on is hidden from the
>> user. What do others think?
>>
>>
>> P.S.: You mean?
>> pm_printf(ALPM_LOG_DEBUG, _("skipping duplicate target: %s\n"), target);
> 
> I realized after sending this that adding a duplicate could actually
> be an error if they are two separate packages with the same name.  So,
> I'm going to say to add a check in add_pkg for whether the duplicate
> is actually the same package (simple pointer cmp) and, if they are the
> same, log a debug message and return success, if they are different
> return the error as we do now.  process_pkg in pacman should then
> treat that error just like any other instead of printing a warning and
> continuing on.  alpm_remove_pkg should just log the debug message and
> return success.
> 
> apg
> 

Ok, I will send a patch the next days.