[pacman-dev] [patch] makepkg3 -- cleanup dep check functions

Andrew Fyfe andrew at neptune-one.net
Fri Jun 1 12:16:05 EDT 2007

Dan McGee wrote:
> On 6/1/07, Andrew Fyfe <andrew at neptune-one.net> wrote:
>> Dan McGee wrote:
>>> On 5/30/07, Andrew Fyfe <andrew at neptune-one.net> wrote:
>>>> Dan McGee wrote:
>>>>> On 3/24/07, Andrew Fyfe <andrew at neptune-one.net> wrote:
>>>>>> Another patch :)
>>>>>> This one targets the 3 dep check functions (handledeps, resolvedeps,
>>>>>> removedeps). The bulk of the changes are whitespace and removing the
>>>>>> repetitive if..then...else checks. Also some bits have been reordered to
>>>>>> stop functions/external programs being called unnecessarily.
>>>>>> This still includes the fakeroot hacks, if you need it rediffed without
>>>>>> the fakeroot hacks, I'll post a new patch.
>>>>>> Andrew
>>>>> Cleaning out the email inbox...if this still applies and we haven't
>>>>> already done the cleanups, can you push it to your git tree?
>>>>> -Dan
>>>> Yes, I'll apply this to my working branch and push it tonight.
>>>> Andrew
>>> Andrew,
>>> This one disappeared on me before I had a chance to look it over...any
>>> chance you still have it sitting around in a branch? If not, 'git
>>> lost-found' may be your friend, it may help you find it even if you
>>> thought you deleted it.
>>> -Dan
>> Oops, I was using working for something else, I've put the makepkg
>> branch up, all the patches are there.
>> Andrew
> Yet another email on this thread...oh well. I just wanted to point out
> that I like the following notation:
> I think having individual error codes defined may be a bit extreme,
> but having a notation why we are bailing out is a good idea at least
> in the code. I don't want to get too many of these codes, but things
> like missing program, bad PKGBUILD, build failure, missing deps, etc.
> are all good but not too-specific reasons. Perhaps you could submit a
> patch with a brief comment header containing these so people hacking
> on makepkg know what they have to choose from?
> -Dan
> _______________________________________________
> pacman-dev mailing list
> pacman-dev at archlinux.org
> http://archlinux.org/mailman/listinfo/pacman-dev

Sure, off the top of my head the original list I had was

$E_USER_ABORT (yn_question() never went in so this isn't needed).

As far as error codes go as long as we keep 2 for build failures, 
everything else can use 1. Build failures are one thing people need to 
know about when using makepkg in a script so giving it it's own error 
code is a good idea.

I'll do up a patch with some description of the error codes.


More information about the pacman-dev mailing list