Re: Depends and build depends in receipt

[Christopher Rogers <slaxemulator@xxxxxxxxx>]

> 3. Proposal standard: Minimum automatism but remove redudency since
> they are implicit, you need libusb if libusb-dev is in build dep:
>
> DEPENDS="libusb pcre perl"
> BUILD_DEPENDS="libusb-dev pcre-dev gtk+-dev"

Why not this:

DEPENDS="libusb pcre perl"
BUILD_DEPENDS="gtk+-dev"

It should look add depends and add the there build depends. At least that will fix some of the problems.  And be more line with what tazwok does.

On Tue, May 10, 2011 at 11:50 AM, Christopher Rogers <slaxemulator@xxxxxxxxx> wrote:

I think know the problem. Its tazwok update-wok is it? That may make since cause that run everything time to update the wok.

The reason update-wok take so long was cause it was design for woks that were not mercurial repos. It checked all files in wok and made a md5sum list.

So i think a tazwok update-wok --hg would fix the problem. It will just copy update the clean-wok and copy everything over. Then it will only update the md5sum file in wok when its being about to be cooked.

On Tue, May 10, 2011 at 11:24 AM, Rohit Joshi <rj.rohit@xxxxxxxxx> wrote:

I get that tazwok take longer to do some things but i need to keep wok compabile for tazwok and the wok. Not doing so will just brake everything me and Goklayeh have been working on.

Most of the time taking is cause of tazwok gen-wok-db. The cookorder.txt file takes a while to be made. Also that should only be run like once a week or something. Not every hour.

The default rule should be that if any feature consumes a lot of time, then modify it or DROP it. 

I don't understand why changes made by Pankso will break Gokhlayeh's code. If you need a code for any special purpose, then please follow Slitaz's good coding practices, e.g., keep all your changes coded in a "function" and call that function.