Re: Re: why bloated packages are bad (acaseforsplittingup large packages)
| From: | Greg Beaver | Date: | Mon, 15 Aug 2005 19:58:47 +0000 |
| Subject: | Re: Re: why bloated packages are bad (acaseforsplittingup large packages) | ||
| References: | 1 2 3 4 5 6 7 8 9 | Groups: | php.pear.core |
| Request: | Send a blank email to pear-core+get-3571@lists.php.net to get a copy of this message | ||
Lukas Smith wrote:
> No, the other way around. All the recent PEAR releases had to do with
> bugs in the installer code and not in PEAR.php, PEAR_Error .. not even
> for the development steps of ErrorStack. Then again I dont know if
> updates where delayed due to waiting for the next PEAR release.
> However in the few versions I clicked through the number of installer
> related fixed clearly outweight the number of fixes to other
> infrastructure code.
When I add small things like a method "staticPop()" to PEAR_ErrorStack,
I don't feel justified in releasing the entire PEAR package, so yes, I
have delayed releasing this package for about 4-5 months in the stable
version of PEAR.
> So I am wondering if we will really "safe" all that many PEAR releases
> by undbundling ErrorStack which should probably get less likely to be
> changed now that its approach stable state.
That is a good question. My reasoning is that PEAR_ErrorStack really
does not belong in the PEAR installer package - this is debatable, but I
felt this way from the beginning. I put the package in PEAR because
Pierre insisted it should be. I never understood the reasoning, and all
of my fears regarding the package have come true. These include
inability to do quick point releases, and confusion as to its use and
purpose (why is this in the installer package? It isn't PEAR_Error2)
For instance, the PEAR_Delegator class is in its own package (a wise
decision), even though it supersedes PEAR_Autoloader.
After considering the arguments, I see no convincing case for any
negative consequences with the unbundling, so I will continue as
originally planned.
Greg