I created issue 21617 <https://pharo.fogbugz.com/f/cases/21617/Better-implementation-of-package-comments>
2018-03-16 16:54 GMT+01:00 Stephane Ducasse <[email protected]>: > Hi denis > > Both solutions are ok for me. > > STef > > On Fri, Mar 16, 2018 at 2:38 PM, Denis Kudriashov <[email protected]> > wrote: > > And at the end I found that #packageComment: is not really work because > it > > expects already formatted string suitable for source code insertion. > > So you can't just call: > > > > aPackage packageComment: 'my comment'. > > > > Instead all users (only Nautilus exists) should write > > > > aPackage packageComment: 'my comment' printString > > > > I will create pull request with all this changes > > > > 2018-03-14 16:51 GMT+01:00 Denis Kudriashov <[email protected]>: > >> > >> And I found that manifest classes are already commented as package > >> comment. > >> > >> 2018-03-14 16:40 GMT+01:00 Denis Kudriashov <[email protected]>: > >>> > >>> Hi. > >>> > >>> I look at hierarchy of PackageManifest and I was wondering that all > these > >>> classes are commented but most of packages which they describe are not. > >>> > >>> Also interesting that most of manifest comments are same. > >>> > >>> Then I found that package comment is managed as special class side > method > >>> #description which is generated when you call "aPackage packageComment: > >>> 'test'". > >>> > >>> Now the idea: will not it be better to manage package comment as > manifest > >>> class comment? > >>> > >>> Manifest is already responsible to describe the package. So it looks > >>> logical to have "manifest comment = package comment". > >>> Also it will simplify code and remove duplication of texts. > >>> In addition it will add extra place indicating that package needs to be > >>> commented because uncommented classes are marked in the browser. > >>> > >>> What you think? > >>> > >> > > > >
