[as-devel] 1.8 release woes
Sasha_Vasko@osca.state.mo.us
Tue, 1 Feb 2000 14:12:24 -0600
> There has been a great deal of disagreement on the 'upcoming'
>release of 1.8. In my opinion, the latest devel needs to have a
>no-brainer method of going from 1.6 to 1.8. The documentation also needs
>to be 100% _complete_ and _accurate_ - FAQ, man pages, READMEs (Solaris,
>HPUX, etc). The current look/feel/etc files need to be thuroughly
>inspected to verify that they are indeed current with the latest
>'expected' contents (AfterStep:function specification requires "name" in
>[Nop]. is one that is continuously output with default files). The
>test.steprc needs adjusting to work; feels need _every_ cursor definition
>in them. This might not be an inclusive list of things, but it's on my
>TODO.
> As for ascp, it is my opinion that releasing 1.8 w/o ascp is a
>BadThing. Everyone seems to disagreee with mw on this point, but it's my
>feeling. We need, as a minumun if ascp won't be released as 1.8, some
>method to transfer 1.6.x config files to the new format. A lengthy text
>file telling the user what to do is _not_ the answer; there needs to be a
>script - be it perl, shell, java; I really don't care, but a methodology
>needs to exist.
> Now this is in the forefront and open for debate/discussion/law
>suits/etc.
Problem is that it is by far not complete list. Among bugs/features that are
high priority also :
1) Complete AfterStep ignorance of the Gravity requested by applications
Central/South/East gravities are not supported. IMHO it is an extremely
shamefull
thing as it the basic functionality that MUST be supported by WM.
Examples: run aterm -geometry -100-100 , and try changing font size - it does
not keep
-100-100 position.
2) Frames implementation sucks - it's not flexible, does not allow for frames
on menus and does not allow semitransparent frames. Bug of Frames being screwed
on look change is only one implication of that.
3) AfterStep is not yet converted to use parsing code, which must be
accomplished before
ascp will be ready. Even thou Andrew has done quite a piece of work, there
are still all the difficult files to be done, and there are quite a few problems
to be sorted out
on this way.
4) the way virtual desktops are handled is not ICCCM compliant and it is a
miracle
that we did not see many problems with that before, party I think it is due to
AfterStep
being treated as not a mature software by big users.
This is only alist of high importance issues, and does not include many smaller
ones.
Now All this amounts to a rewrite of 40% of AfterStep core code, this is not an
easy fix
and will take month and month of hard work, and it requires even more user
testing.
All this makes it unrealistic to release stable AfterStep that is fixing all the
issues soon.
Accordingly there was a decision back in November to go the way of compromise,
and take 1.7.142+10patches add quick and dirty theme support, clean most nasty
bugs
and release it as 1.8 ASAP, by the Christmas. And at the mean time continue work
on
implementing all this listed issues in 1.9 tree.
And so we did. and IMHO we should finish what we started, nd release 1.8 - maybe
not
perfect, but good enough to be called a stable release, and do it ASAP.
If no, then all the development that was done for fixing those listed issues
(and that is a lot)
must be backported into 1.8 and we should forget about stable release for
another year,
which is a realistic time frame ( says developer :).
Please understand that I want AfterStep to be perfect too, but sorry, its not
possible
to be done soon.
>+--------------------------------------------------------------------+
>| d a v i d @ m i h m reality.sys corrupt! |
Sasha.