This is the mail archive of the cygwin-xfree mailing list for the Cygwin XFree86 project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: run.exe will not work with upgrade from 1.14.4 to 1.16.3


On 5-1-2015 05:04, Larry Hall (Cygwin-X) wrote:
> So I'm guessing with your statement above that English isn't your primary
> language.
>  [..]
> 1. I am not the maintainer of the xinit package. [..]

English is indeed not my first language, but that is no excuse for not
carefully reading your replies and not verifying assumptions about you
being a maintainer. I did not read your replies carefully enough and did
not check my assumptions. Please accept my apologies.

> [..]
> 2. Yaakov is a very capable and prolific contributor to the Cygwin
project
>    and has been for many years. [..]

I do appreciate your (yours, Yaakov's, and others) efforts very much. I
tried to express that in the last paragraph of my original mail [1].
Please tell me if this intent did not come across or was drowned
somehow, so I can phrase it better next time.

> [..]
> You'll notice that sometimes Yaakov is answering
>  the question raised and other times others are doing it.  That's
>  standard operating procedure. [..]

I did read all the previous questions and answers. I interpreted the
focus on having users change their configuration rather than changing
the xinit package as a denial of the problem. However, now I see that
none of the replies were by the package maintainer. I now realize that
you were doing the best you could in helping me and others.

> [..]
>  When I mentioned above that you or others can help out by pointing out
>  where the solutions proposed fall short [..]

As far as I can tell, from the available information, users which meet
any of the following criteria will run into trouble:

- Custom .startxwinrc or .xinitrc
- Using untrusted X11 forwards over SSH (e.g. ssh -X or PuTTY)

My assumption is that this covers the vast majority of xinit users.
Including these which previously complained about the non-standard way
of handling configuration.

As a software engineer I strongly believe in the principle of least
astonishment [2]. At least for the vast majority of users. In this case,
in my opinion at least, that would mean that changing the behaviour of
startxwin should be done in such a way as to provide a seamless way of
users to upgrade. Preferably by maintaining compatibility with existing
configurations, or by automatic conversion, or, if necessary through a
well documented manual transition process.

What I am trying to say is that I don't object to your solutions, but I
would really like this to be solved in a way that provides a create user
experience. For me that would mean that the first step would be to
retract the update (revert back to 1.3.2-1) as to prevent more users
from running into problems. I do realize that that would mean forcing
people who have already converted to go back. But I assume this is a
minority of users. Then I would propose to evaluate what could be done
to provide a smooth transition, possibly over a longer time, popping up
increasingly annoying warnings about the configuration, for example.

I would like to help with this. I think I can assist in figuring out
what kind of configurations are out there, as well as in testing, and in
writing documentation. I could even code the solution, but that would
probably be more efficient of people with more experience in bash
scripting would do that.

> [..]
> So I withdraw my previous request that you not post
> your policy question to the Cygwin main list (since I agree it is a
> general issue) but instead request that you only discuss the policy and
> not overlap with the specifics covered in the xinit package threads [..]

Thank you, I would like to discussion on the general list to be broader.
Because, for me, this issue was highlighted by xinit, but is by no means
related to it. I, and I assume many others, rely on Cygwin and also
assume, possibly incorrectly, that it will continue to function as
expected, after applying (security) updates. An incompatible change in
any other package would also be problematic for me. The reason I
mentioned xinit explicitly is to provide the reader with a background,
and as soft of a 'full disclosure' of my involvement.

I will also make it very explicit on the common list that the post is
not specific to xinit, but that is merely serves as an example.

> A: Yes.
> > Q: Are you sure?
> >> A: Because it reverses the logical flow of conversation.
> >>> Q: Why is top posting annoying in email?

I really like your signature, do you mind if I borrow/steal it?

Laurens

[1] https://cygwin.com/ml/cygwin-xfree/2014-12/msg00060.html
[2] http://en.wikipedia.org/wiki/Principle_of_least_astonishment


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://x.cygwin.com/docs/
FAQ:                   http://x.cygwin.com/docs/faq/


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]