Re: Recent upgrade to wish leads to a problem

2012-03-05 Thread Larry Hall (Cygwin)

On 3/2/2012 1:33 PM, Matt Seitz (matseitz) wrote:

From: Christopher Faylor

For the record, I don't think Yaakov, Corinna, or I are really
interested in spending our time adding some sort of suggestion

mechanism

to setup.exe.  This would have ramifications both for setup.exe, for

the

script which updates setup.ini, and for the genini script.  It would
likely be a fair amount of work.


I thought I had sometimes seen Setup display a dialog box that said
something like The package you have selected requires the following
additional packages, with a check box saying Install these packages
(RECOMMENDED).  What triggers that dialog box to appear?


You get that when you've deselected a package that's a requirement for
another package you selected.

--
Larry

_

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

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



Re: Recent upgrade to wish leads to a problem

2012-03-03 Thread Christopher Faylor
On Fri, Mar 02, 2012 at 10:31:54PM -0600, Robert Miles wrote:
On 3/2/2012 11:43 AM, Christopher Faylor wrote:
 On Fri, Mar 02, 2012 at 01:11:49AM -0600, Robert Miles wrote:
 On 3/1/2012 1:38 PM, Jeremy Bopp wrote:
 On 03/01/2012 01:05 PM, Eric Blake wrote:
 On 03/01/2012 10:53 AM, Matt Seitz (matseitz) wrote:
[snip]
 I suspect not, but I would like to see such recommended and/or 
 suggested packages. Back when I was using Cygwin gcc so I could 
 learn C, m4 was not installed along with it, and I believe that m4 
 should at least be shown as a recommended package to go along with 
 gcc. 
There is no reason to include m4 as a gcc dependency if you are trying
to learn C.  This would not even be a case of a suggestion.  If you
needed m4 for some reason (maybe for autoconf/automake) that really
would be a case where you need to install m4 separately.  You can use
gcc quite happily without having m4 installed.

For the record, I don't think Yaakov, Corinna, or I are really
interested in spending our time adding some sort of suggestion
mechanism to setup.exe.  This would have ramifications both for
setup.exe, for the script which updates setup.ini, and for the genini
script.  It would likely be a fair amount of work.

Given the fact that this has come up repeatedly over the years I am
pretty sure that repeated suggestions that setup.exe should be modified
to do this are likely not going to have any effect.

The examples I used for learning C all used #include, and therefore
required m4 in order to compile.  I believe gcc gave an error message
in this case showing what was missing, and therefore did not slow down
the learning process too much.

Sorry but m4 has nothing to do with #include.  I don't doubt that you
saw some error that mentioned m4 but it undoubtedly came from a package
which used m4.  gcc itself does not use m4 to expand #include.  #include
files are expanded by gcc automatically without the help of an external
program like m4.

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



Re: Recent upgrade to wish leads to a problem

2012-03-02 Thread Andrey Repin
Greetings, Achim Gratz!

 As long as setup.exe doesn't handle recommendations, how about having
 an (empty) package X-application (or whatever better name you come up
 with) that all such applications depend on?  I'm still not sure how to
 get the actual warning to the user, but for starters the postinstall
 script could fail when it doesn't find the Cygwin X server and leave a
 file that cygcheck looks for otherwise.

Nothing stops you from showing a warning/opening a README to the user in the
same postinstall script.
And there's more, and better, ways to check for configured X server, I bet.


--
WBR,
Andrey Repin (anrdae...@freemail.ru) 02.03.2012, 16:22

Sorry for my terrible english...


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



Re: Recent upgrade to wish leads to a problem

2012-03-02 Thread Christopher Faylor
On Fri, Mar 02, 2012 at 01:11:49AM -0600, Robert Miles wrote:
On 3/1/2012 1:38 PM, Jeremy Bopp wrote:
 On 03/01/2012 01:05 PM, Eric Blake wrote:
 On 03/01/2012 10:53 AM, Matt Seitz (matseitz) wrote:
 You don't think that Setup telling the user package xyz requires
 package xinit might at least tip off some users that running xyz now
 requires starting an X server?

 Even if it doesn't reduce the questions, would it at least reduce the
 problem of users now having to run Setup a second time to download and
 install the xinit package?
 While I can't control whether Yaakov adds xinit as a dependency to
 tcl-tk, I personally wouldn't mind adding xinit as a prereq to gitk and
 git-gui, since I maintain that package.  There's still that annoyance
 factor that just depending on xinit won't get X running, but at least it
 will let you download all the bits in one run of setup.exe instead of
 hitting the FAQ and finding out you have to rerun setup.exe.
 I don't think I've ever seen it, but does setup.exe support the notion
 of recommended and/or suggested packages?

 It seems that the main issue is that xinit is not a hard dependency of
 these tools/libraries and thus they should not have a hard dependency on
 xinit defined in the package definition.  However, it might make sense
 to make xinit a recommended dependency that setup.exe could offer to
 install by default or at least notify the user that they may want to
 install.

I suspect not, but I would like to see such recommended and/or
suggested packages.  Back when I was using Cygwin gcc so I could learn
C, m4 was not installed along with it, and I believe that m4 should at
least be shown as a recommended package to go along with gcc.

There is no reason to include m4 as a gcc dependency if you are trying
to learn C.  This would not even be a case of a suggestion.  If you
needed m4 for some reason (maybe for autoconf/automake) that really
would be a case where you need to install m4 separately.  You can use
gcc quite happily without having m4 installed.

For the record, I don't think Yaakov, Corinna, or I are really
interested in spending our time adding some sort of suggestion mechanism
to setup.exe.  This would have ramifications both for setup.exe, for the
script which updates setup.ini, and for the genini script.  It would
likely be a fair amount of work.

Given the fact that this has come up repeatedly over the years I am
pretty sure that repeated suggestions that setup.exe should be modified
to do this are likely not going to have any effect.

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



RE: Recent upgrade to wish leads to a problem

2012-03-02 Thread Matt Seitz (matseitz)
 From: Robert Miles
 
 Does it always require xinit, or only sometimes?  

Sometimes.

 If only sometimes, why
 should it always be installed even for computers where there is no
need
 for it?

It should be always be installed if the benefits of always installing it
outweigh the benefits of not always installing it.



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



RE: Recent upgrade to wish leads to a problem

2012-03-02 Thread Matt Seitz (matseitz)
 From: Christopher Faylor
 
 For the record, I don't think Yaakov, Corinna, or I are really
 interested in spending our time adding some sort of suggestion
mechanism
 to setup.exe.  This would have ramifications both for setup.exe, for
the
 script which updates setup.ini, and for the genini script.  It would
 likely be a fair amount of work.

I thought I had sometimes seen Setup display a dialog box that said
something like The package you have selected requires the following
additional packages, with a check box saying Install these packages
(RECOMMENDED).  What triggers that dialog box to appear?


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



RE: Recent upgrade to wish leads to a problem

2012-03-02 Thread Matt Seitz (matseitz)
 From: Yaakov (Cygwin/X) 
 
 I haven't seen any questions that I haven't already answered (although
 it would help if your mail client was capable of preserving a thread).

Sorry about that.  Hopefully I've cleared up the threading problem now.  Please 
let me know if you are still seeing a problem.

The questions I still have are:

1.  A. If xinit is added as a package dependency, will Setup notify the user 
that xinit will also be installed along with the package?

B.  Will that notification from Setup be enough to give some users a hint that 
they need to start an X server before using the package?

2.  Would adding the xinit dependency spare many users the hassle of having 
to run Setup twice?



RE: Recent upgrade to wish leads to a problem

2012-03-02 Thread Matt Seitz (matseitz)
 From: Achim Gratz [mailto:strom...@nexgo.de]
 
 I do and FTR: I don't want the cygwin Xorg server to be a dependency
to
 all programs that might use X because that would pull in a lot of
 packages that I have no use for on most systems.  

Right, that would be a disadvantage of making xinit a dependency.

My questions are:

1.  How many people would need to install xinit?  How many people
would not need to install xinit?

2.  How much of an inconvenience would installing xinit present to
people who don't actually need it?

3.  Is there some way to make both groups happy?  For example, could we
have a gitk+xinit package to give people a clue that they might want
both (keeping the existing gitk package for people who don't need
xinit)?




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



Re: Recent upgrade to wish leads to a problem

2012-03-02 Thread Robert Miles

On 3/2/2012 11:43 AM, Christopher Faylor wrote:

On Fri, Mar 02, 2012 at 01:11:49AM -0600, Robert Miles wrote:

On 3/1/2012 1:38 PM, Jeremy Bopp wrote:

On 03/01/2012 01:05 PM, Eric Blake wrote:

On 03/01/2012 10:53 AM, Matt Seitz (matseitz) wrote:

[snip]
I suspect not, but I would like to see such recommended and/or 
suggested packages. Back when I was using Cygwin gcc so I could 
learn C, m4 was not installed along with it, and I believe that m4 
should at least be shown as a recommended package to go along with 
gcc. 

There is no reason to include m4 as a gcc dependency if you are trying
to learn C.  This would not even be a case of a suggestion.  If you
needed m4 for some reason (maybe for autoconf/automake) that really
would be a case where you need to install m4 separately.  You can use
gcc quite happily without having m4 installed.

For the record, I don't think Yaakov, Corinna, or I are really
interested in spending our time adding some sort of suggestion mechanism
to setup.exe.  This would have ramifications both for setup.exe, for the
script which updates setup.ini, and for the genini script.  It would
likely be a fair amount of work.

Given the fact that this has come up repeatedly over the years I am
pretty sure that repeated suggestions that setup.exe should be modified
to do this are likely not going to have any effect.

The examples I used for learning C all used #include, and therefore 
required m4 in
order to compile.  I believe gcc gave an error message in this case 
showing what

was missing, and therefore did not slow down the learning process too much.

I remember a second problem with m4 though - some of the #include files 
in the
examples had names starting with blank characters, and m4 did not give 
an error
message that was very useful for identifying this problem and that the 
Cygwin C

libraries were not set up to allow this.

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



Re: Recent upgrade to wish leads to a problem

2012-03-02 Thread Robert Miles

On 3/2/2012 1:02 PM, Matt Seitz (matseitz) wrote:

From: Achim Gratz [mailto:strom...@nexgo.de]

I do and FTR: I don't want the cygwin Xorg server to be a dependency

to

all programs that might use X because that would pull in a lot of
packages that I have no use for on most systems.

Right, that would be a disadvantage of making xinit a dependency.

My questions are:

1.  How many people would need to install xinit?  How many people
would not need to install xinit?

2.  How much of an inconvenience would installing xinit present to
people who don't actually need it?

3.  Is there some way to make both groups happy?  For example, could we
have a gitk+xinit package to give people a clue that they might want
both (keeping the existing gitk package for people who don't need
xinit)?

Would it be adequate to make the installation procedure direct the user 
to read

a text file describing anything that is SOMETIMES needed, and where to find
packages to provide it, and where to find similar text files describing 
any startup

procedures needed?

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Matt Seitz (matseitz)
Yaakov (Cygwin/X) wrote:
 Using X requires user intervention to start an X server first.  No
 amount of automatic dependencies will change this, and therefore I
don't
 expect that the number of questions would change one iota.

You don't think that Setup telling the user package xyz requires
package xinit might at least tip off some users that running xyz now
requires starting an X server?

Even if it doesn't reduce the questions, would it at least reduce the
problem of users now having to run Setup a second time to download and
install the xinit package?

  In the case of packages that have both a console mode and an X11
mode,
  perhaps the package could be split into separate packages, as was
done
  with git, git-gui, and gitk?
 
 Can you provide examples of packages for which this isn't already the
 case?

No, I don't have a specific example in mind.  I was just trying to
anticipate a possible objection of why should we install xinit for a
package that is sometimes used with just a text console?

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Matt Seitz (matseitz)
Christopher Faylor wrote:
 
 In the meantime, if people are piling on to suggest this because they
 think it will cause someone to add xinit as a dependency to something
 please be assured that this will not happen.

OK, what would cause someone to add xinit as a dependency to something?


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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Matt Seitz (matseitz)
Matt Seitz (matseitz) 
Christopher Faylor wrote:
 
 In the meantime, if people are piling on to suggest this because they
 think it will cause someone to add xinit as a dependency to something
 please be assured that this will not happen.

 OK, what would cause someone to add xinit as a dependency to
something?

One reason I keep asking this question is that I don't understand the
rationale against adding xinit.

I love that Setup will automatically install the dependencies I need to
actually use a package.  Setup failing to install a dependency that most
users will want seems like a step backwards.  So I am trying to
understand the rationale behind this position.

One of the frustrations I used to have with Linux was repeated cycles of
Package A needs package B, install B, Package B needs Package C,
install C, Package C needs Package D, etc.  I like that modern package
managers, like Cygwin Setup or Yum make it much easier by letting me
just say I want Package A, and the tool automatically installs all the
dependencies for me.  

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Christopher Faylor
On Thu, Mar 01, 2012 at 09:54:23AM -0800, Matt Seitz (matseitz) wrote:
Christopher Faylor wrote:
 
 In the meantime, if people are piling on to suggest this because they
 think it will cause someone to add xinit as a dependency to something
 please be assured that this will not happen.

OK, what would cause someone to add xinit as a dependency to something?

There is nothing in what I wrote which would suggest that adding xinit as
a dependency was a possibility.

cgf

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Earnie Boyd
On Thu, Mar 1, 2012 at 1:14 PM, Christopher Faylor wrote:
 On Thu, Mar 01, 2012 at 09:54:23AM -0800, Matt Seitz (matseitz) wrote:
Christopher Faylor wrote:

 In the meantime, if people are piling on to suggest this because they
 think it will cause someone to add xinit as a dependency to something
 please be assured that this will not happen.

OK, what would cause someone to add xinit as a dependency to something?

 There is nothing in what I wrote which would suggest that adding xinit as
 a dependency was a possibility.


I agree with Matt.  If TK is requiring the use of the X server then
the X server should be a package dependency of TK.  It wasn't a
dependency before because the X server wasn't needed but now the X
server is needed to use TK.

-- 
Earnie
-- https://sites.google.com/site/earnieboyd

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Christopher Faylor
On Thu, Mar 01, 2012 at 10:07:33AM -0800, Matt Seitz (matseitz) wrote:
Matt Seitz (matseitz) 
Christopher Faylor wrote:
 
 In the meantime, if people are piling on to suggest this because they
 think it will cause someone to add xinit as a dependency to something
 please be assured that this will not happen.

 OK, what would cause someone to add xinit as a dependency to
something?

One reason I keep asking this question is that I don't understand the
rationale against adding xinit.

Yaakov posted the rationale.  You responded to it.  Additional messages
insisting how much you want this are really pretty pointless.

cgf

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Christopher Faylor
On Thu, Mar 01, 2012 at 01:20:49PM -0500, Earnie Boyd wrote:
On Thu, Mar 1, 2012 at 1:14 PM, Christopher Faylor wrote:
 On Thu, Mar 01, 2012 at 09:54:23AM -0800, Matt Seitz (matseitz) wrote:
Christopher Faylor wrote:

 In the meantime, if people are piling on to suggest this because they
 think it will cause someone to add xinit as a dependency to something
 please be assured that this will not happen.

OK, what would cause someone to add xinit as a dependency to something?

 There is nothing in what I wrote which would suggest that adding xinit as
 a dependency was a possibility.

I agree with Matt.  If TK is requiring the use of the X server then
the X server should be a package dependency of TK.  It wasn't a
dependency before because the X server wasn't needed but now the X
server is needed to use TK.

TK does not require an X server running on the same system.  This has
been explained in this very thread.

If you have points to make, don't make them with a aolme too/aol
unless you're planning on addressing the issues that Yaakov raised.
Otherwise you're just wasting bandwidth.

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Earnie Boyd
On Thu, Mar 1, 2012 at 1:21 PM, Christopher Faylor wrote:
 On Thu, Mar 01, 2012 at 10:07:33AM -0800, Matt Seitz (matseitz) wrote:
Matt Seitz (matseitz)
Christopher Faylor wrote:

 In the meantime, if people are piling on to suggest this because they
 think it will cause someone to add xinit as a dependency to something
 please be assured that this will not happen.

 OK, what would cause someone to add xinit as a dependency to
something?

One reason I keep asking this question is that I don't understand the
rationale against adding xinit.

 Yaakov posted the rationale.  You responded to it.  Additional messages
 insisting how much you want this are really pretty pointless.

quote
Here's my advice: it would be a better use of your time to install xinit
and accustom yourself to the wonders of X rather than hopelessly trying
to find a way to continue living in the past.
/quote

quote
Using X requires user intervention to start an X server first.  No
amount of automatic dependencies will change this, and therefore I don't
expect that the number of questions would change one iota.
/quote

But if TK needs an X server started in must be a installed and is a
dependency of TK to have X server.  Yes, one has to start it but why
not require the dependency so that the X server software exists to use
already?  I agree with Matt, I'm not getting the angst for adding the
X server as a dependency for TK which requires it.  Having to start it
is a different matter.

-- 
Earnie
-- https://sites.google.com/site/earnieboyd

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Earnie Boyd
On Thu, Mar 1, 2012 at 1:28 PM, Christopher Faylor wrote:

 TK does not require an X server running on the same system.  This has
 been explained in this very thread.

 If you have points to make, don't make them with a aolme too/aol
 unless you're planning on addressing the issues that Yaakov raised.
 Otherwise you're just wasting bandwidth.


And what Windows user who casually installs Cygwin has access to an X
server?  I sure don't.  But I get where you're coming from, it doesn't
require the Cygwin/X server it just requires DISPLAY to be set to one
that exists somewhere.  Still not a good scenario for the unsuspecting
idiot.

-- 
Earnie
-- https://sites.google.com/site/earnieboyd

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Matt Seitz (matseitz)
Christopher Faylor wrote:
 
 Yaakov posted the rationale.  You responded to it.  Additional
messages
 insisting how much you want this are really pretty pointless.

Sorry, I don't mean to be a pest.  I'll wait and see if Yaakov replies
to my latest questions.





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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Eric Blake
[your mailer doesn't set In-Reply-To correctly, which means you are
starting a bunch of new threads instead of replying in-thread]

On 03/01/2012 10:53 AM, Matt Seitz (matseitz) wrote:
 Yaakov (Cygwin/X) wrote:
 Using X requires user intervention to start an X server first.  No
 amount of automatic dependencies will change this, and therefore I
 don't
 expect that the number of questions would change one iota.
 
 You don't think that Setup telling the user package xyz requires
 package xinit might at least tip off some users that running xyz now
 requires starting an X server?
 
 Even if it doesn't reduce the questions, would it at least reduce the
 problem of users now having to run Setup a second time to download and
 install the xinit package?

While I can't control whether Yaakov adds xinit as a dependency to
tcl-tk, I personally wouldn't mind adding xinit as a prereq to gitk and
git-gui, since I maintain that package.  There's still that annoyance
factor that just depending on xinit won't get X running, but at least it
will let you download all the bits in one run of setup.exe instead of
hitting the FAQ and finding out you have to rerun setup.exe.

-- 
Eric Blake   ebl...@redhat.com+1-919-301-3266
Libvirt virtualization library http://libvirt.org



signature.asc
Description: OpenPGP digital signature


Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Jeremy Bopp
On 03/01/2012 01:05 PM, Eric Blake wrote:
 On 03/01/2012 10:53 AM, Matt Seitz (matseitz) wrote:
 You don't think that Setup telling the user package xyz requires
 package xinit might at least tip off some users that running xyz now
 requires starting an X server?

 Even if it doesn't reduce the questions, would it at least reduce the
 problem of users now having to run Setup a second time to download and
 install the xinit package?
 
 While I can't control whether Yaakov adds xinit as a dependency to
 tcl-tk, I personally wouldn't mind adding xinit as a prereq to gitk and
 git-gui, since I maintain that package.  There's still that annoyance
 factor that just depending on xinit won't get X running, but at least it
 will let you download all the bits in one run of setup.exe instead of
 hitting the FAQ and finding out you have to rerun setup.exe.

I don't think I've ever seen it, but does setup.exe support the notion
of recommended and/or suggested packages?

It seems that the main issue is that xinit is not a hard dependency of
these tools/libraries and thus they should not have a hard dependency on
xinit defined in the package definition.  However, it might make sense
to make xinit a recommended dependency that setup.exe could offer to
install by default or at least notify the user that they may want to
install.

Less sophisticated users of setup.exe could select to always install all
recommended packages by default (this may be the default setting) while
more advanced users may disable that option in order to be more choosy.
 While this wouldn't help the problem of users needing to actually start
the X server before using the dependent tools and libraries, it would at
least give them a chance to get what they need in a single pass of
running setup.exe rather than needing to run it again and try to select
the correct additional packages.

-Jeremy

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread marco atzeri

On 3/1/2012 8:05 PM, Eric Blake wrote:

[your mailer doesn't set In-Reply-To correctly, which means you are
starting a bunch of new threads instead of replying in-thread]

On 03/01/2012 10:53 AM, Matt Seitz (matseitz) wrote:

Yaakov (Cygwin/X) wrote:

Using X requires user intervention to start an X server first.  No
amount of automatic dependencies will change this, and therefore I

don't

expect that the number of questions would change one iota.


You don't think that Setup telling the user package xyz requires
package xinit might at least tip off some users that running xyz now
requires starting an X server?

Even if it doesn't reduce the questions, would it at least reduce the
problem of users now having to run Setup a second time to download and
install the xinit package?


While I can't control whether Yaakov adds xinit as a dependency to
tcl-tk, I personally wouldn't mind adding xinit as a prereq to gitk and
git-gui, since I maintain that package.  There's still that annoyance
factor that just depending on xinit won't get X running, but at least it
will let you download all the bits in one run of setup.exe instead of
hitting the FAQ and finding out you have to rerun setup.exe.



Eric,
I understand the logic of Yaakov, but I expect that 99.9% of user using
git-gui will use the x-server on cygwin.

At least the mantainer of  clisp-clx had the same thought
about its package:

sdesc: The new-clx module linkset for clisp
ldesc: This is the linkset full+new-clx for X.Org.
category: Interpreters
requires: clisp xorg-server libXpm4 libXt6 cygwin

So your proposal for gittk and git-gui fits in the same view

Regards
Marco


Regards
Marco

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Achim Gratz
Earnie Boyd ear...@users.sourceforge.net writes:
 And what Windows user who casually installs Cygwin has access to an X
 server?

I do and FTR: I don't want the cygwin Xorg server to be a dependency to
all programs that might use X because that would pull in a lot of
packages that I have no use for on most systems.  You might recommend
companion packages that are typically needed to be combined with a
specific selected package when installed in an isolated environment, but
please not a dependency.  It is perfectly OK to install X applications
onto a machine that doesn't have X installed.  You don't need a server
there, nor do you need an xinit start script or anything else besides
the actual application and a handful of DLL.


Regards,
Achim.
-- 
+[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]+

SD adaptations for KORG EX-800 and Poly-800MkII V0.9:
http://Synth.Stromeko.net/Downloads.html#KorgSDada


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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Yaakov (Cygwin/X)
On Thu, 2012-03-01 at 10:42 -0800, Matt Seitz (matseitz) wrote:
 Christopher Faylor wrote:
  
  Yaakov posted the rationale.  You responded to it.  Additional
 messages
  insisting how much you want this are really pretty pointless.
 
 Sorry, I don't mean to be a pest.  I'll wait and see if Yaakov replies
 to my latest questions.

I haven't seen any questions that I haven't already answered (although
it would help if your mail client was capable of preserving a thread).


Yaakov



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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Yaakov (Cygwin/X)
On Thu, 2012-03-01 at 21:40 +0100, Achim Gratz wrote:
 Earnie Boyd ear...@users.sourceforge.net writes:
  And what Windows user who casually installs Cygwin has access to an X
  server?
 
 I do and FTR: I don't want the cygwin Xorg server to be a dependency to
 all programs that might use X because that would pull in a lot of
 packages that I have no use for on most systems.  You might recommend
 companion packages that are typically needed to be combined with a
 specific selected package when installed in an isolated environment, but
 please not a dependency.  It is perfectly OK to install X applications
 onto a machine that doesn't have X installed.  You don't need a server
 there, nor do you need an xinit start script or anything else besides
 the actual application and a handful of DLL.

Thank you for reinforcing my point.  While we certainly promote the
Cygwin/X server, forcing xorg-server/xinit as a dependency not only
won't prevent questions (since the user still needs to *start* the X
server), but those who use other X servers will have this very
complaint.

Now please excuse me while I get back to orchestrating the next major
transition for the distro.


Yaakov



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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread marco atzeri
On Fri, Mar 2, 2012 at 2:15 AM, Yaakov (Cygwin/X)  wrote:

 Now please excuse me while I get back to orchestrating the next major
 transition for the distro.


 Yaakov


Thanks very much for this.

Marco

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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Achim Gratz
Yaakov (Cygwin/X) yselkow...@users.sourceforge.net writes:
 Thank you for reinforcing my point.

No, thank you for all your work, because without that we couldn't have
this discussion.

 While we certainly promote the Cygwin/X server, forcing
 xorg-server/xinit as a dependency not only won't prevent questions
 (since the user still needs to *start* the X server), but those who
 use other X servers will have this very complaint.

Note: I'm not opposed to give a stronger hint to folks who might not
know what an X server is and why they need one.  If they don't follow up
on that hint it becomes their problem, but you can't reasonably expect
someone to look for an answer to a question they don't even have.

As long as setup.exe doesn't handle recommendations, how about having
an (empty) package X-application (or whatever better name you come up
with) that all such applications depend on?  I'm still not sure how to
get the actual warning to the user, but for starters the postinstall
script could fail when it doesn't find the Cygwin X server and leave a
file that cygcheck looks for otherwise.  If that file is missing I'll
get an INCOMPLETE package in cygcheck.  If I have an install using an
external X server, I can create that file and have clean cygcheck output
(and getting the postinstall marked done).

 Now please excuse me while I get back to orchestrating the next major
 transition for the distro.

Thanks again.


Regards,
Achim.
-- 
+[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]+

Factory and User Sound Singles for Waldorf rackAttack:
http://Synth.Stromeko.net/Downloads.html#WaldorfSounds


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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Robert Miles

On 3/1/2012 12:07 PM, Matt Seitz (matseitz) wrote:

Matt Seitz (matseitz)

Christopher Faylor wrote:


In the meantime, if people are piling on to suggest this because they
think it will cause someone to add xinit as a dependency to something
please be assured that this will not happen.

OK, what would cause someone to add xinit as a dependency to

something?

One reason I keep asking this question is that I don't understand the
rationale against adding xinit.

I love that Setup will automatically install the dependencies I need to
actually use a package.  Setup failing to install a dependency that most
users will want seems like a step backwards.  So I am trying to
understand the rationale behind this position.

One of the frustrations I used to have with Linux was repeated cycles of
Package A needs package B, install B, Package B needs Package C,
install C, Package C needs Package D, etc.  I like that modern package
managers, like Cygwin Setup or Yum make it much easier by letting me
just say I want Package A, and the tool automatically installs all the
dependencies for me.

Some questions for which answers might help you understand:

Does it always require xinit, or only sometimes?  If only sometimes, why 
should it always be installed even for computers where there is no need 
for it?


Note - I don't have the answers.


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



Re: Recent upgrade to wish leads to a problem

2012-03-01 Thread Robert Miles

On 3/1/2012 1:38 PM, Jeremy Bopp wrote:

On 03/01/2012 01:05 PM, Eric Blake wrote:

On 03/01/2012 10:53 AM, Matt Seitz (matseitz) wrote:

You don't think that Setup telling the user package xyz requires
package xinit might at least tip off some users that running xyz now
requires starting an X server?

Even if it doesn't reduce the questions, would it at least reduce the
problem of users now having to run Setup a second time to download and
install the xinit package?

While I can't control whether Yaakov adds xinit as a dependency to
tcl-tk, I personally wouldn't mind adding xinit as a prereq to gitk and
git-gui, since I maintain that package.  There's still that annoyance
factor that just depending on xinit won't get X running, but at least it
will let you download all the bits in one run of setup.exe instead of
hitting the FAQ and finding out you have to rerun setup.exe.

I don't think I've ever seen it, but does setup.exe support the notion
of recommended and/or suggested packages?

It seems that the main issue is that xinit is not a hard dependency of
these tools/libraries and thus they should not have a hard dependency on
xinit defined in the package definition.  However, it might make sense
to make xinit a recommended dependency that setup.exe could offer to
install by default or at least notify the user that they may want to
install.

I suspect not, but I would like to see such recommended and/or suggested 
packages.
Back when I was using Cygwin gcc so I could learn C, m4 was not 
installed along
with it, and I believe that m4 should at least be shown as a recommended 
package

to go along with gcc.


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



Re: Recent upgrade to wish leads to a problem

2012-02-29 Thread Fergus

 Previously bin/wish was a link to wish84.exe (from memory).
 Recently it was upgraded to wish 8.5.exe. Now, unless X is
 also running, wish fails ... I'm not quite certain which
 recently upgraded package led to this: tcl-tk or tcltk ..?

 The tcltk libraries now require a running X server in order to
 display graphics. This is recent and more importantly intentional
 according to what I've read on this list ...

OK, thanks. I'm really miserable about this advance which has messed 
badly with my preferred MO (amongst other things, not using X). I tried 
to revert but got into a tangle (and in fact ended up with no wish* 
under /bin, at all). I have recovered a non-updated Cygwin from about a 
month ago (i.e. wish - wish84.exe) which of course offers many 
opportunities for updating packages using setup.exe.


It is not clear to me which update opportunity I should decline in order 
to keep wish - wish84.exe and not move to wish - wish8.5.exe. 
Searching for this executable under Search packages yielded


Search Results
Found 1 matches for wish8.5.exe
tcl-tk/tcl-tk-8.5.11-1Tcl X11 toolkit

so I de-selected this. However, watching the update process unfold, it 
is clear that *something else* uninstalled tcl-tk (I saw the phrase 
Uninstalling tcl-tk). Was it expect? python? both of which were 
selected in the update process. So as before, I have ended up with no 
wish* under /bin, at all. (Possibly some dependency is missing in 
setup.ini? Also there appears to be nothing under [prev] under tcl-tk :o( )


I will try to recover the working version from 1 month ago and start again.

Q1 It would be good to know what recent update, separate from tcl-tk 
explicitly listed above, has led to the un-installation of the existing 
tcl-tk. Any ideas?


Q2 In some other contexts Cygwin provides nox versions additionally to 
versions requiring a running X server. Is there any chance that 
tcl-tk-8.4 could be recovered and offered as a nox version?


Thank you.
Fergus

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



Re: Recent upgrade to wish leads to a problem

2012-02-29 Thread Yaakov (Cygwin/X)
On Wed, 2012-02-29 at 08:24 +, Fergus wrote:
 OK, thanks. I'm really miserable about this advance which has messed 
 badly with my preferred MO (amongst other things, not using X).

The old 8.4 win32 tcl/tk was unmaintained and broken in many ways, as
discussed at length on these lists, and now it is gone and isn't coming
back.  There is no way to keep the 8.4 version because it would conflict
with the 8.5 version now used by everything in the distro, and you can't
mix-and-match these.

Here's my advice: it would be a better use of your time to install xinit
and accustom yourself to the wonders of X rather than hopelessly trying
to find a way to continue living in the past.


Yaakov



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



Re: Recent upgrade to wish leads to a problem

2012-02-29 Thread Henry S. Thompson
Fergus writes:

 Q2 In some other contexts Cygwin provides nox versions additionally
 to versions requiring a running X server. Is there any chance that
 tcl-tk-8.4 could be recovered and offered as a nox version?

+1

Please!

ht
-- 
   Henry S. Thompson, School of Informatics, University of Edinburgh
  10 Crichton Street, Edinburgh EH8 9AB, SCOTLAND -- (44) 131 650-4440
Fax: (44) 131 650-4587, e-mail: h...@inf.ed.ac.uk
   URL: http://www.ltg.ed.ac.uk/~ht/
 [mail from me _always_ has a .sig like this -- mail without it is forged spam]

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



Re: Recent upgrade to wish leads to a problem

2012-02-29 Thread Corinna Vinschen
On Feb 29 09:41, Henry S. Thompson wrote:
 Fergus writes:
 
  Q2 In some other contexts Cygwin provides nox versions additionally
  to versions requiring a running X server. Is there any chance that
  tcl-tk-8.4 could be recovered and offered as a nox version?
 
 +1
 
 Please!

If you manage to do it so that it does in NO way collide with the
new release, then there's nothing speaking against it.  Just none
of the existing maintainers is interested in maintaining it.

http://cygwin.com/acronyms/#SHTDI
http://cygwin.com/setup.html


Corinna

-- 
Corinna Vinschen  Please, send mails regarding Cygwin to
Cygwin Project Co-Leader  cygwin AT cygwin DOT com
Red Hat

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



Re: Recent upgrade to wish leads to a problem

2012-02-29 Thread Matt Seitz (matseitz)
Christopher Faylor wrote:
 
 The only thing that apparently needs addressing is that you read the
 list and comprehend what's going on.  I wish we could address that by
 making more people do that.  :-)

Would it help to add xinit to the requirements for tcl-tk and other
packages that now require an X11 server?

I know that there are some use cases where xinit isn't actually
required.  But would the benefit (fewer problem reports from new users)
be worth the cost (installing xinit for some users who don't actually
require it)?

In the case of packages that have both a console mode and an X11 mode,
perhaps the package could be split into separate packages, as was done
with git, git-gui, and gitk?

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



Re: Recent upgrade to wish leads to a problem

2012-02-29 Thread Yaakov (Cygwin/X)
On Wed, 2012-02-29 at 13:43 -0800, Matt Seitz (matseitz) wrote:
 Would it help to add xinit to the requirements for tcl-tk and other
 packages that now require an X11 server?
 
 I know that there are some use cases where xinit isn't actually
 required.  But would the benefit (fewer problem reports from new users)
 be worth the cost (installing xinit for some users who don't actually
 require it)?

Asking the same question in a dozen different ways won't change the
answer.

Using X requires user intervention to start an X server first.  No
amount of automatic dependencies will change this, and therefore I don't
expect that the number of questions would change one iota.

 In the case of packages that have both a console mode and an X11 mode,
 perhaps the package could be split into separate packages, as was done
 with git, git-gui, and gitk?

Can you provide examples of packages for which this isn't already the
case?


Yaakov



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



Re: Recent upgrade to wish leads to a problem

2012-02-29 Thread Christopher Faylor
On Wed, Feb 29, 2012 at 06:57:27PM -0600, Yaakov (Cygwin/X) wrote:
On Wed, 2012-02-29 at 13:43 -0800, Matt Seitz (matseitz) wrote:
 Would it help to add xinit to the requirements for tcl-tk and other
 packages that now require an X11 server?
 
 I know that there are some use cases where xinit isn't actually
 required.  But would the benefit (fewer problem reports from new users)
 be worth the cost (installing xinit for some users who don't actually
 require it)?

Asking the same question in a dozen different ways won't change the
answer.

Using X requires user intervention to start an X server first.  No
amount of automatic dependencies will change this, and therefore I don't
expect that the number of questions would change one iota.

I agree 100% but this now qualifies as a FAQ so maybe we should add an
entry about tcl/tk.

In the meantime, if people are piling on to suggest this because they
think it will cause someone to add xinit as a dependency to something
please be assured that this will not happen.

cgf

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



Re: Recent upgrade to wish leads to a problem

2012-02-24 Thread Jeremy Bopp
On 02/24/2012 08:25 AM, Fergus wrote:
 Previously bin/wish was a link to wish84.exe (from memory). Recently it
 was upgraded to wish 8.5.exe.
 Now, unless X is also running, wish fails with
 
 $ wish
 % Application initialization failed: no display name and no $DISPLAY
 environment variable
 
 I'm not quite certain which recently upgraded package led to this:
 tcl-tk or tcltk or ...
 Have I identified a real glitch and if so can this problem be addressed?

The tcltk libraries now require a running X server in order to display
graphics.  This is recent and more importantly intentional according to
what I've read on this list, so there is nothing to address. :-)

-Jeremy

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



Re: Recent upgrade to wish leads to a problem

2012-02-24 Thread Christopher Faylor
On Fri, Feb 24, 2012 at 09:09:02AM -0600, Jeremy Bopp wrote:
On 02/24/2012 08:25 AM, Fergus wrote:
 Previously bin/wish was a link to wish84.exe (from memory). Recently it
 was upgraded to wish 8.5.exe.
 Now, unless X is also running, wish fails with
 
 $ wish
 % Application initialization failed: no display name and no $DISPLAY
 environment variable
 
 I'm not quite certain which recently upgraded package led to this:
 tcl-tk or tcltk or ...
 Have I identified a real glitch and if so can this problem be addressed?

The tcltk libraries now require a running X server in order to display
graphics.  This is recent and more importantly intentional according to
what I've read on this list, so there is nothing to address. :-)

The only thing that apparently needs addressing is that you read the
list and comprehend what's going on.  I wish we could address that by
making more people do that.  :-)

cgf

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