[PATCH] Clear filename line when uninstalling

[Igor Pechtchanski]

This one-line patch clears the line containing the filename when replacing
packages.  This avoids the last filename of the previous package being
displayed while the next is uninstalled.
Igor
==
ChangeLog:
2003-03-13  Igor Pechtchanski <[EMAIL PROTECTED]>

* install.cc (replace_one): Clear Text3.

-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune
Index: install.cc
===
RCS file: /cvs/cygwin-apps/setup/install.cc,v
retrieving revision 2.57
diff -u -p -r2.57 install.cc
--- install.cc  12 Mar 2003 22:15:25 -  2.57
+++ install.cc  14 Mar 2003 03:41:24 -
@@ -150,6 +150,7 @@ replace_one (packagemeta & pkg)
   int errors = 0;
   Progress.SetText1 ("Replacing...");
   Progress.SetText2 (pkg.name.cstr_oneuse());
+  Progress.SetText3 ("");
   log (LOG_PLAIN, String( "Replacing ")  + pkg.name);
   pkg.uninstall ();
 

Re: Setup cvs-HEAD hangs intermittently when downloading

[Igor Pechtchanski]

On Thu, 13 Mar 2003, Igor Pechtchanski wrote:

> On 14 Mar 2003, Robert Collins wrote:
>
> > If it's a hang and you're breaking in, try stepping through a bit -
> > there may be infinite recursion or some such.
> >
> > If it's segfaulting the problem should be easier to debug.
> > Rob
>
> Unfortunately, it's intermittent and not reliably reproducible.  I'm not
> at all sure what causes it.  I'll try it next time I get it, though.
>
> Frankly, I doubt it's infinite recursion - CPU usage is 0.  I'm more
> inclined to think there's a deadlock somewhere, or it's waiting for
> something that doesn't notify...  And it's not really hanging, either, in
> that it gets Windows events and reacts to buttons correctly...
> Igor

On reflection, it looks like the backtrace I got was for the wrong
thread.  I'll post backtraces for all threads next time I can reproduce
this.
Igor
-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune

Re: Setup cvs-HEAD hangs intermittently when downloading

[Igor Pechtchanski]

On 14 Mar 2003, Robert Collins wrote:

> If it's a hang and you're breaking in, try stepping through a bit -
> there may be infinite recursion or some such.
>
> If it's segfaulting the problem should be easier to debug.
> Rob

Unfortunately, it's intermittent and not reliably reproducible.  I'm not
at all sure what causes it.  I'll try it next time I get it, though.

Frankly, I doubt it's infinite recursion - CPU usage is 0.  I'm more
inclined to think there's a deadlock somewhere, or it's waiting for
something that doesn't notify...  And it's not really hanging, either, in
that it gets Windows events and reacts to buttons correctly...
Igor
-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune

Re: Setup cvs-HEAD hangs intermittently when downloading

[Robert Collins]

If it's a hang and you're breaking in, try stepping through a bit -
there may be infinite recursion or some such.

If it's segfaulting the problem should be easier to debug.

Rob


-- 
GPG key available at: .


signature.asc
Description: This is a digitally signed message part

Re: [PATCH] Add logging of move-on-reboot scheduling

[Robert Collins]

On Fri, 2003-03-14 at 11:22, Max Bowsher wrote:
> Robert Collins wrote:
> > Thanks Igor.
> > 
> > Max, do you have time to apply this? If so, please do, otherwise
> > it'll have to sit in the pending application queue for a bit. 
> 
> Done. Shall I refresh the snapshot?

Please.

> Should I do that without specifically asking in future?

Yes. The policy I use for updating a snapshot is : will the change
affect users significantly. 
i.e. the duplicate All category didn't, so I didn't refresh the
snapshot.
This change will (it will warn win9x users where it didn't before), so
the snapshot should be refreshed.

> What policy should I use for removing old snapshots?

I like to keep a few versions around. I trim versions that are know to
cause problems, or everything prior to the point that we issue an RC.
(From the RC through to the release, I keep everything, subject to the
'known to cause problems' heuristic).
At release, well I'll do that for now :}.

Rob

-- 
GPG key available at: .


signature.asc
Description: This is a digitally signed message part

Re: [PATCH] Add logging of move-on-reboot scheduling

[Max Bowsher]

Robert Collins wrote:
> Thanks Igor.
> 
> Max, do you have time to apply this? If so, please do, otherwise
> it'll have to sit in the pending application queue for a bit. 

Done. Shall I refresh the snapshot?

Should I do that without specifically asking in future?

What policy should I use for removing old snapshots?

Max.

Re: [PATCH] Add logging of move-on-reboot scheduling

[Robert Collins]

Thanks Igor.

Max, do you have time to apply this? If so, please do, otherwise it'll have to sit in 
the pending application queue for a bit.

Rob


signature.asc
Description: This is a digitally signed message part

Re: Setup cvs-HEAD hangs intermittently when downloading

[Igor Pechtchanski]

On Thu, 13 Mar 2003, Max Bowsher wrote:

> Igor Pechtchanski wrote:
> > I'm getting intermittent hangs in latest build of setup.  Gdb
> > backtrace shows:
> > (gdb) thread 1
> > [Switching to thread 1 (thread 3084.0xb18)]#0  0x77e11b80 in USER32!DestroyWindow 
> > () from /cygdrive/c/WINNT/system32/USER32.DLL
> > (gdb) where #0  0x77e11b80 in USER32!DestroyWindow ()
> >from /cygdrive/c/WINNT/system32/USER32.DLL
> > #1  0x77e11e09 in USER32!GetMessageW ()
> >from /cygdrive/c/WINNT/system32/USER32.DLL
> > #2  0x7171f4f6 in FlatSB_EnableScrollBar ()
> > #3  0x7171f2dd in FlatSB_EnableScrollBar ()
> > #4  0x71721ba8 in DrawStatusTextW ()
> > #5  0x0043683c in PropSheet::Create(Window const*, unsigned long) (
> > this=0x22f9b0, Parent=0x0, Style=31840)
> > at /usr/src/setup-cvs/setup/propsheet.cc:193
> > #6  0x004250de in WinMain (h=0x40, hPrevInstance=0x0,
> > command_line=0x2332c6 "", cmd_show=1)
> > at /usr/src/setup-cvs/setup/main.cc:370
> > #7  0x00456a43 in main () at
> > /usr/include/c++/3.2/bits/char_traits.h:147 (gdb)
> > This is apparently somewhere inside the PropertySheet() Windows
> > function.
> >
> > Is this a known issue?  What would be the best way to debug this?
>
> I haven't run into this.
>
> That backtrace looks a little odd - anyone know whether gdb could be
> incorrectly decoding the addresses in Windows DLLs?
>
> Exactly what is happening when the hang occurs?
> Max.

It seems to happen when setup is downloading packages.  The package
progress bar is usually at about 97%, so I'm guessing it has to do with
download completion or switching to a new package to download.  The last
action the log shows (for bison) is

get_url_to_file 
ftp://archive.progeny.com/cygwin/release/bison/bison-20030307-1.tar.bz2 
C:\install\Cygwin/ftp%3a%2f%2farchive.progeny.com%2fcygwin/release/bison/bison-20030307-1.tar.bz2.tmp

but no

2003/03/13 17:45:19 Downloaded 
C:\install\Cygwin/ftp%3a%2f%2farchive.progeny.com%2fcygwin/release/binutils/bison-20030307-1.tar.bz2

as it does with other packages.  Setup simply hangs there, but the
controls work (i.e., I can close it, press "Cancel", etc.).  The package
file
C:\install\Cygwin\ftp%3a%2f%2farchive.progeny.com%2fcygwin/release/binutils/bison-20030307-1.tar.bz2.tmp
is slightly smaller than the one on the mirror (232k local vs. 238k on the
mirror) -- maybe an unflushed buffer?

This is confusing.  Hope this helps someone make sense out of it.
Igor
-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune

Re: [PATCH] Add logging of move-on-reboot scheduling

[Igor Pechtchanski]

On Thu, 13 Mar 2003, Igor Pechtchanski wrote:

> On 14 Mar 2003, Robert Collins wrote:
>
> > On Fri, [EMAIL PROTECTED]:07, Igor Pechtchanski wrote:
> > > On 14 Mar 2003, Robert Collins wrote:
> >
> > > > Please Extract Method on the duplicate code here.
> > > > Cheers,
> > > > Rob
> > >
> > > Umm, shouldn't we do this as two separate steps?  There's duplicate code
> > > galore all over that function...  Or do you want it all in one patch?
> >
> > I was only worried about the code that is being touched by the patch. I
> > think it's sufficiently small that a single patch will be ok.
> >
> > Just describe in the patch log the two steps.
> > Cheers,
> > Rob
>
> Actually, the rest of the function wasn't that bad, and I might as well do
> it [EMAIL PROTECTED], at least for the logging.
> Igor
> ==
> ChangeLog:
> 2003-03-13  Igor Pechtchanski <[EMAIL PROTECTED]>
>
> * install.cc (install_one_source): Add logging for
> successful replace-on-reboot scheduling.  Factor out
> duplicate code.  Set rebootneeded on Win9x.
> (log_ror_failure) New static function.
> (log_ror_success) New static function.

Oops!  Stale patch (doesn't even compile).  I'm properly ashamed.  New one
attached.  ChangeLog still applies.
Igor
-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune
Index: install.cc
===
RCS file: /cvs/cygwin-apps/setup/install.cc,v
retrieving revision 2.57
diff -u -p -r2.57 install.cc
--- install.cc  12 Mar 2003 22:15:25 -  2.57
+++ install.cc  13 Mar 2003 23:30:52 -
@@ -161,6 +161,30 @@ replace_one (packagemeta & pkg)
   return errors;
 }
 
+/* log failed scheduling of replace-on-reboot of a given file. */
+/* also increment errors. */
+static void
+log_ror_failure (String const &fn, int &errors)
+{
+  log (LOG_TIMESTAMP,
+   "Unable to schedule reboot replacement of file %s with %s (Win32 Error %ld)",
+   cygpath (String ("/") + fn).cstr_oneuse(),
+   cygpath (String ("/") + fn + ".new").cstr_oneuse(),
+   GetLastError ());
+  ++errors;
+}
+
+/* log successful scheduling of replace-on-reboot of a given file. */
+/* also set rebootneeded. */
+static void
+log_ror_success (String const &fn, bool &rebootneeded)
+{
+  log (LOG_TIMESTAMP,
+   "Scheduled reboot replacement of file %s with %s",
+   cygpath (String ("/") + fn).cstr_oneuse(),
+   cygpath (String ("/") + fn + ".new").cstr_oneuse());
+  rebootneeded = true;
+}
 
 /* install one source at a given prefix. */
 static int
@@ -252,13 +276,7 @@ install_one_source (packagemeta & pkgm, 
  source, MAX_PATH);
  if (!len || len > MAX_PATH)
{
- log (LOG_TIMESTAMP,
-  "Unable to schedule reboot replacement of file %s with 
%s (Win32 Error %ld)",
-  cygpath (String ("/") + fn).cstr_oneuse(),
-  cygpath (String ("/") + fn +
-   ".new").cstr_oneuse(),
-  GetLastError ());
- ++errors;
+ log_ror_failure (fn, errors);
}
  else
{
@@ -269,29 +287,19 @@ install_one_source (packagemeta & pkgm, 
  dest, MAX_PATH);
  if (!len || len > MAX_PATH)
{
- log (LOG_TIMESTAMP,
-  "Unable to schedule reboot replacement of file %s 
with %s (Win32 Error %ld)",
-  cygpath (String ("/") + fn).cstr_oneuse(),
-  cygpath (String ("/") + fn +
-   ".new").cstr_oneuse(),
-  GetLastError ());
- ++errors;
-
+ log_ror_failure (fn, errors);
}
  else
/* trigger a replacement on reboot */
  if (!WritePrivateProfileString
("rename", dest, source, "WININIT.INI"))
{
- log (LOG_TIMESTAMP,
-  "Unable to schedule reboot replacement of file %s 
with %s (Win32 Error %ld)",
- 

Re: Setup cvs-HEAD hangs intermittently when downloading

[Max Bowsher]

Igor Pechtchanski wrote:
> I'm getting intermittent hangs in latest build of setup.  Gdb
> backtrace shows:
> (gdb) thread 1
> [Switching to thread 1 (thread 3084.0xb18)]#0  0x77e11b80 in
> USER32!DestroyWindow () from /cygdrive/c/WINNT/system32/USER32.DLL
> (gdb) where #0  0x77e11b80 in USER32!DestroyWindow ()
>from /cygdrive/c/WINNT/system32/USER32.DLL
> #1  0x77e11e09 in USER32!GetMessageW ()
>from /cygdrive/c/WINNT/system32/USER32.DLL
> #2  0x7171f4f6 in FlatSB_EnableScrollBar ()
> #3  0x7171f2dd in FlatSB_EnableScrollBar ()
> #4  0x71721ba8 in DrawStatusTextW ()
> #5  0x0043683c in PropSheet::Create(Window const*, unsigned long) (
> this=0x22f9b0, Parent=0x0, Style=31840)
> at /usr/src/setup-cvs/setup/propsheet.cc:193
> #6  0x004250de in WinMain (h=0x40, hPrevInstance=0x0,
> command_line=0x2332c6 "", cmd_show=1)
> at /usr/src/setup-cvs/setup/main.cc:370
> #7  0x00456a43 in main () at
> /usr/include/c++/3.2/bits/char_traits.h:147 (gdb)
> This is apparently somewhere inside the PropertySheet() Windows
> function.
>
> Is this a known issue?  What would be the best way to debug this?

I haven't run into this.

That backtrace looks a little odd - anyone know whether gdb could be
incorrectly decoding the addresses in Windows DLLs?

Exactly what is happening when the hang occurs?

Max.

Re: [PATCH] Add logging of move-on-reboot scheduling

[Igor Pechtchanski]

On 14 Mar 2003, Robert Collins wrote:

> On Fri, [EMAIL PROTECTED]:07, Igor Pechtchanski wrote:
> > On 14 Mar 2003, Robert Collins wrote:
>
> > > Please Extract Method on the duplicate code here.
> > > Cheers,
> > > Rob
> >
> > Umm, shouldn't we do this as two separate steps?  There's duplicate code
> > galore all over that function...  Or do you want it all in one patch?
>
> I was only worried about the code that is being touched by the patch. I
> think it's sufficiently small that a single patch will be ok.
>
> Just describe in the patch log the two steps.
> Cheers,
> Rob

Actually, the rest of the function wasn't that bad, and I might as well do
it all at once, at least for the logging.
Igor
==
ChangeLog:
2003-03-13  Igor Pechtchanski <[EMAIL PROTECTED]>

* install.cc (install_one_source): Add logging for
successful replace-on-reboot scheduling.  Factor out
duplicate code.  Set rebootneeded on Win9x.
(log_ror_failure) New static function.
(log_ror_success) New static function.

-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune
Index: install.cc
===
RCS file: /cvs/cygwin-apps/setup/install.cc,v
retrieving revision 2.57
diff -u -p -r2.57 install.cc
--- install.cc  12 Mar 2003 22:15:25 -  2.57
+++ install.cc  13 Mar 2003 23:30:52 -
@@ -161,6 +161,30 @@ replace_one (packagemeta & pkg)
   return errors;
 }
 
+/* log failed scheduling of replace-on-reboot of a given file. */
+/* also increment errors. */
+static int
+log_ror_failure (String const &fn, int &errors)
+{
+  log (LOG_TIMESTAMP,
+   "Unable to schedule reboot replacement of file %s with %s (Win32 Error %ld)",
+   cygpath (String ("/") + fn).cstr_oneuse(),
+   cygpath (String ("/") + fn + ".new").cstr_oneuse(),
+   GetLastError ());
+  ++errors;
+}
+
+/* log successful scheduling of replace-on-reboot of a given file. */
+/* also set rebootneeded. */
+static int
+log_ror_success (String const &fn, bool &rebootneeded)
+{
+  log (LOG_TIMESTAMP,
+   "Scheduled reboot replacement of file %s with %s",
+   cygpath (String ("/") + fn).cstr_oneuse(),
+   cygpath (String ("/") + fn + ".new").cstr_oneuse());
+  rebootneeded = true;
+}
 
 /* install one source at a given prefix. */
 static int
@@ -252,13 +276,7 @@ install_one_source (packagemeta & pkgm, 
  source, MAX_PATH);
  if (!len || len > MAX_PATH)
{
- log (LOG_TIMESTAMP,
-  "Unable to schedule reboot replacement of file %s with 
%s (Win32 Error %ld)",
-  cygpath (String ("/") + fn).cstr_oneuse(),
-  cygpath (String ("/") + fn +
-   ".new").cstr_oneuse(),
-  GetLastError ());
- ++errors;
+ log_ror_failure (fn, errors);
}
  else
{
@@ -269,29 +287,19 @@ install_one_source (packagemeta & pkgm, 
  dest, MAX_PATH);
  if (!len || len > MAX_PATH)
{
- log (LOG_TIMESTAMP,
-  "Unable to schedule reboot replacement of file %s 
with %s (Win32 Error %ld)",
-  cygpath (String ("/") + fn).cstr_oneuse(),
-  cygpath (String ("/") + fn +
-   ".new").cstr_oneuse(),
-  GetLastError ());
- ++errors;
-
+ log_ror_failure (fn, errors);
}
  else
/* trigger a replacement on reboot */
  if (!WritePrivateProfileString
("rename", dest, source, "WININIT.INI"))
{
- log (LOG_TIMESTAMP,
-  "Unable to schedule reboot replacement of file %s 
with %s (Win32 Error %ld)",
-  cygpath (String ("/") + fn).cstr_oneuse(),
-  cygpath (String ("/") + fn +
-   ".new").cstr_oneuse(),
-  GetLastError ());
- 

Re: [PATCH] Add logging of move-on-reboot scheduling

[Robert Collins]

On Fri, 2003-03-14 at 10:07, Igor Pechtchanski wrote:
> On 14 Mar 2003, Robert Collins wrote:

> > Please Extract Method on the duplicate code here.
> > Cheers,
> > Rob
> 
> Umm, shouldn't we do this as two separate steps?  There's duplicate code
> galore all over that function...  Or do you want it all in one patch?

I was only worried about the code that is being touched by the patch. I
think it's sufficiently small that a single patch will be ok.

Just describe in the patch log the two steps.

Cheers,
Rob

-- 
GPG key available at: .


signature.asc
Description: This is a digitally signed message part

Re: [PATCH] Add logging of move-on-reboot scheduling

[Igor Pechtchanski]

On 14 Mar 2003, Robert Collins wrote:

> On Fri, [EMAIL PROTECTED]:26, Igor Pechtchanski wrote:
> > Hi,
> >
> > Currently the replace-on-reboot files are only logged if unable to
> > schedule them.  I've run into this problem and wanted to know what files
> > will be replaced, and this information is apparently not in the log.  The
> > attached patch adds it.
> >
> > It also does this (untested) for Win9x.  In fact, is there a particular
> > reason rebootneeded is not set to true on Win9x?  Rob?
>
> Uhmm, no :}.

Ok, so I guessed right?

> Please Extract Method on the duplicate code here.
> Cheers,
> Rob

Umm, shouldn't we do this as two separate steps?  There's duplicate code
galore all over that function...  Or do you want it all in one patch?
Igor
-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune

Setup cvs-HEAD hangs intermittently when downloading

[Igor Pechtchanski]

I'm getting intermittent hangs in latest build of setup.  Gdb backtrace
shows:
(gdb) thread 1
[Switching to thread 1 (thread 3084.0xb18)]#0  0x77e11b80 in USER32!DestroyWindow () 
from /cygdrive/c/WINNT/system32/USER32.DLL
(gdb) where
#0  0x77e11b80 in USER32!DestroyWindow ()
   from /cygdrive/c/WINNT/system32/USER32.DLL
#1  0x77e11e09 in USER32!GetMessageW ()
   from /cygdrive/c/WINNT/system32/USER32.DLL
#2  0x7171f4f6 in FlatSB_EnableScrollBar ()
#3  0x7171f2dd in FlatSB_EnableScrollBar ()
#4  0x71721ba8 in DrawStatusTextW ()
#5  0x0043683c in PropSheet::Create(Window const*, unsigned long) (
this=0x22f9b0, Parent=0x0, Style=31840)
at /usr/src/setup-cvs/setup/propsheet.cc:193
#6  0x004250de in WinMain (h=0x40, hPrevInstance=0x0,
command_line=0x2332c6 "", cmd_show=1)
at /usr/src/setup-cvs/setup/main.cc:370
#7  0x00456a43 in main () at /usr/include/c++/3.2/bits/char_traits.h:147
(gdb)
This is apparently somewhere inside the PropertySheet() Windows function.

Is this a known issue?  What would be the best way to debug this?
Igor
-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune

Re: [PATCH] Add logging of move-on-reboot scheduling

[Robert Collins]

On Fri, 2003-03-14 at 09:26, Igor Pechtchanski wrote:
> Hi,
> 
> Currently the replace-on-reboot files are only logged if unable to
> schedule them.  I've run into this problem and wanted to know what files
> will be replaced, and this information is apparently not in the log.  The
> attached patch adds it.
> 
> It also does this (untested) for Win9x.  In fact, is there a particular
> reason rebootneeded is not set to true on Win9x?  Rob?

Uhmm, no :}.

Please Extract Method on the duplicate code here.

Cheers,
Rob
-- 
GPG key available at: .


signature.asc
Description: This is a digitally signed message part

setup needs a utility class..

[Robert Collins]

A really useful utility class would be a setup wizard page that simply
displays scrollable text, with carriage returns honoured.

Requirements:
* Can be (but is not mandatory to be) automatically switched to by the
user. (IOW, it can be accessible via Next/Back, or via setup.exe code
only.
* Multiple instances needed.
* Dynamic title?

This would be useful for:
* Showing README's.
* Showing the results of the log files.
* Showing ChangeLogs from applications.
* Showing the help for --help.

Rob

-- 
GPG key available at: .


signature.asc
Description: This is a digitally signed message part

[PATCH] Add logging of move-on-reboot scheduling

[Igor Pechtchanski]

Hi,

Currently the replace-on-reboot files are only logged if unable to
schedule them.  I've run into this problem and wanted to know what files
will be replaced, and this information is apparently not in the log.  The
attached patch adds it.

It also does this (untested) for Win9x.  In fact, is there a particular
reason rebootneeded is not set to true on Win9x?  Rob?
Igor
==
ChangeLog:
2003-03-13  Igor Pechtchanski <[EMAIL PROTECTED]>

* install.cc (install_one_source): Add logging for
successful replace-on-reboot scheduling.  Also set
rebootneeded on Win9x.

-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune
Index: install.cc
===
RCS file: /cvs/cygwin-apps/setup/install.cc,v
retrieving revision 2.57
diff -u -p -r2.57 install.cc
--- install.cc  12 Mar 2003 22:15:25 -  2.57
+++ install.cc  13 Mar 2003 22:14:07 -
@@ -291,7 +291,15 @@ install_one_source (packagemeta & pkgm, 
   GetLastError ());
  ++errors;
}
-
+ else
+   {
+ log (LOG_TIMESTAMP,
+ "Scheduled reboot replacement of file %s with %s",
+ cygpath (String ("/") + fn).cstr_oneuse(),
+ cygpath (String ("/") + fn +
+  ".new").cstr_oneuse());
+ rebootneeded = true;
+   }
}
}
  break;
@@ -315,7 +323,13 @@ install_one_source (packagemeta & pkgm, 
  ++errors;
}
  else
-   rebootneeded = true;
+   {
+ log (LOG_TIMESTAMP,
+  "Scheduled reboot replacement of file %s with %s",
+  cygpath (String ("/") + fn).cstr_oneuse(),
+  cygpath (String ("/") + fn + ".new").cstr_oneuse());
+ rebootneeded = true;
+   }
  break;
}
}

tcsh postinstall script

[Igor Pechtchanski]

Hi,

This is mostly for Corinna (as the tcsh maintainer)...
I just noticed what the tcsh postinstall script looks like.
Quoting:

> rm -f /usr/bin/csh /usr/bin/csh.exe /usr/bin/csh.lnk
> ln -s tcsh /usr/bin/csh

Isn't the above a bit brutal?  I mean, suppose the user has her own
/bin/csh script that invokes tcsh with some flags?  Or suppose she ported
some other csh and installed it as /usr/bin/csh.exe?  Wouldn't it make
more sense to do what the other postinstall scripts do and only link if
the file doesn't exist?
Igor
-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune

Re: [PATCH] Fix iterator bug in choose.cc

[Robert Collins]

On Fri, 2003-03-14 at 08:07, Igor Pechtchanski wrote:
> On 14 Mar 2003, Robert Collins wrote:
> 
> > On Fri, [EMAIL PROTECTED]:47, Igor Pechtchanski wrote:
> > > I noticed there is a bug in an iterator in choose.cc that causes "All" to
> > > be printed twice in the list of categories in setup.log.full.  The
> > > attached patch fixes that.  Note that, like the original, the patch will
> > > not work if a package is not in any category (but that shouldn't happen
> > > anyway, right?).
> > >   Igor
> > > ==
> > > 2003-03-13  Igor Pechtchanski <[EMAIL PROTECTED]>
> > >
> > >   * choose.cc (ChooserPage::OnNext): Fix iterator bug.
> >
> > Nice catch. Wouldn't
> > = pkg.categories.begin ();
> > ->  String all_categories = *(i++);
> >
> > Be an easier fix?
> 
> It's shorter, but I dislike the use of increment operators outside of
> loops.  Leads to obscure errors.  Personal preference, y'know... :-)
> But it's your preference that counts, so feel free to tweak it.
>   Igor

I will :}.

The reason in this case, is that (IMO) it's more clear to a casual
reader that we copy and increment, rather than copy, then loop beginning
with an increment on each cycle. And whats clears to a casual reader is
blindingly obvious to us :}.

Rob

-- 
GPG key available at: .


signature.asc
Description: This is a digitally signed message part

Re: [PATCH] Fix iterator bug in choose.cc

[Igor Pechtchanski]

On 14 Mar 2003, Robert Collins wrote:

> On Fri, [EMAIL PROTECTED]:47, Igor Pechtchanski wrote:
> > I noticed there is a bug in an iterator in choose.cc that causes "All" to
> > be printed twice in the list of categories in setup.log.full.  The
> > attached patch fixes that.  Note that, like the original, the patch will
> > not work if a package is not in any category (but that shouldn't happen
> > anyway, right?).
> >   Igor
> > ==
> > 2003-03-13  Igor Pechtchanski <[EMAIL PROTECTED]>
> >
> >   * choose.cc (ChooserPage::OnNext): Fix iterator bug.
>
> Nice catch. Wouldn't
> = pkg.categories.begin ();
> ->  String all_categories = *(i++);
>
> Be an easier fix?

It's shorter, but I dislike the use of increment operators outside of
loops.  Leads to obscure errors.  Personal preference, y'know... :-)
But it's your preference that counts, so feel free to tweak it.
Igor
-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune

Bug: multiple subsequent << to LogFile are lost?

[Igor Pechtchanski]

There seems to be an obscure stream bug in the LogFile class.  I'm getting
messages like

Warning, the default trust level for package Chooser results...
^

Note that the above is comprised of 2 parts.  Part 1 (underlined) is from
package_version.cc:384, and, apparently, the rest of the message is eaten
up somehow.  I have a feeling it's because the wrong ostream is returned
from operator<< (so subsequent <<'s are lost), but don't understand custom
streams well enough to fix that.
Igor
-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune

Re: [PATCH] Fix iterator bug in choose.cc

[Robert Collins]

On Fri, 2003-03-14 at 07:47, Igor Pechtchanski wrote:
> I noticed there is a bug in an iterator in choose.cc that causes "All" to
> be printed twice in the list of categories in setup.log.full.  The
> attached patch fixes that.  Note that, like the original, the patch will
> not work if a package is not in any category (but that shouldn't happen
> anyway, right?).
>   Igor
> ==
> 2003-03-13  Igor Pechtchanski <[EMAIL PROTECTED]>
> 
>   * choose.cc (ChooserPage::OnNext): Fix iterator bug.

Nice catch. Wouldn't
= pkg.categories.begin ();
->  String all_categories = *(i++);

Be an easier fix?
-- 
GPG key available at: .


signature.asc
Description: This is a digitally signed message part

[PATCH] Fix iterator bug in choose.cc

[Igor Pechtchanski]

I noticed there is a bug in an iterator in choose.cc that causes "All" to
be printed twice in the list of categories in setup.log.full.  The
attached patch fixes that.  Note that, like the original, the patch will
not work if a package is not in any category (but that shouldn't happen
anyway, right?).
Igor
==
2003-03-13  Igor Pechtchanski <[EMAIL PROTECTED]>

* choose.cc (ChooserPage::OnNext): Fix iterator bug.

-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune
Index: choose.cc
===
RCS file: /cvs/cygwin-apps/setup/choose.cc,v
retrieving revision 2.112
diff -u -p -r2.112 choose.cc
--- choose.cc   10 Mar 2003 06:22:16 -  2.112
+++ choose.cc   13 Mar 2003 20:10:19 -
@@ -604,8 +604,8 @@ ChooserPage::OnNext ()
  set ::const_iterator i 
= pkg.categories.begin ();
  String all_categories = *i;
- while (i != pkg.categories.end ())
-   all_categories += String (", ") + *(i++);
+ while (++i != pkg.categories.end ())
+   all_categories += String (", ") + (*i);
 
  log (LOG_BABBLE) << " categories=" << all_categories << endLog;
}

Re: New release of setup.exe (2.249.2.10)

[Robert Collins]

On Fri, 2003-03-14 at 00:58, Pierre A. Humblet wrote:
> Robert Collins wrote:
> > 
> > I'm happy to announce a new release of setup.exe. This version will
> > correctly set permissions for use with the ntsec setting of cygwin,
> > which is the default now.
> 
> Rob, that's too optimistic. It will *improve* the display of 
> permissions in the majority of cases. More fine tuning may
> be needed but my calls for exhaustive testing have met the same
> fate as the calls to test the snapshots.
> 
> Two things worry me: domain users and the Windows popup asking
> if setup should run as Administrator.

Well, I'm glad I snafued and got the wrong list...

Rob

-- 
GPG key available at: .


signature.asc
Description: This is a digitally signed message part

Re: Synthesized missing ChangeLog entries for setup/zlib

[Robert Collins]

On Fri, 2003-03-14 at 01:07, Max Bowsher wrote:
> Dug these out of the cvs log messages and inspection of the files. OK to
> commit?

Sure.
Rob
-- 
GPG key available at: .


signature.asc
Description: This is a digitally signed message part

Ok, to upload grace ? was Re: Pending packages status (10 Mar 2003)

[Pavel Tsekov]

On Thu, 13 Mar 2003, Max Bowsher wrote:

> Volker Quetschke wrote:
> > -BEGIN PGP SIGNED MESSAGE-
> > Hash: SHA1
> >
> > Hi!
> >
> >> Actually, maybe the postinstall *should* use -e, not -f - just in
> >> case someone wants to do some symlink trickery.
> >
> > Changed and uploaded!
> 
> In that case:
> 
> Ready for upload to s.r.c as soon as someone with the relevant permissions
> has a moment.
> 
> BTW, keep the URLs in the message:
> http://www.scytek.de/cygwin/grace-5.1.12-1.tar.bz2
> http://www.scytek.de/cygwin/grace-5.1.12-1-src.tar.bz2
> http://www.scytek.de/cygwin/setup.hint
> 
> 
> Max.
> 
> 

Pending setup patches (issue 9)

[Max Bowsher]

Goodbye generated files!
* Remove autotool-generated files from CVS
No objections received.
Robert, you can zap them at your leisure!

Patches queued for review:
~~
* Igor's postinstall ordering patch
Latest patch here:
http://cygwin.com/ml/cygwin-apps/2003-03/msg00293.html

PATCHES/VERIFICATION SOLICITED FOR RELEASE BLOCKERS:

* Remove *empty* directories on uninstalls. (done?)
* Correctly overwrite -r--r--r-- files. (done?)

Patches being worked on (not required for next release):

* Igor's postinstall logging
Many threads. Latest one starts here:
http://sources.redhat.com/ml/cygwin-apps/2003-03/msg00253.html

* Gary's big chooser patch
http://sources.redhat.com/ml/cygwin-apps/2003-02/msg00020.html
Updated patch in the works.

* Max's Preliminary colour-coding patch
http://sources.redhat.com/ml/cygwin-apps/2003-01/msg00361.html
Needs a key/legend plus design fixes. (Working on it)

Re: Pending packages status (10 Mar 2003)

[Max Bowsher]

Volker Quetschke wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> Hi!
>
>> Actually, maybe the postinstall *should* use -e, not -f - just in
>> case someone wants to do some symlink trickery.
>
> Changed and uploaded!

In that case:

Ready for upload to s.r.c as soon as someone with the relevant permissions
has a moment.

BTW, keep the URLs in the message:
http://www.scytek.de/cygwin/grace-5.1.12-1.tar.bz2
http://www.scytek.de/cygwin/grace-5.1.12-1-src.tar.bz2
http://www.scytek.de/cygwin/setup.hint


Max.

Re: Pending packages status (10 Mar 2003)

[Volker Quetschke]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi!

| Actually, maybe the postinstall *should* use -e, not -f - just in case
| someone wants to do some symlink trickery.
Changed and uploaded!

Volker

- --
PGP/GPG key  (ID: 0x9F8A785D)  available  from  wwwkeys.de.pgp.net
key-fingerprint 550D F17E B082 A3E9 F913  9E53 3D35 C9BA 9F8A 785D
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.1 (MingW32)
Comment: Using GnuPG with Netscape - http://enigmail.mozdev.org
iD8DBQE+cLF9PTXJup+KeF0RAqjfAKC3j8oA1baikh7ypCz2CNzCKx1x7wCeOHDi
Sb8DrLsf8J8x0QZt7BzoTr8=
=RQFr
-END PGP SIGNATURE-

Re: Pending packages status (10 Mar 2003)

[Max Bowsher]

Volker Quetschke wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> Hi Igor,
>
>> Hope the following chunk helps:
>>
>> #!/bin/sh
>> [ ! -e "/etc/gracerc" ] && cp -a /etc/gracerc.default /etc/gracerc
>> [ ! -e "/etc/gracerc.user" ] && cp -a /etc/gracerc.user.default
>> /etc/gracerc.user
>>
>> And include /etc/gracerc.default and /etc/gracerc.user.default in the
>> package (those can get overwritten).
>
> too late ;-) the version which got uploaded half an hour ago already
> contains a similar postinstall script.

Actually, maybe the postinstall *should* use -e, not -f - just in case
someone wants to do some symlink trickery.

Max.

Re: Pending packages status (10 Mar 2003)

[Volker Quetschke]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi Igor,

| Hope the following chunk helps:
|
| #!/bin/sh
| [ ! -e "/etc/gracerc" ] && cp -a /etc/gracerc.default /etc/gracerc
| [ ! -e "/etc/gracerc.user" ] && cp -a /etc/gracerc.user.default
/etc/gracerc.user
|
| And include /etc/gracerc.default and /etc/gracerc.user.default in the
| package (those can get overwritten).
too late ;-) the version which got uploaded half an hour ago already
contains a similar postinstall script.
Volker

- --
PGP/GPG key  (ID: 0x9F8A785D)  available  from  wwwkeys.de.pgp.net
key-fingerprint 550D F17E B082 A3E9 F913  9E53 3D35 C9BA 9F8A 785D
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.1 (MingW32)
Comment: Using GnuPG with Netscape - http://enigmail.mozdev.org
iD8DBQE+cKzyPTXJup+KeF0RAsOjAKCtlSwUK/0ojAn9OXERC3PBMLQLDgCbBDHY
G+OE8sk4PMVbPeowtcP8sCc=
=v1S0
-END PGP SIGNATURE-

Re: Pending packages status (10 Mar 2003)

[Igor Pechtchanski]

On Thu, 13 Mar 2003, Volker Quetschke wrote:

> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> Hi,
>
> I send the last mail before I read your mail Earnie.
>
> |>>> 1. grace
> |>>> ...
> |>> Max did a review in:
> |>> ~ 
> |>>
> |>> and all proposed changes are applied to the packages at the url
> |>> mentioned above.
> |>
> |> OK, I've completed the review I began there. I have the following notes:
> |>
> |> - The warning about gracerc and gracerc.user being overwritten on
> |> reinstall
> |> is in the README. I'm not sure very many people will read that. I suggest
> |> putting it in the comments actually in the files themselves.
> |
> | IIRC, this is a major flaw.  Package configuration files are to not be
> | overwritten upon reinstall.  You need to use postinstall scripts to
> | install initial configuration files and not overwrite exsiting
> | configuration files.
> If you plan to change the default settings you should use user defined
> settings, but yes I see the point.

What if you are a system administrator and want to change settings for all
users at once?  I agree with Earnie.

> |> - You could do change doc to /usr/grace/doc in the README file. This
> |> would
> |> make it more clear to grace newbies where to find the installed
> |> documentation.
> |
> | Uhm, you mean /usr/doc/grace or do you mean /usr/doc/Cygwin/grace.README?
> The original /usr/grace/README which gets copied to
> /usr/doc/grace-5.1.12 only talks about the doc directory, grace.README
> has the full information in it.
>
> |> Neither of these are critical - the current packages could be released
> |> as-is - but both of the above are minor improvements that should be
> |> considered.
> |
> | Not following these conventions are critical IMNSHO.
>
> Ok, if I find a few cycles I will implement the postinstall script.
>
> Volker

Hope the following chunk helps:

#!/bin/sh
[ ! -e "/etc/gracerc" ] && cp -a /etc/gracerc.default /etc/gracerc
[ ! -e "/etc/gracerc.user" ] && cp -a /etc/gracerc.user.default /etc/gracerc.user

And include /etc/gracerc.default and /etc/gracerc.user.default in the
package (those can get overwritten).
Igor
-- 
http://cs.nyu.edu/~pechtcha/
  |\  _,,,---,,_[EMAIL PROTECTED]
ZZZzz /,`.-'`'-.  ;-;;,_[EMAIL PROTECTED]
 |,4-  ) )-,_. ,\ (  `'-'   Igor Pechtchanski
'---''(_/--'  `-'\_) fL a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

Oh, boy, virtual memory! Now I'm gonna make myself a really *big* RAMdisk!
  -- /usr/games/fortune

Re: Pending packages status (10 Mar 2003)

[Max Bowsher]

Volker Quetschke wrote:
> The version that I'm currently uploading contains this README:
> 
> - 
> ~ GRACE - GRaphing, Advanced Computation and Exploration of data
> 
> Recommended reading:
> 
> /usr/doc/Cygwin/grace-5.1.12.README - Notes for the Cygwin package
> 
> COPYRIGHT   - legal stuff
> LICENSE - license terms (GPL)
> CHANGES - chronological list of changes
> /usr/grace/doc/FAQ.html - frequently asked questions (in HTML format)
> 
> The installation procedure is covered in the User's Guide
> (in the /usr/grace/doc directory).
> - 

Looks good! I think this is ready for release.

Max.

Re: Pending packages status (10 Mar 2003)

[Volker Quetschke]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
|>Hmm, I thought the:
|>
|>- -Recommended reading:
|>+Recommended reading: ( files are in /usr/grace/ )
|>
|>would be enough for that.
|
| Here is the file as it is now, |-indented:
|
| |  GRACE - GRaphing, Advanced Computation and Exploration of data
| |
| | Recommended reading: ( files are in /usr/grace/ )
|
| But they aren't!
Sorry, I thought they get installed, but they don't.
| | arch//README   - system-specific important notes
|
| I'm not sure if this even exists for Cygwin.
It does, but grace.README holds all the information, it
would only confuse. (The directory is in the source archive)
The version that I'm currently uploading contains this README:

- 
~ GRACE - GRaphing, Advanced Computation and Exploration of data
Recommended reading:

/usr/doc/Cygwin/grace-5.1.12.README - Notes for the Cygwin package

COPYRIGHT   - legal stuff
LICENSE - license terms (GPL)
CHANGES - chronological list of changes
/usr/grace/doc/FAQ.html - frequently asked questions (in HTML format)
The installation procedure is covered in the User's Guide
(in the /usr/grace/doc directory).
- 
Volker

- --
PGP/GPG key  (ID: 0x9F8A785D)  available  from  wwwkeys.de.pgp.net
key-fingerprint 550D F17E B082 A3E9 F913  9E53 3D35 C9BA 9F8A 785D
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.1 (MingW32)
Comment: Using GnuPG with Netscape - http://enigmail.mozdev.org
iD8DBQE+cKPaPTXJup+KeF0RAmNYAKDH5JDhnn3CEIl5zJRlTYmbGHcaoACgo2/i
xlk9JlDKgC7paYyO6ycZ4/A=
=G+Jd
-END PGP SIGNATURE-

Re: Pending packages status (10 Mar 2003)

[Max Bowsher]

Volker Quetschke wrote:
> Hmm, I thought the:
> 
> - -Recommended reading:
> +Recommended reading: ( files are in /usr/grace/ )
> 
> would be enough for that.

Here is the file as it is now, |-indented:

|  GRACE - GRaphing, Advanced Computation and Exploration of data
| 
| Recommended reading: ( files are in /usr/grace/ )

But they aren't!

| 
| COPYRIGHT   - legal stuff
| LICENSE - license terms (GPL)
| CHANGES - chronological list of changes

These three are in the same dir as README

| doc/FAQ.html- frequently asked questions (in HTML format)

This is in /usr/grace/doc

| arch//README   - system-specific important notes

I'm not sure if this even exists for Cygwin.

| The installation procedure is covered in the User's Guide
| (in the /usr/grace/doc directory).


Max.

Re: Pending packages status (10 Mar 2003)

[Volker Quetschke]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi!

|1. grace
|...
|>>OK, I've completed the review I began there. I have the following
|>>notes:
|>>
|>>- The warning about gracerc and gracerc.user being overwritten on
|>>reinstall is in the README. I'm not sure very many people will read
|>>that. I suggest putting it in the comments actually in the files
|>>themselves.
|>Good idea, done!.
| Actually, now Earnie mentions it, someone *might* want to have all-users
| configuration.
I implemented a postinstall script.

|>>- You could do change doc to /usr/grace/doc in the README file. This
|>>would make it more clear to grace newbies where to find the installed
|>>documentation.
|>
|>I forgot that the README gets copied to /usr/doc/grace-5.1.12/ , its
|>original place is /usr/grace/. I will patch /usr/grace/doc into this
|>file. Done!
|
| You got one occurrence, but missed the other. 'doc/FAQ.html' should become
| '/usr/grace/doc/FAQ.html'
Hmm, I thought the:

- -Recommended reading:
+Recommended reading: ( files are in /usr/grace/ )
would be enough for that.

The last changed version (still with doc/FAQ.html) is uploaded.

Volker

- --
PGP/GPG key  (ID: 0x9F8A785D)  available  from  wwwkeys.de.pgp.net
key-fingerprint 550D F17E B082 A3E9 F913  9E53 3D35 C9BA 9F8A 785D
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.1 (MingW32)
Comment: Using GnuPG with Netscape - http://enigmail.mozdev.org
iD8DBQE+cJjLPTXJup+KeF0RAq4+AKDCCCwG/TkxuLnx/iy5wVCQO3yfVwCgkQ1B
yXsjhrkxOaWd7WKTViyttkc=
=rZvq
-END PGP SIGNATURE-

Synthesized missing ChangeLog entries for setup/zlib

[Max Bowsher]

Dug these out of the cvs log messages and inspection of the files. OK to
commit?

Index: ChangeLog
===
RCS file: /cvs/cygwin-apps/setup/zlib/ChangeLog,v
retrieving revision 2.6
diff -u -p -r2.6 ChangeLog
--- ChangeLog 13 Mar 2003 12:34:48 - 2.6
+++ ChangeLog 13 Mar 2003 14:06:58 -
@@ -21,6 +21,28 @@
  * zutil.c (z_error): Const correctness change.
  * zutil.h (z_error): Ditto.

+2001-04-18  Christopher Faylor  <[EMAIL PROTECTED]>
+
+ * configure.in: Force NM substitution.
+ * configure: Regenerate.
+
+2000-07-07  DJ Delorie  <[EMAIL PROTECTED]>
+
+ * gzio.c: Add gzctell() for progess displays.
+ Make gzsetparams() a no-op returning Z_OK when NO_DEFLATE is defined.
+ * zlib.h: And declare it.
+
+2000-04-02  Christopher Faylor  <[EMAIL PROTECTED]>
+
+ * configure.in: New file.
+ * Makefile.am: New file.
+ * aclocal.m4: New file.
+ * acinclude.m4: New file.
+ * configure: Generate from configure.in, replacing non-autotool
+ version.
+ * Makefile.in: Generate from Makefile.am, replacing non-autotool
+ version.
+
   ChangeLog file for zlib

 Changes in 1.1.4 (11 March 2002)

Re: New release of setup.exe (2.249.2.10)

[Pierre A. Humblet]

Robert Collins wrote:
> 
> I'm happy to announce a new release of setup.exe. This version will
> correctly set permissions for use with the ntsec setting of cygwin,
> which is the default now.

Rob, that's too optimistic. It will *improve* the display of 
permissions in the majority of cases. More fine tuning may
be needed but my calls for exhaustive testing have met the same
fate as the calls to test the snapshots.

Two things worry me: domain users and the Windows popup asking
if setup should run as Administrator.

Pierre

Re: Pending packages status (10 Mar 2003)

[Volker Quetschke]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi,

I send the last mail before I read your mail Earnie.

|>>> 1. grace
|>>> ...
|>> Max did a review in:
|>> ~ 
|>>
|>> and all proposed changes are applied to the packages at the url
|>> mentioned above.
|>
|> OK, I've completed the review I began there. I have the following notes:
|>
|> - The warning about gracerc and gracerc.user being overwritten on
|> reinstall
|> is in the README. I'm not sure very many people will read that. I suggest
|> putting it in the comments actually in the files themselves.
|
| IIRC, this is a major flaw.  Package configuration files are to not be
| overwritten upon reinstall.  You need to use postinstall scripts to
| install initial configuration files and not overwrite exsiting
| configuration files.
If you plan to change the default settings you should use user defined
settings, but yes I see the point.
|> - You could do change doc to /usr/grace/doc in the README file. This
|> would
|> make it more clear to grace newbies where to find the installed
|> documentation.
|
| Uhm, you mean /usr/doc/grace or do you mean /usr/doc/Cygwin/grace.README?
The original /usr/grace/README which gets copied to
/usr/doc/grace-5.1.12 only talks about the doc directory, grace.README
has the full information in it.
|> Neither of these are critical - the current packages could be released
|> as-is - but both of the above are minor improvements that should be
|> considered.
|>
|
| Not following these conventions are critical IMNSHO.
Ok, if I find a few cycles I will implement the postinstall script.

Volker

- --
PGP/GPG key  (ID: 0x9F8A785D)  available  from  wwwkeys.de.pgp.net
key-fingerprint 550D F17E B082 A3E9 F913  9E53 3D35 C9BA 9F8A 785D
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.1 (MingW32)
Comment: Using GnuPG with Netscape - http://enigmail.mozdev.org
iD8DBQE+cITqPTXJup+KeF0RAmPOAKCKPFWgW2uFEUF4LzoIfHj3rGJHWwCggyUY
B4avT/5BMGWeXdHjLl1Egs8=
=E7ed
-END PGP SIGNATURE-

Re: Pending packages status (10 Mar 2003)

[Max Bowsher]

Volker Quetschke wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> Hi!
>
 1. grace

 date   : 25 Nov 2002
 version: 5.1.12-1
 status : updated package available for review
 notes  : http://www.cygwin.com/ml/cygwin-apps/2002-11/msg00322.html
 reviews: http://www.cygwin.com/ml/cygwin-apps/2003-03/msg00254.html
 votes  : 2 (Lapo and Robert)
 url: http://www.scytek.de/cygwin/grace-5.1.12-1.tar.bz2
 http://www.scytek.de/cygwin/grace-5.1.12-1-src.tar.bz2
 http://www.scytek.de/cygwin/setup.hint
>>>
>>> Max did a review in:
>>> ~ 
>>>
>>> and all proposed changes are applied to the packages at the url
>>> mentioned above.
>>
>> OK, I've completed the review I began there. I have the following
>> notes:
>>
>> - The warning about gracerc and gracerc.user being overwritten on
>> reinstall is in the README. I'm not sure very many people will read
>> that. I suggest putting it in the comments actually in the files
>> themselves.
>
> Good idea, done!.

Actually, now Earnie mentions it, someone *might* want to have all-users
configuration.

Better do the "install .default files, and copy into place iff not already
in existance" thing.

>> - You could do change doc to /usr/grace/doc in the README file. This
>> would make it more clear to grace newbies where to find the installed
>> documentation.
>
> I forgot that the README gets copied to /usr/doc/grace-5.1.12/ , its
> original place is /usr/grace/. I will patch /usr/grace/doc into this
> file. Done!

You got one occurrence, but missed the other. 'doc/FAQ.html' should become
'/usr/grace/doc/FAQ.html'

Max.

Re: Pending packages status (10 Mar 2003)

[Volker Quetschke]

-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hi!

|>>1. grace
|>>
|>>date   : 25 Nov 2002
|>>version: 5.1.12-1
|>>status : updated package available for review
|>>notes  : http://www.cygwin.com/ml/cygwin-apps/2002-11/msg00322.html
|>>reviews: http://www.cygwin.com/ml/cygwin-apps/2003-03/msg00254.html
|>>votes  : 2 (Lapo and Robert)
|>>url: http://www.scytek.de/cygwin/grace-5.1.12-1.tar.bz2
|>> http://www.scytek.de/cygwin/grace-5.1.12-1-src.tar.bz2
|>> http://www.scytek.de/cygwin/setup.hint
|>
|>Max did a review in:
|>~ 
|>
|>and all proposed changes are applied to the packages at the url
|>mentioned above.
|
| OK, I've completed the review I began there. I have the following notes:
|
| - The warning about gracerc and gracerc.user being overwritten on
reinstall
| is in the README. I'm not sure very many people will read that. I suggest
| putting it in the comments actually in the files themselves.
Good idea, done!.

| - You could do change doc to /usr/grace/doc in the README file. This would
| make it more clear to grace newbies where to find the installed
| documentation.
I forgot that the README gets copied to /usr/doc/grace-5.1.12/ , its
original place is /usr/grace/. I will patch /usr/grace/doc into this
file. Done!
All changes are in the packages at the url mentioned above.

Volker

- --
PGP/GPG key  (ID: 0x9F8A785D)  available  from  wwwkeys.de.pgp.net
key-fingerprint 550D F17E B082 A3E9 F913  9E53 3D35 C9BA 9F8A 785D
-BEGIN PGP SIGNATURE-
Version: GnuPG v1.2.1 (MingW32)
Comment: Using GnuPG with Netscape - http://enigmail.mozdev.org
iD8DBQE+cIAxPTXJup+KeF0RAp/qAJ4hhrKIYqGneFvvPs43GgiKbLkxdQCfdqLN
i5kwKh0KUAto6mcMovd4NSk=
=cWlL
-END PGP SIGNATURE-

Re: Pending packages status (10 Mar 2003)

[Earnie Boyd]

Max Bowsher wrote:
Volker Quetschke wrote:

Hi!


1. grace

date   : 25 Nov 2002
version: 5.1.12-1
status : updated package available for review
notes  : http://www.cygwin.com/ml/cygwin-apps/2002-11/msg00322.html
reviews: http://www.cygwin.com/ml/cygwin-apps/2003-03/msg00254.html
votes  : 2 (Lapo and Robert)
url: http://www.scytek.de/cygwin/grace-5.1.12-1.tar.bz2
http://www.scytek.de/cygwin/grace-5.1.12-1-src.tar.bz2
http://www.scytek.de/cygwin/setup.hint
Max did a review in:
~ 
and all proposed changes are applied to the packages at the url
mentioned above.


OK, I've completed the review I began there. I have the following notes:

- The warning about gracerc and gracerc.user being overwritten on reinstall
is in the README. I'm not sure very many people will read that. I suggest
putting it in the comments actually in the files themselves.
IIRC, this is a major flaw.  Package configuration files are to not be 
overwritten upon reinstall.  You need to use postinstall scripts to 
install initial configuration files and not overwrite exsiting 
configuration files.

- You could do change doc to /usr/grace/doc in the README file. This would
make it more clear to grace newbies where to find the installed
documentation.
Uhm, you mean /usr/doc/grace or do you mean /usr/doc/Cygwin/grace.README?

Neither of these are critical - the current packages could be released
as-is - but both of the above are minor improvements that should be
considered.
Not following these conventions are critical IMNSHO.

Earnie.

Re: New release of setup.exe (2.249.2.10)

[Pavel Tsekov]

Wrong list ?

On 13 Mar 2003, Robert Collins wrote:

> I'm happy to announce a new release of setup.exe. This version will
> correctly set permissions for use with the ntsec setting of cygwin,
> which is the default now. Additionally it allows the use of http proxies
> that require a username with no password.
> 
> To download this version, simply click on the 'install cygwin now' link
> on the cygwin homepage, or wait for setup.exe to be copied out to your
> preferred mirror.

New release of setup.exe (2.249.2.10)

[Robert Collins]

I'm happy to announce a new release of setup.exe. This version will
correctly set permissions for use with the ntsec setting of cygwin,
which is the default now. Additionally it allows the use of http proxies
that require a username with no password.

To download this version, simply click on the 'install cygwin now' link
on the cygwin homepage, or wait for setup.exe to be copied out to your
preferred mirror.

Cheers,
Rob


2003-03-05  Max Bowsher  <[EMAIL PROTECTED]>

* netio.cc (save_dialog): Allocate an empty string if the password is
null, so that an empty password is accepted elsewhere in setup.
(check_if_enable_ok): An empty password is OK.

2003-03-04  Pavel Tsekov  <[EMAIL PROTECTED]>

* compress_gz.cc (compress_gz::construct): Always initialize key members
to ensure the destructor does the right thing.

2003-02-28  Max Bowsher  <[EMAIL PROTECTED]>

* main.cc (set_default_sec): Add logging.

2002-11-26  Robert Collins  <[EMAIL PROTECTED]>

Backport from HEAD
* desktop.cc (etc_profile): Remove.
(make_etc_profile): Remove.
(uexists): Remove.
(make_passwd_group): Remove.
(do_desktop_setup): Don't call removed functions.

2003-02-28  Max Bowsher  <[EMAIL PROTECTED]>

Backport from HEAD
* Makefile.am (%.o: %.rc): Make windres use $(CC) not hardcoded gcc.
(inilint_LDADD): Remove -lstdc++.
(setup_LDADD): Ditto.

2003-02-28  Pierre Humblet  <[EMAIL PROTECTED]>

Backported by Pavel Tsekov <[EMAIL PROTECTED]>
* main.cc (set_default_dacl): Replace with set_default_sec.
(set_default_sec): Attempt to change the default group to
Users or Administrators if it is None, in addition to setting
Everyone in the default DACL.
(main): Call set_default_sec instead of set_default_dacl.
-- 
GPG key available at: .


signature.asc
Description: This is a digitally signed message part

Re: Pending packages status (10 Mar 2003)

[Max Bowsher]

Volker Quetschke wrote:
> Hi!
>
>> 1. grace
>>
>> date   : 25 Nov 2002
>> version: 5.1.12-1
>> status : updated package available for review
>> notes  : http://www.cygwin.com/ml/cygwin-apps/2002-11/msg00322.html
>> reviews: http://www.cygwin.com/ml/cygwin-apps/2003-03/msg00254.html
>> votes  : 2 (Lapo and Robert)
>> url: http://www.scytek.de/cygwin/grace-5.1.12-1.tar.bz2
>>  http://www.scytek.de/cygwin/grace-5.1.12-1-src.tar.bz2
>>  http://www.scytek.de/cygwin/setup.hint
>
> Max did a review in:
> ~ 
>
> and all proposed changes are applied to the packages at the url
> mentioned above.

OK, I've completed the review I began there. I have the following notes:

- The warning about gracerc and gracerc.user being overwritten on reinstall
is in the README. I'm not sure very many people will read that. I suggest
putting it in the comments actually in the files themselves.

- You could do change doc to /usr/grace/doc in the README file. This would
make it more clear to grace newbies where to find the installed
documentation.

Neither of these are critical - the current packages could be released
as-is - but both of the above are minor improvements that should be
considered.


Max.

Re: nasm-0.98.36-1

[Pavel Tsekov]

On Thu, 13 Mar 2003, Dean Scarff wrote:

> This is the new version of nasm, released on sf March 13.  The pdf issue from the 
> 0.98.35 port is still current: again the pdf has been ommitted.
> 
> setup.hint:
> http://www.polarhome.com:753/~p00ya/up/setup.hint
> src package:
> http://www.polarhome.com:753/~p00ya/up/nasm-0.98.36-1-src.tar.bz2
> bin package:
> http://www.polarhome.com:753/~p00ya/up/nasm-0.98.36-1.tar.bz2
> 
> Please upload.

Uploaded. Please, send an announcement to cygwin-apps in a few hours.

Thanks!

nasm-0.98.36-1

[Dean Scarff]

This is the new version of nasm, released on sf March 13.  The pdf issue from the 
0.98.35 port is still current: again the pdf has been ommitted.

setup.hint:
http://www.polarhome.com:753/~p00ya/up/setup.hint
src package:
http://www.polarhome.com:753/~p00ya/up/nasm-0.98.36-1-src.tar.bz2
bin package:
http://www.polarhome.com:753/~p00ya/up/nasm-0.98.36-1.tar.bz2

Please upload.

Cheers,
Dean Scarff
-- 
___
http://www.operamail.com
Now with OperaMail Premium for only US$29.99/yr

Powered by Outblaze