Keith Marshall | 20 Jun 22:08 2016
Picon
Picon

RFC: <excpt.h>: __try1/__except1 are broken; are they worth fixing?


Guys,

Further reviewing Cesar's clone of our original CVS:
https://sourceforge.net/u/cstrauss/mingwrt/ci/56e6189f267527e4dd710d55a16de39f7bf21c4a/

I stumbled upon:
https://sourceforge.net/p/mingw/bugs/1328/

IMO, this bug report was inappropriately closed, as invalid.  I've
reopened it, accordingly, because, as may be seen from the extensive
analysis which I've appended thereto, this commit:
https://sourceforge.net/u/cstrauss/mingwrt/ci/3eac8de7de5be34a9b58918ae8500697a7ddc1c4/

does nothing even remotely useful to resolve the issue; in fact the
entire __try1/__except1 implementation is so utterly TARFUBAR, that I
wonder if it is worth expending any effort on fixing it, or should it
simply be allowed to disappear without trace?

Thoughts?

--

-- 
Regards,
Keith.

Public key available from keys.gnupg.net
Key fingerprint: C19E C018 1547 DE50 E1D4 8F53 C0AD 36C6 347E 5A3F
Keith Marshall | 14 Jun 16:00 2016
Picon
Picon

Can someone please help me to understand this commit?


Guys,

Looking at Cesar's clone of our original CVS, I'm utterly confused by:
https://sourceforge.net/u/cstrauss/mingwrt/ci/56e6189f267527e4dd710d55a16de39f7bf21c4a/

It appears to have been committed, originally, by Earnie, and I have
too issues with it:

1) The declaration of _locale_t surely belongs in <locale.h>, not in
<_mingw.h>.  Furthermore, as a reference to an opaque type, we could
just as well define it as a pointer to void, (although a pointer to a
single incomplete structure type may make sense); I can certainly see
absolutely no advantage in defining it as an aggregate of *two* other
opaque structural types.

2) I am utterly mystified by the changes to mingwex/tsearch.c, (which
appear to have accidentally have crept into the commit).  AFAICT, the
original code, (which has come from NetBSD), works just fine; I have
no idea what issue these changes were intended to resolve, but once
again AFAICT, they should have absolutely no functional effect whatsoever.

What am I missing?

--

-- 
Regards,
Keith.

Public key available from keys.gnupg.net
Key fingerprint: C19E C018 1547 DE50 E1D4 8F53 C0AD 36C6 347E 5A3F
(Continue reading)

Keith Marshall | 11 Jun 12:22 2016
Picon
Picon

Is WSL still a viable identification choice for our system libraries?


Guys,

I know we used it first, but Microsoft have hijacked WSL, as a TLA
designation for their "Windows Services for Linux".  We certainly
don't want to take on Microsoft over it; do we really want to live
with the potential confusion it may cause, or should we consider an
alternative for our own use?

Any thoughts?  Alternative suggestions?

--

-- 
Regards,
Keith.

Public key available from keys.gnupg.net
Key fingerprint: C19E C018 1547 DE50 E1D4 8F53 C0AD 36C6 347E 5A3F
Keith Marshall | 31 May 19:03 2016
Picon
Picon

RFC: updating CPU identification macros


Guys,

As I've noted at https://sourceforge.net/p/mingw/bugs/2301/, in both
<windows.h> and <winnt.h> I see:

  #if defined(__i686__) && !defined(_M_IX86)
  #define _M_IX86 600
  #elif defined(__i586__) && !defined(_M_IX86)
  #define _M_IX86 500
  #elif defined(__i486__) && !defined(_M_IX86)
  #define _M_IX86 400
  #elif defined(__i386__) && !defined(_M_IX86)
  #define _M_IX86 300
  #endif

This may have been okay when Anders Norlander first wrote it,
(although why he chose to duplicate it -- unnecessarily, as it happens
-- in the two headers noted above is something of a puzzle), but I
doubt that it is good enough today.

I've indicated how I propose to update this, on the 2301 ticket, but
I'd particularly appreciate your thoughts on which, if any, of the
other Microsoft predefined macros, as documented at:
https://msdn.microsoft.com/en-us/library/b0084kay(v=vs.140).aspx
we should consider adding.  Also, how much of:

  #if defined(_M_IX86) && !defined(_X86_)
  #define _X86_
  #elif defined(_M_ALPHA) && !defined(_ALPHA_)
(Continue reading)

Keith Marshall | 18 May 12:48 2016
Picon
Picon

RFC: <sys/stat.h>: S_IFBLK is unsupported on Win32


Guys,

I'd like to put https://sourceforge.net/p/mingw/bugs/1146 to bed.

I'd appreciate your opinions on the options for resolution, which I've
noted at https://sourceforge.net/p/mingw/bugs/1146/#9cc1/9ca0/e11c

--

-- 
Regards,
Keith.

Public key available from keys.gnupg.net
Key fingerprint: C19E C018 1547 DE50 E1D4 8F53 C0AD 36C6 347E 5A3F
Keith Marshall | 12 May 23:40 2016
Picon
Picon

RFC: time for a new mingwrt/w32api release?


Guys,

It's been a long time since we last released an update for this pair
of packages.  I'd like to prepare a new combined release, preferably
with synchronized version numbers, but given the problems inherent in
the last combined release, (v4.0), I think that any new release should
be derived from the current "legacy" branch, and should skip
immediately to a synchronized version number of 5.0.

To prepare for this, I'm considering:

1) Rename the existing "5.0-dev" branch as "5.0-deferred".
2) Create a new "5.0-dev" branch, based on the tip of "legacy".
3) Implement an integrated build infrastructure, preserving the
   existing separate package configuration, but building both
   together, with enforced version number synchronization.
4) Create an early 5.0 release, substantially unchanged from
   the current "legacy" tip, (possibly with a few new features
   which are already "good to go").

Any thoughts?  Objections?

--

-- 
Regards,
Keith.

Public key available from keys.gnupg.net
Key fingerprint: C19E C018 1547 DE50 E1D4 8F53 C0AD 36C6 347E 5A3F
(Continue reading)

Keith Marshall | 13 Mar 13:05 2016
Picon
Picon

Fwd: MinGW Updates [Win32]


Guys,
I received the following messages, directly from Alex Jones.

Alex,
Please don't mail me directly; you are subscribed to MinGW-Dvlpr, so
please use it.  That way, everyone gets to know what you are planning
to work on, so can both offer advice, and avoid duplicating your work.

. . .

Morning Keith. We have been emailing back & forth for a little while.
But want to start helping with MinGW. Willing to help anywhere needed.
But prefer to work with Windows API/Runtime updates. Started looking
at existing tickets. Want to help with the following tickets:

   - #2236 Touch and gestures functionality
     - Assume I need a Microsoft touch device e.g. Microsoft Surface
   - #2230 GetLayout() missing in w32api libgdi32.a
   - #2274 Various modern Win32 API definitions missing
   - #2262 [patch] add Registry API functions and macros introduced
     in Windows Vista

Would like to add the following tickets also:
   - Add ODS_HOTLIGHT [0x40] in winuser.h
   - Add QWORD [#define QWORD DWORD64]

Please let me know how to set up my development environment to most
efficiently help with MinGW tickets. E.g. Git/VirtualBox/MinGW/...

(Continue reading)

waterlan | 13 Mar 00:14 2016
Picon
Picon

ncurses for msys in mintty

Hi,

I built ncurses 6.0 for MSYS for use in mintty. I built it in the first 
place for my own wcd project. I'm willing to share this. The files could 
be put under MSYS/Contributed/ncurses/. The MinGW installer doesn't 
support contributed libraries for msys, but people could copy the files 
manually.

Is this interesting for msys?

regards,

--

-- 
Erwin Waterlander
http://waterlan.home.xs4all.nl/

------------------------------------------------------------------------------
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://pubads.g.doubleclick.net/gampad/clk?id=278785111&iu=/4140
arthur | 11 Jan 20:50 2016
Picon

git installation

Happy New Year, List!

I used to install my PC with mingw-get but recently I tried a fresh 
installation of MinGW with the GUI installer. I installed the 
'mingw32-base', 'msys-base' and 'mingw-deleloper-toolkit'. Then in MSYS 
I have cvs (from MSYS) and svn (from my TortoiseSVN) but not git.

I know I can install cyg-win or git-for-windows to get git to my PC, but 
I think there may be a more clean way to add only git to my MinGW32/MSYS 
without other MSYS/MinGW64 files. What is the recommended way to add git.

Thanks,

Arthur

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
waterlan | 9 Jan 09:35 2016
Picon
Picon

porting scripts

Hi,

I looked at the latests mingw32 package upgrades (gettext and binutils) 
and I could not find a copy of the porting scripts in the src packages. 
I think they should be in there, because without them it becomes extra 
difficult for other people to create an upgrade.

I like the approach of the msys2 project, where all porting scripts are 
in a git repository. This makes it easy for people to contribute. It 
also makes it a lot easier if you would like to do a mass rebuild with a 
new gcc version.

regards,

--

-- 
Erwin Waterlander
http://waterlan.home.xs4all.nl/

------------------------------------------------------------------------------
Site24x7 APM Insight: Get Deep Visibility into Application Performance
APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
Monitor end-to-end web transactions and take corrective actions now
Troubleshoot faster and improve end-user experience. Signup Now!
http://pubads.g.doubleclick.net/gampad/clk?id=267308311&iu=/4140
waterlan | 5 Jan 07:05 2016
Picon
Picon

ugrade mingw32-libunistring problem

Hi,

I upgraded mingw32-libunistring to version 0.9.6, but in the catalogue 
the version sticks to 0.9.3.

I updated file contrib/mingw32-libunistring.xml, and did a normal build.
Only contrib/issue.log changed, and there was only one unpublished file: 
contrib/unpublished/mingw32-libunistring.xml.lzma.
I'm not sure if this is normal. Normally common/issue.log also changes, 
and there are new packet list lzma files generated.

I did upload everything.

regards,

--

-- 
Erwin Waterlander
http://waterlan.home.xs4all.nl/

------------------------------------------------------------------------------

Gmane