Re: Tests 1060, 1061 fail

[Spork Schivago]

Daniel,

I think that's probably one of the most important things about programming.
  Making the code as easy to read and understand as possible.   I tried
coming up with a solution using a few different styles of loops, to see if
I could do away with your goto (after I sent in my first message) but I
couldn't find away to do it that was nice either.   Granted, I'm not nearly
as good as you guys at writing code and I haven't really written serious
code in years.

Anyway, thanks for sharing!

On Mon, Jan 9, 2017 at 5:44 PM, Daniel Stenberg  wrote:

> On Mon, 9 Jan 2017, Spork Schivago wrote:
>
> That's what I thought you were doing originally.  Using the goto just as a
>> simple fix, to see if it'd correct the problem.  But then I thought I saw a
>> post where you were going to commit it to the main branch or something
>> along those lines.
>>
>
> In the end I went with the goto solution anyway and I've now pushed that
> to git. Changing it to a while loop turned out to end up with a lot more
> code that certainly was not easier to follow or understand than the goto
> approach...
>
>
> --
>
>  / daniel.haxx.se
> ---
> List admin: https://cool.haxx.se/list/listinfo/curl-library
> Etiquette:  https://curl.haxx.se/mail/etiquette.html
>
---
List admin: https://cool.haxx.se/list/listinfo/curl-library
Etiquette:  https://curl.haxx.se/mail/etiquette.html

Re: Removing */Makefile.vc[num] !

[Daniel Stenberg]

On Mon, 9 Jan 2017, Daniel Stenberg wrote:

I'd like to once again propose that we remove these makefiles in order to 
simplify and confuse our users less.


I'm attaching a patch showing what I'd like to do:

 Makefile.dist| 449 +--
 lib/Makefile.am  |   4 +-
 lib/Makefile.vc6 | 691 

 maketgz  |  10 +-
 src/Makefile.am  |   4 +-
 src/Makefile.vc6 | 550 
-
 6 files changed, 12 insertions(+), 1696 deletions(-)

--

 / daniel.haxx.seFrom 8d98083b9f98b6788dbbea7b6b62bec4fc1d02c6 Mon Sep 17 00:00:00 2001
From: Daniel Stenberg 
Date: Tue, 10 Jan 2017 00:29:14 +0100
Subject: [PATCH] VC: remove the makefile.vc6 build infra

All hail winbuild/
---
 Makefile.dist| 449 +---
 lib/Makefile.am  |   4 +-
 lib/Makefile.vc6 | 691 ---
 maketgz  |  10 +-
 src/Makefile.am  |   4 +-
 src/Makefile.vc6 | 550 ---
 6 files changed, 12 insertions(+), 1696 deletions(-)
 delete mode 100644 lib/Makefile.vc6
 delete mode 100644 src/Makefile.vc6

diff --git a/Makefile.dist b/Makefile.dist
index 329e4f3a8..8577c8a14 100644
--- a/Makefile.dist
+++ b/Makefile.dist
@@ -3,11 +3,11 @@
 #  Project ___| | | |  _ \| |
 # / __| | | | |_) | |
 #| (__| |_| |  _ <| |___
 # \___|\___/|_| \_\_|
 #
-# Copyright (C) 1998 - 2015, Daniel Stenberg, , et al.
+# Copyright (C) 1998 - 2017, Daniel Stenberg, , et al.
 #
 # This software is licensed as described in the file COPYING, which
 # you should have received as part of this distribution. The terms
 # are also available at https://curl.haxx.se/docs/copyright.html.
 #
@@ -18,12 +18,10 @@
 # This software is distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY
 # KIND, either express or implied.
 #
 ###
 
-VC=vc6
-
 all:
 	./configure
 	make
 
 ssl:
@@ -85,370 +83,17 @@ mingw32-examples%:
 
 mingw32%:
 	$(MAKE) -C lib -f Makefile.m32 CFG=$@
 	$(MAKE) -C src -f Makefile.m32 CFG=$@
 
-vc-clean: $(VC)
-	cd lib
-	nmake -f Makefile.$(VC) clean
-	cd ..\src
-	nmake -f Makefile.$(VC) clean
-
-vc-all: $(VC)
-	cd lib
-	nmake -f Makefile.$(VC) cfg=release
-	nmake -f Makefile.$(VC) cfg=release-ssl
-	nmake -f Makefile.$(VC) cfg=release-zlib
-	nmake -f Makefile.$(VC) cfg=release-ssl-zlib
-	nmake -f Makefile.$(VC) cfg=release-ssl-dll
-	nmake -f Makefile.$(VC) cfg=release-zlib-dll
-	nmake -f Makefile.$(VC) cfg=release-ssl-dll-zlib-dll
-	nmake -f Makefile.$(VC) cfg=release-dll
-	nmake -f Makefile.$(VC) cfg=release-dll-ssl-dll
-	nmake -f Makefile.$(VC) cfg=release-dll-zlib-dll
-	nmake -f Makefile.$(VC) cfg=release-dll-ssl-dll-zlib-dll
-	nmake -f Makefile.$(VC) cfg=debug
-	nmake -f Makefile.$(VC) cfg=debug-ssl
-	nmake -f Makefile.$(VC) cfg=debug-zlib
-	nmake -f Makefile.$(VC) cfg=debug-ssl-zlib
-	nmake -f Makefile.$(VC) cfg=debug-ssl-dll
-	nmake -f Makefile.$(VC) cfg=debug-zlib-dll
-	nmake -f Makefile.$(VC) cfg=debug-ssl-dll-zlib-dll
-	nmake -f Makefile.$(VC) cfg=debug-dll
-	nmake -f Makefile.$(VC) cfg=debug-dll-ssl-dll
-	nmake -f Makefile.$(VC) cfg=debug-dll-zlib-dll
-	nmake -f Makefile.$(VC) cfg=debug-dll-ssl-dll-zlib-dll
-
-vc: $(VC)
-	cd lib
-	nmake /f Makefile.$(VC) cfg=release
-	cd ..\src
-	nmake /f Makefile.$(VC)
-
-vc-x64: $(VC)
-	cd lib
-	nmake /f Makefile.$(VC) MACHINE=x64 cfg=release
-	cd ..\src
-	nmake /f Makefile.$(VC) MACHINE=x64 cfg=release
-
-vc-sspi: $(VC)
-	cd lib
-	nmake /f Makefile.$(VC) cfg=release WINDOWS_SSPI=1
-	cd ..\src
-	nmake /f Makefile.$(VC) cfg=release WINDOWS_SSPI=1
-
-vc-x64-sspi: $(VC)
-	cd lib
-	nmake /f Makefile.$(VC) MACHINE=x64 cfg=release WINDOWS_SSPI=1
-	cd ..\src
-	nmake /f Makefile.$(VC) MACHINE=x64 cfg=release WINDOWS_SSPI=1
-
-vc-idn: $(VC)
-	cd lib
-	nmake /f Makefile.$(VC) cfg=release USE_IDN=1
-	cd ..\src
-	nmake /f Makefile.$(VC) cfg=release USE_IDN=1
-
-vc-x64-idn: $(VC)
-	cd lib
-	nmake /f Makefile.$(VC) MACHINE=x64 cfg=release USE_IDN=1
-	cd ..\src
-	nmake /f Makefile.$(VC) MACHINE=x64 cfg=release USE_IDN=1
-
-vc-idn-sspi: $(VC)
-	cd lib
-	nmake /f Makefile.$(VC) cfg=release USE_IDN=1 WINDOWS_SSPI=1
-	cd ..\src
-	nmake /f Makefile.$(VC) cfg=release USE_IDN=1 WINDOWS_SSPI=1
-
-vc-x64-idn-sspi: $(VC)
-	cd lib
-	nmake /f Makefile.$(VC) MACHINE=x64 cfg=release USE_IDN=1 WINDOWS_SSPI=1
-	cd ..\src
-	nmake /f Makefile.$(VC) MACHINE=x64 cfg=release USE_IDN=1 WINDOWS_SSPI=1
-
-vc-zlib: $(VC)
-	cd lib
-	nmake /f Makefile.$(VC) cfg=release-zlib
-	cd ..\src
-	nmake /f Makefile.$(VC) cfg=release-zlib
-
-vc-x64-zlib: $(VC)
-	cd lib
-	nmake /f Makefile.$(VC) MACHINE=x64 cfg=release-zlib
-	cd ..\src
-	nmake /f Makefile.$(VC) MACHINE=x64 

Re: Tests 1060, 1061 fail

[dev_user]

On 01/09/2017 10:44 PM, Daniel Stenberg wrote:

On Mon, 9 Jan 2017, Spork Schivago wrote:


That's what I thought you were doing originally.  Using the goto just
as a simple fix, to see if it'd correct the problem.  But then I
thought I saw a post where you were going to commit it to the main
branch or something along those lines.


In the end I went with the goto solution anyway and I've now pushed that
to git. Changing it to a while loop turned out to end up with a lot more
code that certainly was not easier to follow or understand than the goto
approach...



I looked at it .. the goto works.  Fixed.  Those tests are sort of odd
 anyways.


dc
---
List admin: https://cool.haxx.se/list/listinfo/curl-library
Etiquette:  https://curl.haxx.se/mail/etiquette.html

Re: Tests 1060, 1061 fail

[Daniel Stenberg]

On Mon, 9 Jan 2017, Spork Schivago wrote:

That's what I thought you were doing originally.  Using the goto just as a 
simple fix, to see if it'd correct the problem.  But then I thought I saw a 
post where you were going to commit it to the main branch or something along 
those lines.


In the end I went with the goto solution anyway and I've now pushed that to 
git. Changing it to a while loop turned out to end up with a lot more code 
that certainly was not easier to follow or understand than the goto 
approach...


--

 / daniel.haxx.se
---
List admin: https://cool.haxx.se/list/listinfo/curl-library
Etiquette:  https://curl.haxx.se/mail/etiquette.html

Re: Tests 1060, 1061 fail

[Spork Schivago]

Daniel,

That's what I thought you were doing originally.   Using the goto just as a
simple fix, to see if it'd correct the problem.   But then I thought I saw
a post where you were going to commit it to the main branch or something
along those lines.   That caught my eye for some reason and reminded me of
my college days.   Thank you for taking the time to answer my questions
about this.   On some mailing lists, I'm sure there'd have been some real
negative comments.   I've always felt welcomed here.  Although I don't
generally post to the list, I do read all the emails and I learn from a lot
of them.

Thanks!

On Sun, Jan 8, 2017 at 5:54 PM, Daniel Stenberg  wrote:

> On Sun, 8 Jan 2017, Spork Schivago wrote:
>
> I'm glad you responded to me and told me they're okay in a situation like
>> this.  Thank you for taking the time to educate me a little.  It's much
>> appreciated.
>>
>
> I'm generally fine with gotos when used with care and the alternatives
> turn into worse solutions.
>
> In this particular case however, I used the goto only to make a first
> quick test patch to make sure this was the reason for the discussed
> problem. I'll probably land the fix with a do-while loop instead.
>
> --
>
>  / daniel.haxx.se
>
> ---
> List admin: https://cool.haxx.se/list/listinfo/curl-library
> Etiquette:  https://curl.haxx.se/mail/etiquette.html
>
---
List admin: https://cool.haxx.se/list/listinfo/curl-library
Etiquette:  https://curl.haxx.se/mail/etiquette.html

Removing */Makefile.vc[num] !

[Daniel Stenberg]

Hey all,

I'd like to once again propose that we remove these makefiles in order to 
simplify and confuse our users less.


We added winbuild/* in 2011 and that seems like better infrastructure to do VC 
makefiles. It is about time we ditched the older setup.


Previously, it has been mentioned that we have autobuilds using the old 
makefiles, but with some combined efforts I think we should be able to migrate 
those.


Thoughts? What else will break? Anyone thinks they can't move, if so why not?

--

 / daniel.haxx.se
---
List admin: https://cool.haxx.se/list/listinfo/curl-library
Etiquette:  https://curl.haxx.se/mail/etiquette.html