In the past there's been confusion as to what the release packages contain with regards to EOLs <https://svn.boost.org/trac/boost/ticket/2641>. My idea for dealing with the confusion is to tag the packages with an EOL indicator (as explained in the bug). Any other ideas? -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org (msn) - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim,yahoo,skype,efnet,gmail
On Saturday 14 February 2009 15:04, Rene Rivera wrote:
In the past there's been confusion as to what the release packages contain with regards to EOLs <https://svn.boost.org/trac/boost/ticket/2641>. My idea for dealing with the confusion is to tag the packages with an EOL indicator (as explained in the bug). Any other ideas?
Couldn't you just let svn use the svn:eol-style to get the end-of-line characters right on files that are specifically for unix (for example the configure script mentioned in the bug report) in all packages? I checked in svn, and the configure script does have its svn:eol-style set properly to LF.
Frank Mori Hess wrote:
On Saturday 14 February 2009 15:04, Rene Rivera wrote:
In the past there's been confusion as to what the release packages contain with regards to EOLs <https://svn.boost.org/trac/boost/ticket/2641>. My idea for dealing with the confusion is to tag the packages with an EOL indicator (as explained in the bug). Any other ideas?
Couldn't you just let svn use the svn:eol-style to get the end-of-line characters right on files that are specifically for unix (for example the configure script mentioned in the bug report) in all packages? I checked in svn, and the configure script does have its svn:eol-style set properly to LF.
That wouldn't solve the problem for environments/toolsets that can use the configure, Makefiles, etc. in CRLF form. In particular it would cause grief for MinGW and Cygwin/CRLF users (I'm one of those). -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org (msn) - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim,yahoo,skype,efnet,gmail
On Saturday 14 February 2009 15:30, Rene Rivera wrote:
Couldn't you just let svn use the svn:eol-style to get the end-of-line characters right on files that are specifically for unix (for example the configure script mentioned in the bug report) in all packages? I checked in svn, and the configure script does have its svn:eol-style set properly to LF.
That wouldn't solve the problem for environments/toolsets that can use the configure, Makefiles, etc. in CRLF form. In particular it would cause grief for MinGW and Cygwin/CRLF users (I'm one of those).
Oh, in that case should the svn:eol-style for the configure script, etc be corrected to "native"?
Frank Mori Hess wrote:
Couldn't you just let svn use the svn:eol-style to get the end-of-line characters right on files that are specifically for unix (for example the configure script mentioned in the bug report) in all packages? I checked in svn, and the configure script does have its svn:eol-style set properly to LF.
I set svn:eol-style for shell scripts in 1.38. This issue had come up several times and everyone seemed to think it was a good idea. 2009/2/14 Rene Rivera <grafikrobot@gmail.com>:
That wouldn't solve the problem for environments/toolsets that can use the configure, Makefiles, etc. in CRLF form. In particular it would cause grief for MinGW and Cygwin/CRLF users (I'm one of those).
How does it cause grief? Daniel
Daniel James wrote:
Frank Mori Hess wrote:
Couldn't you just let svn use the svn:eol-style to get the end-of-line characters right on files that are specifically for unix (for example the configure script mentioned in the bug report) in all packages? I checked in svn, and the configure script does have its svn:eol-style set properly to LF.
I set svn:eol-style for shell scripts in 1.38. This issue had come up several times and everyone seemed to think it was a good idea.
2009/2/14 Rene Rivera <grafikrobot@gmail.com>:
That wouldn't solve the problem for environments/toolsets that can use the configure, Makefiles, etc. in CRLF form. In particular it would cause grief for MinGW and Cygwin/CRLF users (I'm one of those).
How does it cause grief?
If you try to run an LF shell script on either of those it gives errors, or does nothing, because it looks like a single long line to them. -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org (msn) - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim,yahoo,skype,efnet,gmail
Rene Rivera wrote:
Daniel James wrote:
for MinGW and Cygwin/CRLF users (I'm one of those).
How does it cause grief?
If you try to run an LF shell script on either of those it gives errors, or does nothing, because it looks like a single long line to them.
Scratch that... It seems MSYS dynamically adjusts to the EOL of the file itself. -- -- Grafik - Don't Assume Anything -- Redshift Software, Inc. - http://redshift-software.com -- rrivera/acm.org (msn) - grafik/redshift-software.com -- 102708583/icq - grafikrobot/aim,yahoo,skype,efnet,gmail
participants (3)
-
Daniel James -
Frank Mori Hess -
Rene Rivera