Mail Archives: cygwin/2004/02/04/17:58:16
First, emacs:
On Tue, 3 Feb 2004, Thomas L Roche wrote:
>> [snip]
>> Note that I run emacs via
>> emacs --debug-init &
>> from bash in an xterm which I launch with
>> start XWin -multiwindow -clipboard -emulate3buttons
Igor Pechtchanski 02/04/2004 04:21:29 PM (rearranged)
> Not quite. Judging from your cygcheck output, you use startxwin.bat.
Yes, quite :-) My startxwin.bat does
start XWin -multiwindow -clipboard -emulate3buttons
But I should probably not edit the default, and instead copy/mod my
own. Especially now that I know to _also_
>> [snip]
>> (Note that cygcheck shows
>> d:\cygwin\bin
>> d:\cygwin\usr\X11R6\bin
>> at the head of my PATH; those dirs don't exist, and they're not in my
>> windows path.)
> This (and the existence of CYGWIN_ROOT in your environment) is an
> artifact of using startxwin.bat. Unless you have a default directory
> for the Cygwin root, you can't use that file OOTB -- you'll have to
> edit it and replace "\cygwin" with the value of your Cygwin root (in
> your case, "d:\ProgramFiles\Cygwin".
Doh! Will do, thanks for the advice.
>> [snip]
>> However I'm also, again only since upgrading, having problems with
>> unzip, e.g.
>>> inflating:
/d/eclipse/builds/20040121_1953-WB213-AD-V512D-00/eclipse/plugins/com.ibm.etools.j2ee.ui/build/pluginBuild/commonbld/com/ibm/etools/ejb/ui/presentation/EJBEditorOutlineContentProvider.class
>>> 735 [main] unzip 1820 cmalloc: cmalloc returned NULL
>>> error: zipfile probably corrupt (segmentation violation)
>> which I get even unzip'ing a known-good file (i.e. one that I unzip'ed
>> successfully before upgrading from 1.5.5-1).
> Do you have a small testcase that would allow others to reproduce
> this? Perhaps a (small) sample .zip file?
Alas, no: the smallest I have is 1.3GB.
> I'd suspect the unzip issue has to do with the path length to the
> file. You can try confirming it by creating a file in a deeply
> nested directory, putting that one file in a .zip, and then trying
> to unzip it into some place even deeper in the directory hierarchy.
OK, will do.
>> [snip]
>> So it seems plausible that there is also a (base) cygwin problem, which
>> might be causing the emacs and unzip problems. Or that the emacs and
>> unzip problems are unrelated.
> They most likely are unrelated.
Hmm: note that
* both (emacs and unzip) problems occurred after upgrading cygwin.
Your response of "post hoc, non ergo propter hoc" is understandable :-)
_But_
* unzip appears to be unchanged: it reports its version as "UnZip 5.50
of 17 February 2002" (guess I should look @ its cygwin changelog,
though)
>>> inflating:
/d/eclipse/builds/20040121_1953-WB213-AD-V512D-00/eclipse/plugins/com.ibm.etools.j2ee.ui/build/pluginBuild/commonbld/com/ibm/etools/ejb/ui/presentation/EJBEditorOutlineContentProvider.class
>>> 735 [main] unzip 1820 cmalloc: cmalloc returned NULL
>>> error: zipfile probably corrupt (segmentation violation)
* doesn't "cmalloc returned NULL" sound pretty base-ic? I.e. like
something happening underneath unzip?
* emacs is also (after upgrading) plagued by segfaults
But I will attempt to
* test your path-length hypothesis, and acquire more data regarding it
* make a personal copy of startxwin.bat, and fix it
* get IBM to cough up the $ for a support contract :-)
--
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
Problem reports: http://cygwin.com/problems.html
Documentation: http://cygwin.com/docs.html
FAQ: http://cygwin.com/faq/
- Raw text -