Re: Strange fstatat / stat behavour on directories causing tar "file changed as we read it" error

2011-01-07 Thread Steven Hartland
- Original Message - From: "Larry Hall (Cygwin)" Here it takes about 2 - 5mins for what ever is causing the 0 size after a find to start to happen. Prior to that after the find all dirs show 8192 for size in an ls. Ah, that's interesting. I see no such time-lag here. Just to be 100

Re: Strange fstatat / stat behavour on directories causing tar "file changed as we read it" error

2011-01-07 Thread Larry Hall (Cygwin)
On 1/7/2011 9:39 PM, Steven Hartland wrote: Here it takes about 2 - 5mins for what ever is causing the 0 size after a find to start to happen. Prior to that after the find all dirs show 8192 for size in an ls. Ah, that's interesting. I see no such time-lag here. $ ls -l total 0 drwxr-xr-x

Re: Strange fstatat / stat behavour on directories causing tar "file changed as we read it" error

2011-01-07 Thread Steven Hartland
- Original Message - From: "Larry Hall (Cygwin)" Just NTFS I'm afraid nothing special. You can see the behaviour with ls -l as well, as you would expect. A simpler, which may help is:- ls -l drwxr-xr-x 1 test test 0 Oct 20 14:09 testdir find testdir > /dev/null # lots of files ls -l drw

Re: Strange fstatat / stat behavour on directories causing tar "file changed as we read it" error

2011-01-07 Thread Larry Hall (Cygwin)
On 1/7/2011 9:00 PM, Steven Hartland wrote: - Original Message - From: "Eric Blake" On 01/07/2011 11:21 AM, Steven Hartland wrote: What file system is this on? Someone else reported the same behavior for "Samba share on QNX through Virtual PC." - if the problem is limited to just a subse

Re: Bad rebaseall interaction with new mingw

2011-01-07 Thread Daniel Colascione
On Fri, Jan 7, 2011 at 4:25 PM, JonY wrote: > IMHO rebaseall shouldn't be interacting with mingw dlls at all. Maybe it > can check for dependencies on cygwin1.dll before rebasing? That's the point. The mingw DLLs need to be added to rebaseall's filter pattern. -- Problem reports: http://cy

Re: Strange fstatat / stat behavour on directories causing tar "file changed as we read it" error

2011-01-07 Thread Steven Hartland
- Original Message - From: "Eric Blake" On 01/07/2011 11:21 AM, Steven Hartland wrote: What file system is this on? Someone else reported the same behavior for "Samba share on QNX through Virtual PC." - if the problem is limited to just a subset of (known-buggy) file systems, it would b

Re: New version of w3m 0.5.2 Crashes - GDB help needed (SOLVED)

2011-01-07 Thread Eric Blake
On 01/07/2011 06:09 PM, nyc4...@aol.com wrote: > It appears that it crashes if LANG is not defined > > When function term.c:check_cygwin_console:239 is called and > the following is evaluated with environment variable LANG undefined: > > if (strncmp(getenv("LANG"), "ja", 2) == 0) { Yep, ge

Re: New version of w3m 0.5.2 Crashes - GDB help needed (SOLVED)

2011-01-07 Thread nyc4bos
nyc4...@aol.com writes: > Bob Heckel writes: > >> On Wed, Jan 5, 2011 at 16:22, Christopher Faylor wrote: >>> On Wed, Jan 05, 2011 at 04:16:37PM -0500, nyc4bosaol.com wrote: nyc4bosaol.com writes: Bob created a debug version of w3m for me. Here's what I see: The cra

Re: Bad rebaseall interaction with new mingw

2011-01-07 Thread JonY
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 1/8/2011 06:06, Daniel Colascione wrote: > I recently ran into this issue between the new mingw64 libraries and > rebaseall (mandatory on every Win7/Server2008R2 64-bit system I've > seen so far, alas). A patch is described at > http://blog.brev.nam

Bad rebaseall interaction with new mingw

2011-01-07 Thread Daniel Colascione
I recently ran into this issue between the new mingw64 libraries and rebaseall (mandatory on every Win7/Server2008R2 64-bit system I've seen so far, alas). A patch is described at http://blog.brev.name/2010/09/nodejs-on-windows-7-under-cygwin.html and works fine. Could it be merged? Thanks, Daniel

Re: Please don't resolve native symbolic links in cygwin_conv_path

2011-01-07 Thread Daniel Colascione
On Fri, Jan 7, 2011 at 7:29 AM, Eric Blake wrote: > On 01/06/2011 06:39 PM, Daniel Colascione wrote: >> If a POSIX path supplied to cygwin_conv_path ends in a symbolic link, >> the returned path refers to the target of that link. Normally, that's >> a good thing because native programs can't under

Re: Strange fstatat / stat behavour on directories causing tar "file changed as we read it" error

2011-01-07 Thread Eric Blake
On 01/07/2011 11:21 AM, Steven Hartland wrote: > It turns out that the value of st_size returned by a call > to fstatat on a directory can change even though there > have been no changes at all to said directory or its > children. What file system is this on? Someone else reported the same behavi

Re: skipping file `...', as it was replaced while being copied

2011-01-07 Thread Eric Blake
On 01/07/2011 01:06 PM, Nellis, Kenneth wrote: > Using Cygwin 1.7.7, the following behavior started recently, > but I can't think of any change that may be the culprit: I can. Most likely, you recently ran setup.exe and upgraded tar. Upstream tar includes a patch to make it more picky about stat

Re: skipping file `...', as it was replaced while being copied

2011-01-07 Thread Matt Seitz
Would either of the following help? -collect a network capture using Wildshark while running both the Cygwin "cp" command and the Windows Explorer "copy" operation (or a simple Windows cmd "cp" command). -use Microsoft's "Process Monitor" to collect a Win32 trace while running both the Cygwin "c

Re: What version of cygwin is considered stable on Windows 7?

2011-01-07 Thread Eric Blake
On 01/07/2011 01:12 PM, Larry W. Virden wrote: > When considering building a basically "frozen" version of cygwin - that is to > say, downloading, configuring, and building a disk image, then turning that > disk > image into a MSI for installation purposes (in an environment where this is > bei

Re: What version of cygwin is considered stable on Windows 7?

2011-01-07 Thread Larry Hall (Cygwin)
On 1/7/2011 3:12 PM, Larry W. Virden wrote: When considering building a basically "frozen" version of cygwin - that is to say, downloading, configuring, and building a disk image, then turning that disk image into a MSI for installation purposes (in an environment where this is being done because

What version of cygwin is considered stable on Windows 7?

2011-01-07 Thread Larry W. Virden
When considering building a basically "frozen" version of cygwin - that is to say, downloading, configuring, and building a disk image, then turning that disk image into a MSI for installation purposes (in an environment where this is being done because users will not have Windows 7 permissions

skipping file `...', as it was replaced while being copied

2011-01-07 Thread Nellis, Kenneth
Using Cygwin 1.7.7, the following behavior started recently, but I can't think of any change that may be the culprit: ===BEGIN=SNIPPET knel...@cobqdppj1 ~ $ cp /cygdrive/q/knellis/xyz ~ cp: skipping file `/cygdrive/q/knellis/xyz', as it was replaced while being copied knel...

Strange fstatat / stat behavour on directories causing tar "file changed as we read it" error

2011-01-07 Thread Steven Hartland
Just been debugging a very strange issue where tar was reporting "file changed as we read it" for directories which aren't actually seeing any changes. It turns out that the value of st_size returned by a call to fstatat on a directory can change even though there have been no changes at all to s

man size

2011-01-07 Thread Nellis, Kenneth
With "LANG=C.UTF-8", "man size" writes U+23AA (CURLY BRACKET EXTENSION) characters in place of vertical bars (ASCII 0x7C). Not only does this Unicode "Miscellaneous Technical" character not display properly, it also seems a poor character choice for this context. Possibly not related, after se

Re: Please don't resolve native symbolic links in cygwin_conv_path

2011-01-07 Thread Eric Blake
On 01/06/2011 06:39 PM, Daniel Colascione wrote: > If a POSIX path supplied to cygwin_conv_path ends in a symbolic link, > the returned path refers to the target of that link. Normally, that's > a good thing because native programs can't understand Cygwin links. > But this behavior is unwanted when