[dim] dim-0.5.5-mv2-pb1
dim-0.5.5-mv2-pb1 Do not complain if root node already exists when importing. Suppressed trap on EXIT in cd_import_tmpdir(), that caused import to fail, because it fired when exiting the function ! -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.5-pb2
dim-0.5.5-pb2
Symlink support limited to dim {add,del,diff}.
Index update is the only supported feature.
Work in progress - Do not use this version!
--
Archive: http://www.mail-archive.com/dim@ml.free.fr/
To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.5-pb4
dim-0.5.5-pb4
Symlink support for 'dim {diff,ldiff,wdiff}'.
Work in progress - Do not use this version!
--
Archive: http://www.mail-archive.com/dim@ml.free.fr/
To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.5-pb3
dim-0.5.5-pb3
Fixed a bug in update_sum() for symlinks.
To find regular file changes, we use the following expression:
find versiondir -type f '(' -cnewer scanfile -o -links 1 ')'
The empty file scanfile stores the time of the last scan.
The version directory is maintained by 'dim add|del'.
Files in the version directory are hard linked to the
corresponding file in the clone directory. After the user
changes the contents of a file in the clone, either
1. the hard link is preserved, and the change is detected by
-cnewer scanfile
2. or the hard link is broken, and the change is detected by
-links 1
For symlinks, we cannot use the hard link trick, because
hardlinking symlinks is not portable. We have to use an
absolute scan each time:
find versiondir -type l
Work in progress - Do not use this version!
--
Archive: http://www.mail-archive.com/dim@ml.free.fr/
To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.5-pb1
dim-0.5.5-pb1 Fixed a typo in 'dim export' for protocol rsync. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.4-mv11-pb1
dim-0.5.4-mv11-pb1 Experimental rewrite of 'dim export' for protocols file and rsync. Optimized 'dim get' and 'dim put' for degenerate library file:.dimlib. Note: this is unstable and untested code. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.4-mv8-pb1
This is a bug: under current policy, cloning into an empty directory is supposed to copy. The job/clone root directory is not seen as empty because of the presence of the metadata directory .dimlib (which should be ignored but is not). Philippe dim-ow...@ml.free.fr wrote on 11/12/2009 14:45:35: > > - "Philippe Bergheaud" a écrit : > > > dim-0.5.4-mv8-pb1 > > Fixed the bug that caused clone to fail at the root directory of a > > job. > Yes, works, but in that case the 'copy' operation is not performed > by 'clone' and must be done manually. > > Marc -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.4-mv7
Understood, it was not an infinite loop, it was the filter 'worldline'
reading from the keyboard !
Replacing 'wordline ${old|new}file' by 'wordline <${old,new}file' fixes
the new version of wd.
Philippe
dim-ow...@ml.free.fr wrote on 11/12/2009 13:18:44:
>
> - "Philippe Bergheaud" a écrit :
>
> > The new version of wd enters an infinite loop in gawk 3.1.5.
>
> I will revert this change. Does this bug happen in any case or for a
> particular data file?
>
> Marc
--
Archive: http://www.mail-archive.com/dim@ml.free.fr/
To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.4-mv8-pb1
dim-0.5.4-mv8-pb1 Fixed the bug that caused clone to fail at the root directory of a job. Philippe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.4-mv7
The new version of wd enters an infinite loop in gawk 3.1.5. Philippe dim-ow...@ml.free.fr wrote on 11/11/2009 16:57:20: > dim-0.5.4-mv7 > Determine shell executable only if dynamic patch and rerun is > required. > Replace tr(1) by a awk script in wd. This removes a dependency. > -- -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.4-mv1-pb1
dim-0.5.4-mv1-pb1 Fixed error in protocol HTTP/1.1. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv14-pb6
dim-0.5.2-mv14-pb6 'dim diff' returns 'no item' when that is the case. Suppressed an erroneous error message when unexporting with protocol file:. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv14-pb5
dim-0.5.2-mv14-pb5 Fixed a bug in file_getnewversion(), called by 'dim export' for the protocol 'file:', which caused dim to enter an infinite loop when the directory '.version' was not writable by the user. Suppressed an inappropriate error message in 'dim import'. Fixed a ksh88 'test' parsing bug in 'dim put'. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv14-pb4
dim-0.5.2-mv14-pb4 'dim copy -f' stays silent about loosing data. Fixed a typo in 'dim unexport' for protocol 'file:'. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv14-pb3
dim-0.5.2-mv14-pb3 'dim clone' populates an empty clone (with 'dim copy'). -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv14-pb2
dim-0.5.2-mv14-pb2 Renamed getnewvers() family of functions into getnewversion(). Fixed a bug in getnewversion(), which did not branch off a foreign branch. Fixed lget(), which now returns false when the attr file is empty. Fixed a bug in file_getnewversion(), where argument #2 of getneversion() was the checksum, instead of the ancestor version. Fixed noclobber pattern to suppress the error message: s:'echo >file 2>/dev/null':'(echo >file) 2>/dev/null': -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv14-pb1
dim-0.5.2-mv14-pb1 Fixed 'dim ancestor -', when n > 1. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.2-mv14
Hi Marc, 'dim get' downloads a corrupted empty file named 'dim-0.4.5-mv4--0.4.' Could you check the archive directory on server www.oligem.com, and remove the empty file if necessary ? Thank you, Philippe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv13-pb6
dim-0.5.2-mv13-pb6 'dim clone Version' returns an error message when Version does not exist. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv13-pb5
dim-0.5.2-mv13-pb5 Fixed parseopt for ksh88 (at least on AIX). -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv13-pb4
dim-0.5.2-mv13-pb4 Created command 'dim unclone', which operates on current clone. 'dim mkjob file:' creates a job which is its own file: library. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv13-pb3
dim-0.5.2-mv13-pb3 Forced Http_proto=0 (Http_proto=1 breaks 'dim get' and 'dim put'). -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.2-mv13-pb2
dim-0.5.2-mv13-pb2 Fixed a bug in update_sum(), which returned 1 instead of 0 after an empty incremental scan. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv13-pb1
dim-0.5.2-mv13-pb1 Fixed a bug in 'dim add', which could operate on $HOME rather than $PWD. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv8-pb3
dim-0.5.2-mv8-pb3 Fixed a bug in 'dim reclone', which could duplicate anc as anc2. Fixed reparenting of desc2 in 'dim unimport'. 'dim unexport' operates on the exported ancestor of current clone. This is the first step towards saving log and data of the unexported version in the current clone (after having saved it if necessary). -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv8-pb2
dim-0.5.2-mv8-pb2 'dim unexport' refuses to remove a version with an exported descendant. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.2-mv8-pb1
dim-0.5.2-mv8-pb1 Fixed a bug in 'dim unexport', which was not removing archives. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv6-pb8
dim-0.5.2-mv6-pb8 Fixed a typo in reclone(). Fixed dar(), which omitted 'anc2' and 'log' since dim-0.5.2-mv1. getiv() now defines item from $Archive/$item or .dimlib/item/$item. copy() suppresses all ancestors 2 of the current clone. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv6-pb7
dim-0.5.2-mv6-pb7 Fixed a bug in ldel(), broken since dim-0.5.2-mv6-pb4. 'dim unexport' refuses to unexport version 0. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv6-pb6
dim-0.5.2-mv6-pb6 Fixed mkitem (previous fix was wrong). -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv6-pb5
dim-0.5.2-mv6-pb5 Fixed mkitem(), broken in the previous release. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv6-pb4
dim-0.5.2-mv6-pb4 Renamed anc2() to ancestor2(). Renamed rmanc2() to delanc2(). Created new functions operating on attr[item,version], stored in $Job/.dimlib/item/$item/.dim/$version/$attr: setattr attr item version text - assign text to attr getattr attr item version - return 1st line of attr addattr attr item version list - add list of words to 1rst line of attr delattr attr item version list - delete list of words from 1rst line of attr Fixed a bug in getanc2(), which was broken since dim-0.5.1-mv4 ! -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv6-pb3
dim-0.5.2-mv6-pb3 Fixed a bug in do_export. Fixed a bug in merge2. Changed Http_proto from 1 to 0. Suppressed message "ignoring missing ancestor2". -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.2-mv6-pb2
For the record, 'dim remove' could be replaced by clone/unclone # + save/unsave # *--* import/unimport # exported version export/unexport # exported version, never replicated, no exported descendant, user == author If we forced unexport to operate on the ancestor of the current job, one could save log and data of the unexported version in the clone.. Philippe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv6-pb2
dim-0.5.2-mv6-pb2 Fixed: 'dim unexport V1' does not save history in descendants of V1. 'dim unexport V1' removes V1 from the anc2 list of each desc2 of V1. To do: 'dim unexport V1' does not yet check that V1 has no exported descendant. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv6-pb1
dim-0.5.2-mv6-pb1 s/chanc/reclone/ -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.2-mv3-pb9
I forgot to say that 'dim clone V1' now always calls 'dim add .' The previous behaviour was too dangerous: 'dim copy' without 'dim add' could remove existing files in the clone without warning. Philippe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv3-pb9
dim-0.5.2-mv3-pb9 Suppressed global variable Narg. Fixed a typo in dag(). Suppressed getitem(), replaced by getiv(). Changed getiv(), which honours N1, and returns R1=item R2=version R3=versiondir R4=clonedir R5=entry. 'getiv Path' now excludes Jobdir/.dimlib. Renamed setv1v2() into getiv2(). Renamed urlenc() into urlencode(). Changed clone: 'dim clone' Print current clone 'dim clone V1'Make current directory a clone of V1 'dim clone Item' List all clones of Item 'dim clone -a'List all clones of the current job clone() makes +/path and +/dir relative to Jobdir. Created getclonedir(), which returns the absolute clonedir. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv3-pb8
dim-0.5.2-mv3-pb8 s/umount/remove/ -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv3-pb7
dim-0.5.2-mv3-pb7 Fixed 'dim import -n' (and oversimplifying it). Always 'cd /' before attempting to remove a directory. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv3-pb6
dim-0.5.2-mv3-pb6 Suppressed 'dim mkclone'. Use 'dim clone' followed by 'dim copy'. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv3-pb5
dim-0.5.2-mv3-pb5 Dropped the mount analogy for users, back to clones and ancestors. s/mount/clone/ s/remount/chanc/ s/ on / is ancestor of / mkclone is an alias for clone+copy, nor sure if this is a good idea. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv3-pb4
dim-0.5.2-mv3-pb4 Fixed a typo in getiv (s/gawk/awk/). -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv3-pb3
dim-0.5.2-mv3-pb3 Suppressed argument Dir of 'dim remount'. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv3-pb2
dim-0.5.2-mv3-pb2 Fixed 'dim item -a' when there is no current item. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv3-pb1
dim-0.5.2-mv3-pb1 Fixed 'dim get -a', which did not work without item. Suppressed argument Dir of 'dim mount' and 'dim mkclone'. Hardened mount: mounted sub-directories are detected. Hardened getiv: . .. and / are forbidden before :. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv2-pb2
dim-0.5.2-mv2-pb2 Suppressed the argument Dir of 'dim mkjob'. Renamed flag 'dim mkjob -v' into 'dim mkjob -l'. Added new options to 'dim job': -R Recursively search sub-directories for jobs. -q Return only the first job found. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2-mv2-pb1
dim-0.5.2-mv2-pb1 Suppressed the optional Path argument of 'dim lib', 'dim job' and 'dim item'. Fixed 'dim import'. The flag -a is mandatory when there is no current item. Fixed 'dim remove'. One cannot remove the version 0 of an item. getiv() dies when there is no item (instead of returning with item=). setjob() sets the global variable Jobdir. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.2 has 36 ancestors
Did not work at the office, working fine at home (!?) Philippe dim-ow...@ml.free.fr wrote on 02/12/2009 17:08:25: > I can not reproduce your error. Works ok here. > > Marc > > Philippe Bergheaud wrote: > > I have done that, I still get the same error. > > > > dim-ow...@ml.free.fr wrote on 02/12/2009 16:59:00: > > > >> You have to remove the previous archive first, then get again: > >> > >> $ rm $(dim job)/.dimlib/archive/dim/dim-0.5.2.tar.gz > >> $ dim get > >> > >> Marc > >> > >> Philippe Bergheaud wrote: > >>> 'dim get' returns the following error message: > >>> > >>> get archive/dim/dim-0.5.2.tar.gz (45373 bytes) > >>> dim error: wrong size of dim-0.5.2.tar.gz: 45448 > >>> > >>> Philippe > >>> > >>> dim-ow...@ml.free.fr wrote on 02/12/2009 16:18:58: > >>> > >>>> I fixed and updated the archive of dim-0.5.2. Please reset your jobs > >>>> pointing on http://www.oligem.com/dim/wdim.php. > >>>> > >>>> A fix of dim tool will follow. > >>>> > >>>> Thanks, > >>>> > >>>> Marc > >>> -- > >>> Archive: http://www.mail-archive.com/dim@ml.free.fr/ > >>> To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe > >> -- > >> Archive: http://www.mail-archive.com/dim@ml.free.fr/ > >> To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe > > -- > > Archive: http://www.mail-archive.com/dim@ml.free.fr/ > > To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe > > -- > Archive: http://www.mail-archive.com/dim@ml.free.fr/ > To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.2 has 36 ancestors
I have done that, I still get the same error. dim-ow...@ml.free.fr wrote on 02/12/2009 16:59:00: > You have to remove the previous archive first, then get again: > > $ rm $(dim job)/.dimlib/archive/dim/dim-0.5.2.tar.gz > $ dim get > > Marc > > Philippe Bergheaud wrote: > > 'dim get' returns the following error message: > > > > get archive/dim/dim-0.5.2.tar.gz (45373 bytes) > > dim error: wrong size of dim-0.5.2.tar.gz: 45448 > > > > Philippe > > > > dim-ow...@ml.free.fr wrote on 02/12/2009 16:18:58: > > > >> I fixed and updated the archive of dim-0.5.2. Please reset your jobs > >> pointing on http://www.oligem.com/dim/wdim.php. > >> > >> A fix of dim tool will follow. > >> > >> Thanks, > >> > >> Marc > > -- > > Archive: http://www.mail-archive.com/dim@ml.free.fr/ > > To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe > > -- > Archive: http://www.mail-archive.com/dim@ml.free.fr/ > To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.2 has 36 ancestors
'dim get' returns the following error message: get archive/dim/dim-0.5.2.tar.gz (45373 bytes) dim error: wrong size of dim-0.5.2.tar.gz: 45448 Philippe dim-ow...@ml.free.fr wrote on 02/12/2009 16:18:58: > I fixed and updated the archive of dim-0.5.2. Please reset your jobs > pointing on http://www.oligem.com/dim/wdim.php. > > A fix of dim tool will follow. > > Thanks, > > Marc -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Fw: dim-0.5.2 has 36 ancestors
Erratum: non-exported versions are listed in .dimlib/item/dim/.dim/0.5.2/anc2, but not displayed. - Forwarded by Philippe Bergheaud/France/IBM on 02/12/2009 08:12 - dim-ow...@ml.free.fr wrote on 02/12/2009 08:04:33: > 'dim dag dim-0.5.2' displays 36 ancestors, including non-exported > versions. > > Ancillary question: is there a command to list the ancestors displayed in > red? > > Philippe > -- > Archive: http://www.mail-archive.com/dim@ml.free.fr/ > To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2 has 36 ancestors
'dim dag dim-0.5.2' displays 36 ancestors, including non-exported versions. Ancillary question: is there a command to list the ancestors displayed in red? Philippe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.2 is not imported
$ dim get dim get archive/dim/dim-0.5.2.tar.gz (45448 bytes) $ dim import dim# does not import dim-0.5.2. $ dim import dim-0.5.2 dim-0.5.2 -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.1-mv16-pb1
dim-0.5.1-mv16-pb1 Expérimental demo version: mount with a mount table. 'dim mount' does not block when mounting on a subdir of a clone. 'dim mount' unconditionally appends mount info to .dimlib/mount. 'dim ngetitem' queries .dimlib/mount to identify the current clone. Note that ngetitem is the only client of the mount table for now. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.1-mv13-pb2
dim-0.5.1-mv13-pb2 All commands but 'dim mkjob' 'dim job' and 'dim lib' operate on the job defined by the current directory, which can be changed with 'dim -D directory'. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.1-mv13-pb1
Right, the problem has nothing to do with the clone mount directory. Here is the script illustrating the problem: $ dim -V dim-0.5.1-mv13 $ cd job# this is a job $ dim mkclone dim-0.5.1-mv13 dim cd dim $ dim save -n dim-0 # should be dim-0.5.1-mv13 $ dim export -n dim-0 # should be dim-0.5.1-mv13 Philippe dim-ow...@ml.free.fr wrote on 02/08/2009 11:24:03: > Philippe Bergheaud wrote: > > > Note that R3 is broken: do not try to save your work > > from any other directory than the clone mount directory. > > Note also that this version has not been extensively tested. > > What is the exact problem? Is it in your version only? I tried to save > i.e. from dim/contrib (changed file was ../dim), and it worked normally > with dim-0.5.1-mv13. > > Marc -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.1-mv13-pb1
dim-0.5.1-mv13-pb1 It has been decided that dim commands will only operate on one job at a time, defined by the global variable Job (which strangely points to a directory called 'item'). Therefore, I have made the following changes: Suppressed getjob, replaced by setjob, which sets Job=/.dimlib/item Archive=/.dimlib/archive Url= Changed the signature of getjiv, which now returns Job=/.dimlib/item Archive=/.dimlib/archive Url= R1=item R2=version R3= R4= Note that R3 is broken: do not try to save your work from any other directory than the clone mount directory. Note also that this version has not been extensively tested. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.1-mv10-pb1
dim-0.5.1-mv10-pb1 Fixed a typo in 'dim log'. The arguments of dag, getdir, getitem and getjiv defaults to '.'. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5.1-mv10-pb1 (Warning)
The problem has been fixed. Philippe dim-ow...@ml.free.fr wrote on 02/07/2009 14:28:47: > Warning: the archive dim-0.5.1-mv10-pb1--0.5.1-mv10.tar.gz is corrupted on > www.oligem.com. > Please refrain from 'dim get'ting it until further notice. > > Philippe > -- > Archive: http://www.mail-archive.com/dim@ml.free.fr/ > To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.1-mv10-pb1 (Warning)
Warning: the archive dim-0.5.1-mv10-pb1--0.5.1-mv10.tar.gz is corrupted on www.oligem.com. Please refrain from 'dim get'ting it until further notice. Philippe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.1-mv3-pb1
dim-0.5.1-mv3-pb1 Fixed dim export -n. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5.1-pb1
dim-0.5.1-pb1 Fixed a bug in mkjob. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5-mv28-pb4
dim-0.5-mv28-pb4 First version of the internal command dim2dim usage: dim2dim Spec Translate a dim specification Spec into dim variables. A dim specification is [Jobdir]:[Item][Version], where Jobdir is a job directory (default: current job) Item is the name of an item (default: current item) Version is a version of the item (default: current version) Note: work in progress, do no use this version of dim. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5-mv28-pb3
dim-0.5-mv28-pb3 Undone changes made to 'umount' in the previous release. My attempt to fix 'umount Dir' had broken 'umount Vers'. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5-mv28-pb2
dim-0.5-mv28-pb2 Fixed clone number allocation algorithm in mount. Fixed umount when current directory is a job. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5-mv28-pb1
dim-0.5-mv28-pb1 Suppressed an error message from 'ln' in merge, when a file of the current clone is overwritten by a file of the version being merged. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5-mv25-pb2
dim-0.5-mv25-pb2 Hardened mkjob (including a simple heuristic to avoid a full find when possible). -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5-mv25-pb2
dim-0.5-mv25-pb2 Hardened mkjob (including a simple heuristic to avoid a full find when possible). -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5-mv25-pb1
dim-0.5-mv25-pb1 'history' is an builtin alias in ksh88, unalias it. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5-mv20-pb2
dim-0.5-mv20-pb2 delta now works with AIX cpio, which accepts 'cpio -pdul', but rejects 'cpio -pdu -l'. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5-mv20-pb1
dim-0.5-mv20-pb1 Ash ldiff and wdiff work again. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5-mv14-pb3
dim-0.5-mv14-pb3 Renamed experimental command path2dim into unix2dim. Added support for clones mounted on the job directory itself. Fixed a few scoping anomalies. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5-mv14-pb2
dim-ow...@ml.free.fr wrote on 01/18/2009 21:23:57: > Version not in repository. 'dim put' missing ? Done (3 times yesterday evening), once this morning (?!). Silence yesterday (neither error message, nor success message). Success message this morning. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5-mv14-pb2
dim-0.5-mv14-pb2 This is an experimental rewrite of the getvers family of function. I have created a test command called path2dim, which can be called from the command line. It translates its (optional) path argument into the following dim variables: jobdir # job path (unix) item # item name (dim) version # item version name (dim) versiondir # item version path in library (unix) clone# item clone number (dim) clonedir # item clone path in job (unix) relpath # path of argument relative to # versiondir or clonedir (unix) A job is a unix directory with a job tag (currently .dimlib). A item is a unix directory with an item tag (currently .dim). A version is a direct subdirectory of an item. A version named +n is associated to the clone number n. A clone is a subdirectory of the job. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] Re: dim-0.5-mv14-pb1
dim-0.5-mv14-pb1 'dim item Path' works when Path points out of the current job. 'dim job Path' and 'dim lib Path' also work in that case. Note 'dim version Path' does not yet work in that case. I have also made a few cosmetics changes. In particular, I propose that a library should *serve* jobs, rather than being linked to, or associated with, or mounted on, them. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] dim-0.5-mv13-pb1
dim-0.5-mv13-pb1 mount options auto/noauto have been suppressed. mount is now always noauto, i.e. users must run add/del. add/del messages have been made more informative and consistent. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:dim-requ...@ml.free.fr?subject=unsubscribe
[dim] wd-0.6 is broken
wd-0.6 (shipped with dim-0.4.8-mv8) does not colorize big (greater than a screen) blocks of deleted lines. It seems that (at least some of the) suppressed endmark/startmark pairs were not as redundant as thought. Philippe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] Version specification (again)
Inside a dim library namespace, an item version is
Item:Version
Item is
[A-Za-z0-9_][A-Za_z0-9_.]*
Version is one of
[A-Fa-f0-9]{40} # sha1 checksum
+[1-9][0-9]*# clone number
Name[(-Name)+[(--Name)+]]
Name is
[A-Za-z0-9_][A-Za_z0-9_.]*
The syntax Item:Version can be relaxed into one of
Item-Version# used for tarballs
Item
Version
A file or directory Pathname is one of
\.[^\n]*# starts with .
[^.\n]*[^-+A-Za-z0-9_.\n][^\n]*
Pathname syntax can be relaxed into whatever
[\n]+
Ambiguity will be lifted by a heuristic not yet documented.
Philippe
[dim] Re: Version specification
Philippe Bergheaud wrote: Inconsistencies: Library cannot be specified for each argument. Item cannot be specified for each sha1 checksum argument. Version cannot be specified with -V Version. Ambiguities: Relative Unix path without '.' or '/' looks like Item or Version. Main node Version without '.' looks like Item. Idea #1: The syntax [Library]:Item would help. One should also be able to alias Library to a local nickname. In fact, enforcing ':' in all dim namespace specifications really help. :Item, :Version, :Item-Version would discriminate from Unix pathname. This convention would add a new (but decent) restriction: Relative Unix path containing ':' but no '/' would have to be prefixed with './'. Note: getopts(3) also demands '--' or './' prefixing for Unix path starting with '-' or '+'. Idea #2: :[Item]-Version would discriminate Item from Version, Version from flag. :a # Item :a1# Item :-a# Version (assuming the last digit constraint has been lifted) :-a1 # Version :+1# Version -a # flag -a1 # flags (-a -1) +1 # flag a # Relative pathname a1 # Relative pathname 1 # Relative pathname This convention would ease the integration with other tools (expecting standard Unix flag and pathname conventions). Suppressing ambiguities would also simplify argument parsing. Philippe -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] Version specification
A item version is defined by (Library, Item, Version) in the dim namespace. Library is an absolute Unix directory path. Library can be defined by '-L Library' or 'export DIMLIB=Library'. Item is a sequence of letters, digits and underscores. Item can be defined by '-I Item' or specified for each argument. Version is a a sha1 checksum (a sequence of 40 hex digits) or list of node names (Main node, Branch node, ...). Version can be specified for each argument, specified or modified by various command flags. Branch node is a sequence of letters, digits, underscores and dots. A dot can only be used between two digits. Main nodes are: Public, created by 'dim export', prefixed by '-', starting with a digit. Branch nodes can be: Public, created by 'dim export', prefixed by '-', starting with a letter. Private, created by 'dim save', prefixed by '--', starting with a letter. Volatile, created by 'dim mount', prefixed by '+', without letters or underscores. An existing item version can be unambiguously specified by the Unix path of one of its file or directory (at the cost of using '.'n '..' or '/'). The argument of dim commands operating one item version typically defaults to '.'. On the other hand, the current command line syntax used by dim to specify an item version in the dim namespace is inconsistent and ambiguous. There are also some restrictions. Inconsistencies: Library cannot be specified for each argument. Item cannot be specified for each sha1 checksum argument. Version cannot be specified with -V Version. Ambiguities: Relative Unix path without '.' or '/' looks like Item or Version. Main node Version without '.' looks like Item. Restrictions: Item or Main node Version cannot be 40 hex digits. Public node Version must end with a digit. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] dim-0.4.8-mv6-pb9
dim-0.4.8-mv6-pb9 'dim add|del' honours mount points: arguments can belong to arbitrary items. Know bugs: 'dim umount' ignores client and suppliers. Verbosity control has gone out of control. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] dim-0.4.8-mv6-pb8
dim-0.4.8-mv6-pb8 'dim client|supplier' honour flag -a. 'dim add|del' use portable find(1) construct -path. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] dim-0.4.8-mv6-pb7
dim-0.4.8-mv6-pb7 Fixed bug: 'dim mount' updates client and honours suppliers. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] dim-0.4.8-mv6-pb6
dim-0.4.8-mv6-pb6 Suppressed mount table, along with mget, mset and munset. New command: 'dim client'. Know bugs: 'dim mount' updates client, but ignores suppliers. 'dim umount' ignores client and suppliers. 'dim client|supplier' are limited to clones. 'dim save|export' do not create the mount table. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] dim-0.4.8-mv6-pb5
dim-0.4.8-mv6-pb5 Changed mount file format to allow blanks in mount point pathnames. This format change only impacts recursive mounts. New internal functions: mset file key data add (key, data) pair to file mget file print all keys from file mget file key print data of (key, data) pair from file munset file key delete (key, data) pair from file -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] Re: dim-0.4.8-mv6-pb4
dim-0.4.8-mv6-pb4 'dim del' does not exclude mount points - useless in the index tree. Internal function getvers() now always sets the global variable item. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] dim-0.4.8-mv6-pb3
dim-0.4.8-mv6-pb3 'dim add|del' now exclude the mount points of the suppliers. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] dim-0.4.8-mv6-pb2
dim-0.4.8-mv6-pb2 Simplified, robustified (and unified again) 'dim add' and 'dim del'.
[dim] dim-0.4.8-mv6-pb1
Do not use this version (work in progress). dim-0.4.8-mv6-pb1 Merged dim-0.4.8-mv5-pb1. Bug fixes: dim merge removes the scan file. dim umount updates desc2 metadata. Know bugs: dim remove should reparent desc2. Philippe
[dim] dim-0.4.8-mv5-pb1
dim-0.4.8-mv5-pb1 Bug fixes: dim del operates on the index tree. dim index updates the index file before using it. dim mount checks that mountpoint is not already mounted. dim mount checks that item is not already mounted in clone. Functional changes: New syntax dim add . (replaces dim add '*'). dim add|del|mount support empty directories. dim add ignores scan file (like dim del|mount). Restrictions: Do not use this version. Recursive mounts have not been tested. Mount options other than noauto have not been tested. Dim commands other than add del mount umount have not been tested. Known bugs: dim del does not handle parent directories of deleted entries. When empty, those are incorrectly indexed as empty directories, instead of being deleted. Philippe.
[dim] Re: dim-0.4.8-mv3
Marc, did you forget to put it on www.oligem.com? Philippe
[dim] dvs2dim
dim-0.4.8-pb2 Added contrib/dvs2dim. This program creates a dim replica of a dvs item. -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] dim-0.4.7-mv4 export bug
$ dim mklib file:/home/felix/selflib /home/felix/selflib $ export DIMLIB=/home/felix/selflib $ dim mkuser $ dim mkitem rr $ dim mount rr . $ dim export -n rr-0-pb1# good $ dim export -n -M rr-1# good $ dim export -n -M -B 37 rr-37 # good $ dim export -n -B 37 rr-0--37# bad
[dim] Re: dim-0.4.7-mv2-pb1
'dim mount' should be quiet by default. Do not know how to code this change. Philippe Bergheaud wrote: dim-0.4.7-mv2-pb1 Bugfix: the existence of the sum file also required for update_sum(). -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED] -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] dim-0.4.7-mv2-pb1
dim-0.4.7-mv2-pb1 Bugfix: the existence of the sum file also required for update_sum(). -- Archive: http://www.mail-archive.com/dim@ml.free.fr/ To unsubscribe: mailto:[EMAIL PROTECTED]
[dim] Re: New spec proposal for mount index update
Attempt #2, for 'dim mount' only. dim mount item|version dir Link all files from the mount point to the clone index. Options persist until 'dim umount'. Options can be changed with 'dim remount'. OPTIONS -a | -o auto Automatically update the index to reflect the mount point contents. -o noauto Do not automatically update the index to reflect the mount point contents (default). -o cvsignore Honour .cvsignore files when automatically updating the index. -o nocvsignore Do not honour .cvsignore files (default). -o gitignore Honour .gitignore files when automatically updating the index. -o nogitignore Do not honour .gitignore files (default).
[dim] New spec proposal for mount index update
dim [re]mount item|version dir Link all files from the mountpoint to the clone index. OPTIONS -a | -o auto Automatically update the index to reflect the mount point contents. -o noauto Do not automatically update the index to reflect the mount point contents (default). -g | -o gitignore Honour .gitignore files when automatically updating the index. -o nogitignore Do not honour .gitignore files (default).
[dim] dim-0.4.6-mv9 export bugs
Tested with dim-0.4.6-mv9. $ dim lib -l http://host/dim/wdim.php on /tmp/dimlib.http $ dim version -a linux-0 $ dim export -q -M -B 2.6.10 -n linux-2.6.10# good $ dim export -q -M -B 2.6.10 new anc of +1: 2.6.10 # good linux-2.6.10# good $ dim lib -l file:/tmp/nsdir on /tmp/dimlib.file $ dim version -a linux-0 $ dim export -q -M -B 2.6.10 -n linux-2.6.10 $ dim export -q -M -B 2.6.10 new anc of +1: 2.6.10 # good linux-2.6.10# good $ dim lib -l file:/tmp/dimlib.self on /tmp/dimlib.self $ dim version -a linux-0 $ dim export -q -M -B 2.6.10 -n linux-2.6.10# good $ dim export -q -M -B 2.6.10 linux-0--1 # bad $ dim lib -l rsync:[EMAIL PROTECTED]:dimlib on /tmp/dimlib.rsync $ dim version -a linux-0 $ dim export -q -M -B 2.6.10 -n linux-2.6.10# good $ dim export -q -M -B 2.6.10 new anc of +1: 2.6.9.1 # bad linux-2.6.9.1 # bad Philippe
[dim] Re: dim-0.4.6-mv5-pb1
dim-0.4.6-mv5-pb1
First implementation of clones based on an index tree in the library.
dim {add,list,del}' not yet coded (should operate on the index tree).
To use this incompatible and experimental version, users have to
umount
and remount all their clones. Note: dim {add,list,del} not yet coded.
Philippe
