Re: [Barry-devel] perl contributions, was: GUI on gentoo
On Fri, Jul 04, 2008 at 11:40:53PM -0500, ashley willis wrote: you're welcome. i escaped the binary strings (without testing *all* my changes) as well as modified barrybkup to only attempt to tar directories with data. i also added a -p option to the relevant programs to select the device when more than one is installed, but i can't test it as i only have one device. i've tested barrybkup. comparing the new output to the old ouput, but i haven't tested the changes to the other programs (same changes, copy and paste, *should* work). Thanks for these changes. I took a quick look at the package, and noticed that now the -p option is required, otherwise you end up with an empty -p command line option, if I'm reading the code right. The -p option should only be added if the user specifies it. one odd thing i've noticed: the record type on *all* databases is usually 08 00, but sometimes they are *all* 07 00 (while in your file bb_task_format.txt you list 09 00). any idea what this is about? The rec type in bb_task_format.txt is mis-named. That is actually what I call the socket number, and can change at the whim of the device. It is usually consistent, but the device tells btool what number to use, so it could change for some reason internal to the device. The socket number is merely part of the communications protocol, and should not be saved in the backup. The bb_task_format.txt file was a patch sent before tasks parsing was supported, as I recall, and it is probably out of date. The best place to find the format of task records would be in the r_task.h and r_task.cc files. (sorry, it's code, not english) :-) I took another look at btool's DataDumpParser, and it dumps the whole response packet to stdout, instead of just the data portion. This probably makes the work harder for you, when the offset to the actual data is sitting right there in the ParseFields() arguments. It wouldn't be hard to display the offset too, but I suspect that would break your script. - Chris - Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! Studies have shown that voting for your favorite open source project, along with a healthy diet, reduces your potential for chronic lameness and boredom. Vote Now at http://www.sourceforge.net/community/cca08 ___ Barry-devel mailing list Barry-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/barry-devel
Re: [Barry-devel] perl contributions, was: GUI on gentoo
On Sat, 5 Jul 2008, Chris Frey wrote: Thanks for these changes. I took a quick look at the package, and noticed that now the -p option is required, otherwise you end up with an empty -p command line option, if I'm reading the code right. The -p option should only be added if the user specifies it. actually, it should otherwise probe for the pin if the -p option is not given, but this only really needs to be done in barrybkup, not the other programs. and is the pin always an 8 digit hex number (or is that just my pin)? if so, i could add some error-checking. i'd also like to see the output of just btool when 2 and 3 devices are plugged in. one odd thing i've noticed: the record type on *all* databases is usually 08 00, but sometimes they are *all* 07 00 (while in your file bb_task_format.txt you list 09 00). any idea what this is about? The rec type in bb_task_format.txt is mis-named. That is actually what I call the socket number, and can change at the whim of the device. It is usually consistent, but the device tells btool what number to use, so it could change for some reason internal to the device. The socket number is merely part of the communications protocol, and should not be saved in the backup. The bb_task_format.txt file was a patch sent before tasks parsing was supported, as I recall, and it is probably out of date. The best place to find the format of task records would be in the r_task.h and r_task.cc files. (sorry, it's code, not english) :-) thanks for explaining. I took another look at btool's DataDumpParser, and it dumps the whole response packet to stdout, instead of just the data portion. This probably makes the work harder for you, when the offset to the actual data is sitting right there in the ParseFields() arguments. It wouldn't be hard to display the offset too, but I suspect that would break your script. probably. but i'm not following what you're saying. if i print the offset, i always get 'f', which means nothing to me however i look at it. i mentioned in another email that i've written a significantly faster version, but it's not quite ready for distribution, and requires modification of data.cc and data.h (to display binary instead of hex dumps). -ashley -- No one can make you feel inferior without your consent. Eleanor Roosevelt - Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! Studies have shown that voting for your favorite open source project, along with a healthy diet, reduces your potential for chronic lameness and boredom. Vote Now at http://www.sourceforge.net/community/cca08 ___ Barry-devel mailing list Barry-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/barry-devel
Re: [Barry-devel] perl contributions, was: GUI on gentoo
On Sat, Jun 28, 2008 at 09:15:27PM -0500, ashley willis wrote: if anyone has downloaded this already, download it again. i made a non-destuctive patch which adds an -n option to btool to use the null parser, and modified my programs to reflect this. it might not be an ideal patch, as i don't really know what i'm doing, but it works afaict. Hi Ashley, I've applied your btool null parser patch to the latest CVS, so your perlbarry package doesn't need a special patch anymore. I've also added your perlbarry package under contrib/perlbarry. I'm working through testing the scripts, and so far found just a few small issues: - if two devices are plugged in at once, the scripts get confused... this isn't too big a deal, but would be nice to have handled better - some of the scripts embed binary ASCII characters in the sources... this confuses tools like diff, making it think the files are binary... could those characters be substituted for escaped codes? - when running the barrybkup script, some database directories end up empty, and tar gives a blast of errors in those cases Also, in testing your btool null parser patch, I found a bug in the null parser itself, which caused the record unique IDs to be incorrect. I fixed this as well. Thanks very much for your scripts and for making them available to be added to Barry. I'm quite pleased to see the IPD file format in perl script form. - Chris - Sponsored by: SourceForge.net Community Choice Awards: VOTE NOW! Studies have shown that voting for your favorite open source project, along with a healthy diet, reduces your potential for chronic lameness and boredom. Vote Now at http://www.sourceforge.net/community/cca08 ___ Barry-devel mailing list Barry-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/barry-devel