Re: [casper] Compiling error
Morning James I've added some more debugging to the script, so check out the latest version. Have you used the debug logging before? Try setting: casper_log_groups={'all'} And rerunning your scripts. Cheers Paul On 1 June 2015 at 15:24, James Smith jsm...@ska.ac.za wrote: Hi Paul, I wasn't sure what debug information I'd be able to get from this script directly, but I found that it gets called by write_info_table.m in the same folder. So at around line 16 in the script I added a try .. catch around the call to the sblock_to_info function, with e.stack and error(['Error occurred when processing ', tag, '.']) Should that tell me a little more info about where the problem occurs? Or is there another more elegant way to get that information? Regards, James On Mon, Jun 1, 2015 at 10:54 AM, Paul Prozesky paul.proze...@gmail.com wrote: Hi James You can see that your error is happening in this section of that script: if isempty(strtrim(value)) == 1, error('Empty value?'); end So add some debug logging there and rerun the gen_xps_add_design_info part of the build. Cheers Paul On 1 June 2015 at 10:39, James Smith jsm...@ska.ac.za wrote: Hi Paul, Thanks, I have found that script. I can't think of a way to go about finding out which part it has a problem with though, except possibly starting several different compiles, removing a component from each, and seeing which ones pass / fail. Is this a sound approach? James On Mon, Jun 1, 2015 at 10:34 AM, Paul Prozesky paul.proze...@gmail.com wrote: Hi James Check the sblock_to_info script in the +design_info (no, the plus isn't a mistake, it's a matlab thing) folder in your xps_library folder. Something is that script is not happy with one of the blocks in your model file. Cheers Paul On 1 June 2015 at 10:27, James Smith jsm...@ska.ac.za wrote: Hi all, I'm working on a simple design for a two-stage high-resolution spectrometer, which almost gets finished compiling, sysgen seems to finish, and then at the end fails with a most uninformative error message: ## ## Creating EDK files ## ## Error using design_info.sblock_to_info (line 18) Empty value? Anyone else seen this problem before? The sysgen*.log files contain basically zero information, and I've grepped for as many combinations of the design_info.sblock_to_info string as I can think of, to no avail. If anyone has some insight that can point me in the right direction, I'd appreciate it. I can send the .slx file to anyone who would like to look at it, I didn't think the mailing list would be an appropriate place to attach it (about 7 MB). Regards, James
Re: [casper] Compiling error
Hi all, Matt - thanks, that fix seems to have worked. The design gets past this now, ultimately to fail because it's too big for the device... but that's another story. Paul - I don't have that much familiarity with Matlab (I presume the debug logging to which you're referring is a Matlab feature?). Thanks for that - I'll fetch the latest changes from Github and see what you've put in. Will report whether it helps soon. Regards, James On Tue, Jun 2, 2015 at 12:06 PM, Paul Prozesky paul.proze...@gmail.com wrote: Morning James I've added some more debugging to the script, so check out the latest version. Have you used the debug logging before? Try setting: casper_log_groups={'all'} And rerunning your scripts. Cheers Paul On 1 June 2015 at 15:24, James Smith jsm...@ska.ac.za wrote: Hi Paul, I wasn't sure what debug information I'd be able to get from this script directly, but I found that it gets called by write_info_table.m in the same folder. So at around line 16 in the script I added a try .. catch around the call to the sblock_to_info function, with e.stack and error(['Error occurred when processing ', tag, '.']) Should that tell me a little more info about where the problem occurs? Or is there another more elegant way to get that information? Regards, James On Mon, Jun 1, 2015 at 10:54 AM, Paul Prozesky paul.proze...@gmail.com wrote: Hi James You can see that your error is happening in this section of that script: if isempty(strtrim(value)) == 1, error('Empty value?'); end So add some debug logging there and rerun the gen_xps_add_design_info part of the build. Cheers Paul On 1 June 2015 at 10:39, James Smith jsm...@ska.ac.za wrote: Hi Paul, Thanks, I have found that script. I can't think of a way to go about finding out which part it has a problem with though, except possibly starting several different compiles, removing a component from each, and seeing which ones pass / fail. Is this a sound approach? James On Mon, Jun 1, 2015 at 10:34 AM, Paul Prozesky paul.proze...@gmail.com wrote: Hi James Check the sblock_to_info script in the +design_info (no, the plus isn't a mistake, it's a matlab thing) folder in your xps_library folder. Something is that script is not happy with one of the blocks in your model file. Cheers Paul On 1 June 2015 at 10:27, James Smith jsm...@ska.ac.za wrote: Hi all, I'm working on a simple design for a two-stage high-resolution spectrometer, which almost gets finished compiling, sysgen seems to finish, and then at the end fails with a most uninformative error message: ## ## Creating EDK files ## ## Error using design_info.sblock_to_info (line 18) Empty value? Anyone else seen this problem before? The sysgen*.log files contain basically zero information, and I've grepped for as many combinations of the design_info.sblock_to_info string as I can think of, to no avail. If anyone has some insight that can point me in the right direction, I'd appreciate it. I can send the .slx file to anyone who would like to look at it, I didn't think the mailing list would be an appropriate place to attach it (about 7 MB). Regards, James
Re: [casper] Compiling error
Hi James You can see that your error is happening in this section of that script: if isempty(strtrim(value)) == 1, error('Empty value?'); end So add some debug logging there and rerun the gen_xps_add_design_info part of the build. Cheers Paul On 1 June 2015 at 10:39, James Smith jsm...@ska.ac.za wrote: Hi Paul, Thanks, I have found that script. I can't think of a way to go about finding out which part it has a problem with though, except possibly starting several different compiles, removing a component from each, and seeing which ones pass / fail. Is this a sound approach? James On Mon, Jun 1, 2015 at 10:34 AM, Paul Prozesky paul.proze...@gmail.com wrote: Hi James Check the sblock_to_info script in the +design_info (no, the plus isn't a mistake, it's a matlab thing) folder in your xps_library folder. Something is that script is not happy with one of the blocks in your model file. Cheers Paul On 1 June 2015 at 10:27, James Smith jsm...@ska.ac.za wrote: Hi all, I'm working on a simple design for a two-stage high-resolution spectrometer, which almost gets finished compiling, sysgen seems to finish, and then at the end fails with a most uninformative error message: ## ## Creating EDK files ## ## Error using design_info.sblock_to_info (line 18) Empty value? Anyone else seen this problem before? The sysgen*.log files contain basically zero information, and I've grepped for as many combinations of the design_info.sblock_to_info string as I can think of, to no avail. If anyone has some insight that can point me in the right direction, I'd appreciate it. I can send the .slx file to anyone who would like to look at it, I didn't think the mailing list would be an appropriate place to attach it (about 7 MB). Regards, James
Re: [casper] Compiling error
Hi James Check the sblock_to_info script in the +design_info (no, the plus isn't a mistake, it's a matlab thing) folder in your xps_library folder. Something is that script is not happy with one of the blocks in your model file. Cheers Paul On 1 June 2015 at 10:27, James Smith jsm...@ska.ac.za wrote: Hi all, I'm working on a simple design for a two-stage high-resolution spectrometer, which almost gets finished compiling, sysgen seems to finish, and then at the end fails with a most uninformative error message: ## ## Creating EDK files ## ## Error using design_info.sblock_to_info (line 18) Empty value? Anyone else seen this problem before? The sysgen*.log files contain basically zero information, and I've grepped for as many combinations of the design_info.sblock_to_info string as I can think of, to no avail. If anyone has some insight that can point me in the right direction, I'd appreciate it. I can send the .slx file to anyone who would like to look at it, I didn't think the mailing list would be an appropriate place to attach it (about 7 MB). Regards, James
Re: [casper] Compiling error
Hi Paul, Thanks, I have found that script. I can't think of a way to go about finding out which part it has a problem with though, except possibly starting several different compiles, removing a component from each, and seeing which ones pass / fail. Is this a sound approach? James On Mon, Jun 1, 2015 at 10:34 AM, Paul Prozesky paul.proze...@gmail.com wrote: Hi James Check the sblock_to_info script in the +design_info (no, the plus isn't a mistake, it's a matlab thing) folder in your xps_library folder. Something is that script is not happy with one of the blocks in your model file. Cheers Paul On 1 June 2015 at 10:27, James Smith jsm...@ska.ac.za wrote: Hi all, I'm working on a simple design for a two-stage high-resolution spectrometer, which almost gets finished compiling, sysgen seems to finish, and then at the end fails with a most uninformative error message: ## ## Creating EDK files ## ## Error using design_info.sblock_to_info (line 18) Empty value? Anyone else seen this problem before? The sysgen*.log files contain basically zero information, and I've grepped for as many combinations of the design_info.sblock_to_info string as I can think of, to no avail. If anyone has some insight that can point me in the right direction, I'd appreciate it. I can send the .slx file to anyone who would like to look at it, I didn't think the mailing list would be an appropriate place to attach it (about 7 MB). Regards, James
Re: [casper] Compiling error
Hi Paul, I wasn't sure what debug information I'd be able to get from this script directly, but I found that it gets called by write_info_table.m in the same folder. So at around line 16 in the script I added a try .. catch around the call to the sblock_to_info function, with e.stack and error(['Error occurred when processing ', tag, '.']) Should that tell me a little more info about where the problem occurs? Or is there another more elegant way to get that information? Regards, James On Mon, Jun 1, 2015 at 10:54 AM, Paul Prozesky paul.proze...@gmail.com wrote: Hi James You can see that your error is happening in this section of that script: if isempty(strtrim(value)) == 1, error('Empty value?'); end So add some debug logging there and rerun the gen_xps_add_design_info part of the build. Cheers Paul On 1 June 2015 at 10:39, James Smith jsm...@ska.ac.za wrote: Hi Paul, Thanks, I have found that script. I can't think of a way to go about finding out which part it has a problem with though, except possibly starting several different compiles, removing a component from each, and seeing which ones pass / fail. Is this a sound approach? James On Mon, Jun 1, 2015 at 10:34 AM, Paul Prozesky paul.proze...@gmail.com wrote: Hi James Check the sblock_to_info script in the +design_info (no, the plus isn't a mistake, it's a matlab thing) folder in your xps_library folder. Something is that script is not happy with one of the blocks in your model file. Cheers Paul On 1 June 2015 at 10:27, James Smith jsm...@ska.ac.za wrote: Hi all, I'm working on a simple design for a two-stage high-resolution spectrometer, which almost gets finished compiling, sysgen seems to finish, and then at the end fails with a most uninformative error message: ## ## Creating EDK files ## ## Error using design_info.sblock_to_info (line 18) Empty value? Anyone else seen this problem before? The sysgen*.log files contain basically zero information, and I've grepped for as many combinations of the design_info.sblock_to_info string as I can think of, to no avail. If anyone has some insight that can point me in the right direction, I'd appreciate it. I can send the .slx file to anyone who would like to look at it, I didn't think the mailing list would be an appropriate place to attach it (about 7 MB). Regards, James