Hi Joep,
Back on this topic... Last revision of jallib.py has the following features:
* when auto-generate samples from *all* board files and *all* test files,
it checks if the board file has the following tag: "@jallib preferred". If
so, it generates samples. If not, it skips it. This means only one board per
PIC must have this tag, else it would overwrite samples
* you can generate all sample, given a board file:
jallib sample -a path/to/sample/dir -b path/to/board/file
It generates samples, whether there's the "@jallib preferred" tag or not.
So, now we have all the granularity:
* all samples from all board and test files
* all sample from one board file and all test files
* one sample from one board file and one test file
We still have to deal with the sample's prefix. I suggested "sample_", but I
as said, already used by Eur. Now the magic weapon: let's consider Eur has
to rename his samples ! :)
Cheers,
Seb
2008/12/15 Sebastien Lelong <[email protected]>
> To deal with multiple board files, your idea of setting a tag seems nice.
> But it also shouldn't prevent user from generating lots of samples with
> their board, with as few typing as possible.
>
> Here's what I plan to implement:
>
> * deal with a "@jallib preferred-board" tag within board files, as you
> suggest
> * when using "-a" option, will preferably use these board files (it won't
> look for duplicate board files, just select the ones having these tags. this
> means if there's only one version of a board for a given PIC, and if this
> board doesn't have this tag, it won't be considered
> * when using "-b", "-t" and "-o" option, you can force to generate
> whatever you want. To save some typing, I'll add the following: if "-b"
> alone is used, it will use this board file whether there's a preferred tag
> or not, search for tests and generates and potential samples (output files
> being automatically saved in the correct place)
>
>
> Finally, about the prefix, for now, it produces "sample_<test name>.jal"
> output files. Eur uses "sample_" prefix, so there might be collisions. I
> suggest: Eur renames his samples, we use "sample_" prefix as a reserved one.
>
>
> Seb
>
> 2008/12/15 Joep Suijs <[email protected]>
>
>
>> Hi Seb,
>>
>> I tried this but does not run on my computer, because it is not so
>> easy to invoke JAL with my setup. Later this week (or in the next two
>> weeks of holiday :) I'll try and setup a directory with all required
>> settings and retry.
>>
>> I do like the general idea that all board/test combinations are
>> generated. There are some issue's to concider though:
>> 1. there could be multiple board files for one pic-type (if others
>> start to use board/test files). I think there should only be one of
>> them used for sample generation. Maybe put a keyword in the board file
>> to indicate it should be used for sample generation?
>>
>> 2. Compile is a good way to filter out samples that won't work for
>> sure. And we agreed that we can't test everything. But if we discover
>> that a certain combination won't work, what can we do? Fix it is the
>> most obvious. But if we can't within a short time?
>> The best option i see now is to check for known not-working devices in
>> the lib and throw a compile error.
>> And to reduce the number of incorrect (compilable but not working
>> examples) is to check for required bits in stead of registers. ADC on
>> 16f88 is an example - setup is not properly supported, but since the
>> register name is the same (but not the bit fields within the chip) it
>> compiles and most of adc works. Setting the bits would break
>> compilation (is that what we want in this case?) but would also help
>> to prevent incorrect samples...
>> Bbut now I am back at the famous hardware version issue and this is
>> not directly related to sample generation.
>>
>> But the short version of my message: this is a mayor step in the right
>> direction. In time we should be able to handle multiple board files
>> for the same chip.
>>
>> Joep
>>
>> >>
>>
>
>
> --
> Sébastien Lelong
>
> http://www.sirloon.net
> http://sirbot.org
>
--
Sébastien Lelong
http://www.sirloon.net
http://sirbot.org
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"jallib" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/jallib?hl=en
-~----------~----~----~----~------~----~------~--~---
