aboutsummaryrefslogtreecommitdiffstats
path: root/TODO.template
blob: 89a0d5ecadcc3d9485cdef5b4f9cd6d599add84d (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43

switch to .EXPORT_ALL_VARIABLES and/or .ONESHELL (as a refactor/cleanup)?
    or is that too gmake specific...

BUG: synth still seems to continue even if first build (verilog compile)
     fails

add .PRECIOUS for intermediate files we don't want to get deleted

for fpga_editor:
    DISPLAY=`echo $DISPLAY |sed s/'\.0'//` fpga_editor <.ncd file>

effort levels seem high by default:
    Overall effort level (-ol):   High 
    Router effort level (-rl):    High 

impact:
    impact -mode bscan -b build/sp605.bit -port auto -autoassign (needs testing)

requests from AJ:
    anything related to not rebuilding all the coregen when not necessary.

    e.g. want a make clean equivalent to removing the build dir but not the
    cores. This should actually be the default behavior, with different
    operator for nixing the cores.

    not auto rebuilding the cores when switching branches/commits if not
    strictly necessary. Because git touches all the files, this may be
    difficult.

    make isim/simulate will run to completion even if there were errors on
    the build.

    In the case of make isim, isim will load and run the previously valid
    simulation. Unless you happen to see the error go by in the build log,
    you will unknowingly be simulating your previous build, whereas your
    current build failed to compile.

    The solution is to have make simulate begin by deleting the previous
    simulation executable, so that it must successfully create a new one
    before loading isim.

Add linting to tests (aka, tb/*_tb.v)