> On Thu, 2013-12-26 at 16:34 +0100, [email protected] wrote: > > Hello, > > > > > As Tristan has been busy, I checked out the testsuite and tried > > > it... > > > > > > First observations... > > > > > > I noticed one "SKIPPED" test in the "gna" tests : > > > testsuite/gna/sr3028/testsuite.sh > > > It seems to be a good candidate for "analyze_failure" > > Fine with me. > > I plan to rewrite the fix for this issue, because I think it > > doesn't > > catch all the possible cases. > > OK, but it wasn't my intention to make MORE work for you by > submitting > half-baked patches! If you want to handle it yourself that's fine, > but > if you want to hint how I could improve it, I'll have a go.
I think large arrays should be prevented before hitting the constraint error. Never good to rely on such checks. > > > Test: 368 vhdl-93/ashenden/compliant/ch_13_fg_13_26.vhd > > > ghdl -a --std=93c vhdl-93/ashenden/compliant/ch_13_fg_13_26.vhd > > > vhdl-93/ashenden/compliant/ch_13_fg_13_26.vhd:33:20: port "a" > > > already > > > associated in primary binding > > > ...does it fail with previous revisions? > > > > I tested with ghdl_mcode, and not with ghdl_gcc... > > okay, but I can't see why they would be different. I'll put it down > to a > problem at my end for now. IIRC, ghdl_mcode doesn't try to generate code during analysis. > > > Then I modified the vests/testsuite.sh script to actually > > > elaborate > > > and run the tests ... > > > gnatmake get_entities.adb > > > > Great! > > get_entities has a bug (where "end entity" is on the last line : > fixed > here, not checked in) but there are other issues with getting some of > these tests to run. I'll work at some of them for a few more days, > and > report back when I have something a bit more coherent. I have just pushed a change to get_entities: trailing spaces are removing and only the last entity is printed. Thus the spec of that tool has changed a little bit, but that was necessary for all the tests that declare several entities. > > > This next one (second version) may actually be a ghdl failure ... > > > Test: 71 > > > vhdl-93/clifton-labs/compliant/functional/textio/read/character_array_read.vhdl > > > ... > > > ghdl -r character_array_read --assert-level=error > > > ./character_array_read:internal error: file: IO error > > > ghdl: compilation error > > > > The format of binary files is not defined by the standard. We > > should use the > > output of a previous write. > > Aah. As the contents are ASCII it didn't occur to me that this was a > binary file! But that makes sense. I can pursue that one now. I have reordered some tests, so that the files are written before being read (+ 2 new pseudo tests to generate missing files). We could remove the files from the repo. > > > The last, dynamic_package_procedure_for_loop.vhdl, now fails at > > > runtime) > > > > My fault. Fix in progress. > > Thanks! I think it's done. > As I say it seems worthwhile to try making these tests run (where > possible) - it might take some work but I'm happy for this to be my > cat > (if you and David approve!) Argh, I think that's now done: the vests test suite is clean for me. Do not hesitate to second check. Tristan. _______________________________________________ Ghdl-discuss mailing list [email protected] https://mail.gna.org/listinfo/ghdl-discuss
