[Csnd-dev] Errors with arrays
| Date | 2026-01-02 21:15 |
| From | Eduardo Moguillansky |
| Subject | [Csnd-dev] Errors with arrays |
Hi, With latest develop branch, this code generates an error for me kA[] fillarray 1, 2, 3 kB[] fillarray 2, 4, 6 kC[] = (kA * kB) * 0.5 error: error: opcode '##mul' for expression with arg types :k[;c not found, line 17 SEMERR: opcode '##mul' for expression with arg types :k[;c not found, line 17 This works: kA[] fillarray 1, 2, 3 kB[] fillarray 2, 4, 6 kC[] = (kA * kB) kC *= 0.5 Any multi term expression throws the same error, actually: for example, this fails too kC[] = kA * 2 + 0.5 |
| Date | 2026-01-02 22:24 |
| From | Victor Lazzarini <000010b17ddd988e-dmarc-request@LISTSERV.HEANET.IE> |
| Subject | Re: [Csnd-dev] [EXTERNAL] [Csnd-dev] Errors with arrays |
yes, that’s a regression introduced in the struct arrays PR. The issue is that the type string is wrong, ":k[;c” should be “k[]c”, so the opcode is not found. It seems to only happen in expression analysis. I did a bisect and found two possible commits where this regression was introduced. Can you file a ticket? We’ll probably have to ask Hlolli about it. I’m doing what I can to track where this is hailing from, but the code in that area has changed a lot from what I knew. best ======================== Prof. Victor Lazzarini Maynooth University Ireland > On 2 Jan 2026, at 21:15, Eduardo Moguillansky |
| Date | 2026-01-02 23:02 |
| From | vlz |
| Subject | Re: [Csnd-dev] [EXTERNAL] [Csnd-dev] Errors with arrays |
Never mind, I found the issue and made a PR. Thanks for spotting this. > On 2 Jan 2026, at 22:24, Victor Lazzarini <000010b17ddd988e-dmarc-request@LISTSERV.HEANET.IE> wrote: > > yes, that’s a regression introduced in the struct arrays PR. The issue is that the type > string is wrong, ":k[;c” should be “k[]c”, so the opcode is not found. It seems to only > happen in expression analysis. > > I did a bisect and found two possible commits where this regression was introduced. > > Can you file a ticket? We’ll probably have to ask Hlolli about it. > I’m doing what I can to track where this is hailing from, but the code in that area > has changed a lot from what I knew. > > best > ======================== > Prof. Victor Lazzarini > Maynooth University > Ireland > > > > > > >> On 2 Jan 2026, at 21:15, Eduardo Moguillansky |