| Summary: | app-portage/genlop-0.30.10-r1 -c average time calculation doesn't limit compilation ETA time package merge history to the ongoing packages category | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | peter <peter> |
| Component: | Current packages | Assignee: | Portage Tools Team <tools-portage> |
| Status: | UNCONFIRMED --- | ||
| Severity: | minor | CC: | jstein, leio, peter |
| Priority: | Normal | ||
| Version: | unspecified | ||
| Hardware: | All | ||
| OS: | Linux | ||
| See Also: | https://bugs.gentoo.org/show_bug.cgi?id=686144 | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
What's the bug here? You asked for rust, you got results for dev-lang/rust and virtual/rust, total 4 entries (2 for dev-lang/, 2 for virtual). You asked for dev-lang/rust, you got the results for just dev-lang/rust, thus 2 entries. (In reply to Mart Raudsepp from comment #1) > What's the bug here? > You asked for rust, you got results for dev-lang/rust and virtual/rust, > total 4 entries (2 for dev-lang/, 2 for virtual). > You asked for dev-lang/rust, you got the results for just dev-lang/rust, > thus 2 entries. Yes, I included that to show why the -c calculation was wrong. I quote: $ genlop -c Currently merging 8 out of 16 * dev-lang/rust-1.29.1 current merge time: 13 minutes and 38 seconds. ETA: 1 minute and 58 seconds. That calculation has been skewed by the two virtual/rust emerges of a few seconds each. The ETA could not possibly have been only 1:58, because the two real emerges had taken 29:26 and 32:44. I think I was mistaken to include 'or "genlop -t"' in the bug report. Sorry if that's caused confusion. OK, I see what you mean, but the summary wording suggests something completely differently. I'll try to fix that up and leave the rest for our regular wranglers. So you mean that -c was calculating the ETA based on the average of dev-lang/rust and virtual/rust combined, instead of correctly just dev-lang/rust Otherwise you might find golop and emlop of interest as potential genlop replacements. There's also qlop from portage-utils, but I don't think qlop has the equivalent of genlop -c. (In reply to Mart Raudsepp from comment #3) > OK, I see what you mean, but the summary wording suggests something > completely differently. I'll try to fix that up and leave the rest for our > regular wranglers. > So you mean that -c was calculating the ETA based on the average of > dev-lang/rust and virtual/rust combined, instead of correctly just > dev-lang/rust Just so. Or, it was calculating the ETA from the average of all known <cat>/rust packages. > Otherwise you might find golop and emlop of interest as potential genlop > replacements. There's also qlop from portage-utils, but I don't think qlop > has the equivalent of genlop -c. Thank you Mart. I'll have a look at those. |
When I run "genlop -c" or "genlop -t" it doesn't include the whole cat/pkg name in its calculations. $ genlop -c Currently merging 8 out of 16 * dev-lang/rust-1.29.1 current merge time: 13 minutes and 38 seconds. ETA: 1 minute and 58 seconds. $ genlop -t rust | grep second merge time: 29 minutes and 26 seconds. merge time: 4 seconds. merge time: 10 seconds. merge time: 32 minutes and 44 seconds. $ genlop -t dev-lang/rust | grep second merge time: 29 minutes and 26 seconds. merge time: 32 minutes and 44 seconds.